01.31.2005 20:36

Many links

On your Mac, try this:
  Highlight some text in your web browser
  Apple-C copy it
  Go to a terminal or xterm window and run

pbpaste dumps the contexts of the clipboard to standard out. Use pbcopy to put text into the clipboard (aka pasteboard). I did not know that there are 4 different pasteboards. Crazy. What other little handy programs are laying around the Mac OSX filesystem?

It has been a busy day with sampling two more Santa Barbara core sections, measuring them on the kappa bridge and present my initial results to the Driscoll group.

The Griffin PowerMate looks like a fun little input device. Might may a good complement to the space mouse if I can ever finish the driver for 3D stuff. Would be great for editing video and audio if it works smoothly with everything. On sale for $39.99!

Looks like Thursday sugarbush will be getting a new Mandrake install. That will be nice to get a clean start. There is a new interview with the president of Mandrake here

Exxon Mobil profits exceed $25bn, Revenues hit a record $298bn. Wow. (BBC)

The U.S. Army's field manual guide to Cryptanalysis

white_dune is a graphical VRML97 editor, simple NURBS/Superformula 3D modeler and animation tool. Even has an OSX package. I need to try this.

Why do I need shell sql to use sqlite from the shell? I am already doing that. Oh... it lets me batch commands and then commit them. Might be some serious performance gains over my super slow shell scripts.

Posted by Kurt | Permalink

01.31.2005 06:22

MiNews RSS

The screenshots for MiNews looks more like NewsFire. Seems to take my opml file from NewsFire once I deselected slashdot. Then crashed. This is not useable yet.

Also kind of cool, but does not yet support Nannoblogger: BloGTK

Posted by Kurt | Permalink

01.30.2005 20:56

gnuplot multiplot

libferris - Interesting virtual filesystem that lets you see inside db and tar files too. Nifty idea.

I was saved by USENET today. I was trying to figure out how to get multiple tall skinny plots next to each other in a row. I posted via Google groups and 2 hours later I had the answer to check out multiplots.
  set size 0.7,0.7
  set origin 0.1,0.1
  set multiplot
  set size 0.4,0.4
  set origin 0.1,0.1
  plot sin(x)
  set size 0.2,0.2
  set origin 0.5,0.5
  plot cos(x)
  unset multiplot

Posted by Kurt | Permalink

01.30.2005 12:14


Another thing that I have been wanting to write for the last 2 years is a low-level segy library. Something that is simple enough that most c++ programmers could open up a segy file and start working with it. Having to know field numbers is truly insane and not productive. I was thinking that it needed to have a system for people to define how each segy type really is. This would, for example, with the SIO xstar data say how fields really are mapped. There really does not seem to be a segy standard in practice. In sioseis, people are always remapping fields between each other. These config files could also have an entry like the Unix magic/file database for detecting data formats by providing finger prints. There would be a truly standard compliant spec file provided with the library plus format files for each vendor device that writes out funky stuff. This would allow extraction of fields from the unstructured text block in the initial header. Basic operation should be simple. I should be able to verify the segy file and find out what type it is. Get a bounding box and shot range. Be able to dumb to ASCII a particular shot, etc. Then whatever applications could be layered on top. Jeff D. could write his Matlab trace viewer easily. We could make a python version of the basic features in sioseis with just a couple of days. The list goes on.

The trigger for writing this entry was thinking about how I could use libsegy to import all the header info into an SQL database. Then we could start making interpretation programs. It would not be too hard to start building tools that helped you define a layer by picking from processed images. I had talked to Jeff Dingler about using g3data to pick time/shot pairs. We could then drop this back into a database. Then build a mesh surface between lines for this picked layer.

The basic idea is that when we have a rock solid and flexible read/write open source library (LGPL?) people could build whatever suited them quickly.

Random slashdot link: Hints to new consultants

Posted by Kurt | Permalink

01.30.2005 11:39

MetaWeb Log API

I am not really sure what the MetaWeblog API is, but it is supported by Archipelago, Blojsom, Radio Userland, NetNewsWire, Python Desktop Server, and Zoe. Which I put here more as a list of programs that do blogging. Manila and Moveable Type are in the pricy non-open source catagory. What about Blosxom, Conversant, TypePad, and WordPress?

Checking out newnewswire lite which says it is freeware with donations. Looks more like NewsFire or is it the other way around?

Posted by Kurt | Permalink

01.30.2005 08:16

X-Core requirements & NewsMac

Check out these dunes on Mars! (Photojournal PIA07306)

On the RSS News Reader front, all you FireFox RSS news readers are going to shake your heads at me. "If you just used an open source tool..." I have to say that I appreciate the simplicity of NewsFire. I tried NewsMac for about 15 minutes and got totally frustrated by the three level hierarchy. I could not figure out how to nuke Spanish, French, Japanese, Sports. I just want MY feeds. Import from OPML was easy. But the interface was so cluttered. It is easy to tell that the authors of NewsMac have worked very hard on the program. I just do not have the energy to figure it all out.

I am going to try something new with Nanoblogger: an attachments directory with year-month subdirectories. I often have stuff that I do not really know where to put on the web and maybe the best thing is to order them by date. http://schwehr.org/blog/attachments/. My first attachment will be an opml export of my feeds that I am using in NewsFire. Some of these are broken, but this is what I have in there. If you want to see what I have in my feed, look at newsfire-29Jan2005.opml

Now back to my thesis. I have been thinking more about what I would like to do in the next version of X-Core when I get a chance. The main thing I need is to have a better core description, data entry, and export tool. Right now, I am resorting to g3data and emacs to do data entry along with producing a parallel Excel spreadsheet. The Excel spreadsheet really only serves the purpose of making handwriting legible, allowing the addition of a photo without a glue stick, and printing/sharing the data. The xls files are not good for getting data into anything else. Here is what one looks like:

Here are my ideas of what this mythical open source core description program should be able to do:
  1. Be able to import and register images. This includes images from multi-spectral core logger CCD/CMOS strips, SIO style overhead multi-section photos, and oblique digital camera images. With the SIO style core photos, this is a matter of picking out the rulers and the corners and edges of each section on the table. For the oblique or otherwise distorted images, it should be able to ortho-rectify the image using say, the edge of the core liner that have been measured.
  2. Allow the user to designate points, lines, and polygons on the core and attach either a description, photo, or measurement, or category.
  3. Categories might be like the ODP lithology types.
  4. Be able to generate color or greyscale graphs from the core photo. The ability to skip strips of the core are critical for getting around things like shell fragments when looking at the characteristics of the mud.
  5. The ability to do multilayer imaging would be awesome. If I could also drop in x-rays and other types of imager data and have it be like layers in Photoshop that would be a big help.
  6. Be able to import/export simple ASCII csv files, xml, and sql would make integration with a wide range of tools a snap.
  7. It has got to be cross platform. I can not force people into one OS. The age of Windows only tools must die.
  8. It must be open source so that folks can customize it at all levels. It also can not cost or hardly anyone will use it at the university level. This has to be a part of standard operating procedure. This can not be something that could possibly go away based on corporate whims!
  9. No way can it be web based. It needs to be a simple stand alone application that a user can grab and use where ever. This might be in Haughton Crater, on a ship, in the lab, and who knows where else!
  10. I would really like to hook this up to a 3D viewer. Click on the core in the 3D world and it brings up the core in this app. In the app, select a core and tell it to have the core be the focus in the 3D world.
  11. Import/export from an online repository. Allow sharing of this data. If I have finished and published a paper on a core, let me hit one button and upload this to a server so that anyone can download the core.
  12. REVISION CONTROL! Be able to take a photo each time the core is sampled and have this as a layer. Then when someone wants to know the status of the core, the look in the viewer and see that the section they wanted is down to nothing with only core liner showing.
  13. USERS! Be able to have each person enter their comments on a core. Then if we can see who described what and when. If you see comments from me on micro fossils and later comments that say something else from someone like Dick Norris or Chris Charles, you will know to trust comments by those guys over me!
Miscellaneous link for this entry: simple-glade-codegen.py. If you use glade, this will help generate code for pygtk. Not that I have tried glade in years. It is going to take fink an hour or two to build glade2 while I do other stuff.

Posted by Kurt | Permalink

01.29.2005 22:18

Oil in Libya

Libya has some pretty cool geology, not that I want to visit. When I heard last year that American companies were back in that this would be HUGE business. Here is a BBC article about US Oil companies getting back in there. We used Libya as our final exam area for Basin Analysis and all I can say is wow!

Posted by Kurt | Permalink

01.29.2005 16:51


Newsfire just went shareware asking for $20. Doh. Guess I had better find something else, preferably open source. Newsfire is a nice program. Very clean.

I am going to give NewsMac a try. It says it has ipod syncing support. That would be nice.

Posted by Kurt | Permalink

01.29.2005 12:53

BPSIO core descriptions

Just entered bpsio-5g-sec2 description into Excel. I then dumped a pdf via print. This is the first sheet for this cruise to have the photograph inserted in the left hand column. That is easy to do now that I have processCore working again and the photo xpd files already created. Now, how do I control excel so the pdf is just one page. This 4 page business is painful!
Core Descriptions Directory
PNG of 5g-sec2

Posted by Kurt | Permalink

01.29.2005 10:26

Added the static index

I should now have a static index that gets updated when I tell it to. This is from an ugly shell script called nb-build-index.bash. Check out the links on the right hand column of the front page.

Posted by Kurt | Permalink

01.29.2005 07:49

More cleaning - free math help

I am down to 1800 emacs in my inbox. Whew! Some serious cruft in there.

The free math help is not from me! It is from the UCSD math department:
Conference poster advice

SIO Tsunami webpage

Making a sphere in parametric mode with gnuplot:
  set parametric
  set urange [-pi/2:0]
  set vrange [0:2*pi]
  set isosamples 10,10
  splot  'as2-slump-sample-100.xyz' \
       ,'as2-slump-sample-100.xyz' using 4:5:6\
       ,'as2-slump-sample-100.xyz' using 7:8:9\
       , cos(u)*cos(v)/3,cos(u)*sin(v)/3,sin(u)/3
From Jeff Love at the USGS: Software for paleo-vector analysis (directions and intensities)
  Love, J. J. & Constable, C. G., 2003. Gaussian statistics for
  palaeomagnetic vectors, Geophys. J. Int., 152, 515-565.

A good example of a research cruise web page:

Posted by Kurt | Permalink

01.28.2005 19:35

Cleaning house

There are 2600 emails sitting in my inbox going back to July 2003. It is time to clean house a bit. So far I have gotten rid of 300 messages. I will use this log entry to cache some of the more interesting items I find that are of public interest.

Community Model for Coastal Sediment Transport - Chris Sherwood who put up the java applets for basic water parameters like settling velocity. Would be nice to keep adding all we can to a large community body of software for analyzing these types of problems.

ArsDigita University online computer science courses. I should really watch Database Management Systems

Danah Boyd recommended looking at David Laidlaw's work for inspiration on scientific visualization.

NGDC Coastal Data for the SoCal area.

ODP Lat Lon search

ODP "hole trivia" - Where the holes are. However, they are only to the nearest minute. Ouch.

ODP Database front door.

Geometrics sells a lot of the towed magnetometers that we use. They have a new digital streamer of hydrophones at AGU that looks interesting called the GeoEel

Killer Sediment waves in the EuroSTRATAFORM area - Bottom center of the top image.

ipod hacks

Data Structures and Algorithms with Object-Oriented Design Patterns in Python an online book by B. R. Preiss.

SIO Pier Chlorophyll from a CTD sensor sampling fluorescence.

Fiona's recommended Chicken Tikka Masala

GemSys makes magnetometers.

Posted by Kurt | Permalink

01.28.2005 14:02


I have always had a strong dislike for Fortran. It only got worse after seeing what the compiler vendors did with F90 and F95. Why would I want to pay $900 for a junky Fortran compiler? I have tried 3 different compilers at about that price each. I only use Fortran for legacy code at this point.

How Not to Write FORTRAN in Any Language

Of course there is a rebuttal from JPL

The f90 homepage

Check out snopes on the crazy deep sea creatures.

Posted by Kurt | Permalink

01.28.2005 09:59

Faxing from OSX

Just tried sending my first Fax from OSX. I opened a scanned image in Preview and selected print. I then inserted the fax number from my address book, discovering that a 1- in the number makes life easier. Why does it insist on defaulting to a bluetooth modem when I have bluetooth turned off? I select the internal modem and type in a cover sheet being carful to not press return and use Apple-V to past in carrage returns. Then send it. The "Internal Modem" printer window pops up and starts dialing. The first one went fine until it said sent and then just hang then looking busy and saying finished, but then not going on two document 2. I stopped the jobs, deleted the first one, and restarted to stat doc 2 would go. That dialed, got the fax machine and just sat there saying connecting. Hey Apple, how about some more debugging info. Not much int the window, so I went into /var/log/system.log, where I get this kind of thing:
Jan 28 09:30:46 localhost kernel: AppleModemOnHoldNub::start--registered service
Jan 28 09:30:46 localhost kernel: AppleModemOnHoldService: Initializing
Jan 28 09:30:46 localhost kernel: AppleModemOnHoldService: Probing
Jan 28 09:30:46 localhost kernel: AppleModemOnHoldService: Starting
Jan 28 09:30:46 localhost /System/Library/Extensions/IOSerialFamily.
OnHold.app/Contents/MacOS/AppleModemOnHold: ModemOnHoldOpen >> 
Jan 28 09:30:46 localhost /System/Library/Extensions/IOSerialFamily.
OnHold.app/Contents/MacOS/AppleModemOnHold: MOHServiceAddedProc >> 
Jan 28 09:30:46 localhost /System/Library/Extensions/IOSerialFamily.
OnHold.app/Contents/MacOS/AppleModemOnHold: MOHServiceAddedProc <<< 
Jan 28 09:34:07 localhost kernel: AppleModemOnHoldService: Stopping
Jan 28 09:34:07 localhost kernel: AppleModemOnHoldNub::stop()
Jan 28 09:34:07 localhost kernel: AppleModemOnHoldService: Freeing
That does not tell me much.

Posted by Kurt | Permalink

01.28.2005 08:51

The iPod photo

I just found out that the iPod Photo can not actually take pictures. Crazy. I thought it had a built in camera. Looking at the iPod accessories, I do not see a camera attachment. Yo Apple, what is the deal? Cameras a so cheap that it is going to be hard to find a cell phone without one soon. That will cause more problems with places not letting you take your cell phone along, but that is another issue.

The Griffin iTalk rocks one you add a microphone to it. How about Griffin making a camera attachment? Stick the image into a wav file or something to shoehorn it in there. Or make a camera that can directly dump frames to the iPod. Or do both.

Posted by Kurt | Permalink

01.28.2005 07:01

Ames rover program

In this month's Wired magazine one of their "wired" terms is "Blogging your book." One thing that I would like to write about is the history of the rover program at NASA Ames. I certainly do not know all the details since I was only there for a while but I'll start with a time line of events that I know.
  • 1991 - Bill Borucki sent me over N-244 to meet Butler Hine
  • 1992 - TROV goes to the south shore of Lake Tahoe
  • 1992 - TROV goes to Antarctica under the ice in Lake Hoare, Dry Valleys
  • 1993 - I take the Stanford Space Systems Engineering class. Marsokhod comes to the US with its Russian engineering team. Visits Stanford.
  • 1994 - Virtual Planetary Exploration Lab with McGreevy
  • 1995 - MAPS program take TROV to Mono Lake.
  • 1996 - I start working in the Intelligent Mechanisms Group
  • 1996 - Desert96: Marsokhod goes to Arizona for a field test.
  • 1997 - CMU Nomad rover goes to the Atacama Desert
  • 1997 - Ames stereo pipeline (STP), VEVI 5, and MarsMap
  • 1997 - 4th Planet builds Vermilion viewer with Java/VRML/EAI. Those technologies turn out to seriously suck despite heroic efforts by 4th Planet.
  • 1997 - Laurent and I create Viz to follow on to VEVI and MarsMap
  • 1998 - CMU/Ames team does to Antarctica to prepare for Robotic Antarctic Meteorite Search
  • 1998 - First start participating in Pascal's Haughton Mars Project on the impact crater on Devon Island.
  • 1998 - Take the stereo rig and pipeline to Yellowstone hot springs and geysers.
  • 1998 - I move to CMU for the Intelligent Robotics Program
  • 1999 - Marsokhod field test - first with Lisp on board
  • 1999 - Mars Polar Lander, DS-? (the penetrators) and ? crash
  • 2000 - I go to grad school and try to not think about robots for a while
  • 2004 - MER A & B land on Mars and kick butt!
Things I am missing on dates:
  • Marsokhod field test at Mars Hill
  • Marsokhod field test in Hawaii
  • TROV dives in McMurdo Sound. 1994 or 1995?
  • RoCS - Rover Control Station task tries to integrate VEVI and WITS
  • Dex/Poindexer by Hans, Dave
  • Snake robots
  • Santa Monica rover demo
  • Chernobyl/CMap project
  • And tons of other stuff
  • Rocky 7 field test at Silver Lake
Even with this list, I am leaving out things that did not have direct impact on me. There are a ton of Rocky field tests, Dave Miller switching JPL to smaller robots, Ambler, Dante I, Dante II, the NavLabs that all had an impact on robotics at Ames.

I am also missing things about the evolution of the computation hardware and software solutions that we tried to put on these vehicles. Things like when TROV got VxWorks and how. Jay Steele putting Control Shell on the Marsokhod and Mel robots. Dan and I taking Control Shell off of Marsokhod so it was just NDDS. Anne and ? taking NDDS off of Marsokhod and switching to Ace/Dao. When we switched from VME to the ruggedized zillion PCI slot x86 machines. The evil attempt to use CompactPCI (yuck). That little floating ball inside and outside the space station that Hans and Sib made. Working with the upstairs AI guys on many projects. The influence of TCA and RTC on networking and planning.

Think of a giant Amoeba that streaches over NASA Ames, CMU, Stanford, JPL, and more. That would be the best way to describe the progress of these systems and the knowledge of working on them.

From the SIO list serv:
Two weeks ago I sent out the following advertisement for 3
undergraduate volunteers to join us during a three week cruise to
Samoa this coming April. I sent it out to the entire list of
undergraduates with an interest in Earth Sciences and on the Volcano
listserver. Remarkably I have not received one application or inquiry
from any of our UCSD students, while the response from around the US
has been very good, even when considering the fact that our cruise may
fall together with finals week on most campuses.  Maybe either one of
you knows of a good student that we might want to take on our cruise?
The students should have a good background in Petrology of basalts, if
possible. Note that our deadline is this coming Saturday. Today I sent
out a reminder to the UCSD undergraduates as well.
I am very surprised that students here have not jumped all over this. When I was an undergraduate, I would have jumped all over this kind of thing. I went on trips to Yellowstone, Mono Lake, and Eastern Nevada. I never got asked to go on an ocean going cruise or I would have. For my TTN136B cruise to the Humboldt area, I got three SUPER undergrads from Humboldt State who went along and helped core and run the hydrosweep.

Posted by Kurt | Permalink

01.27.2005 15:53


IRM-manual PDF - where IRM == the Institute for Rock Magnetism at the University of Minnesota. It is handy to have a lab manual. I want to convert volume susceptibility to mass normalized susceptibility. This should be assuming a 10cc size. So to convert is should be something like MassΧ = VolΧ * 10 cc / mass in g, where Χ is Chi or bulk susceptibility.

LinkQuest of San Diego sold two TrakLink systems for AUVs to University of Southern Mississippi and the Memorial University of Newfoundland. Both of these will by on ISE Explorer AUVs. ISE products

This morning I was just a block away from the LinkQuest office and did not know it. They are no where near the water.

everyvideogame.com - Things to do after I finish my thesis!

Bazaar del Mundo

To cook a Leek:
Leek recipe
Cut down center but not all the way
Swish in water until all grit/dirt is out
Cover in butter, sprinkle with sweet basil, thyme, and parsley
Wrap in foil
Cook at 380degF for 25 minutes in the toaster oven
Worked pretty well

Posted by Kurt | Permalink

01.27.2005 10:40

Starting back into TeX/LaTex

I started writing my next paper and have gone back to using TeX. I need to figure out how to export EndNote to BibTeX. No more word. I hate writing with word. Need to figure out how I want to do TeX editing. Lisa uses TeXShop. I prefer X based tools where I can use emacs.

BRL-CAD - "constructive solid geometry solid modeling system that includes an interactive geometry editor, ray tracing support for rendering and geometric analysis, network distributed framebuffer support, and image and signal-processing tools." And has Mac OSX support. Need time to try some of these links I keep finding.

samldap - Samba + ldap. Maybe this will help when I setup a lab network someday along with LDAP Account Manager or maybe Apple will put their XServe GUI tools on all Macs. I tried to setup OpenLDAP about 5 years ago and to call it confusing is an understatement.

Apples Mail.app just considered my AGU convener email as spam. Grrr.

Sometimes when I burn CDs on my laptop, Mac OSX leave the temp volume around on /Volumes and it takes a reboot for it to disappear. It will not umount and I am a bit nervous to do an "rm -rf".

DHL seems a bit scetchy. I had a DHL package I wanted to get out today. I found the nearest drop box and it was about a 4 mile drive. When I got there, their drop says "Airborn Express" with ZERO indication that they are DHL. No yellow or anything. Then I opened the drop box and there were DHL yellow mailers and all. Come on guys! How hard is it to slap a sticker on the outside of the drop box that says DHL?

Another link - Python code metrics with PyMetrics Linux Journal article

Condor Earth Technologies came by the USGS sometime during the 94-95 timeframe showing off milspec tablet PC's with the Leica laser range finding binoculars with magnetic compass. At the Dec AGU, I saw that people are starting to pick up this idea again as these things get cheaper. I just wish people would stop using MS Windows on these things.


Posted by Kurt | Permalink

01.27.2005 06:44

POVRAY modeler

yaprm - Yet Another POV-Ray Modeller

gas - analyze your cars gas consumption

Log Application Usage - Track what OSX applications you use for that human factors paper you were thinking of writing. Lots more fun and useful apps. TreeSize is exactly what I was wanting yesterday as my laptop fills up yet again.

Installing Debian

burn to external DVD for iDVD. This link is for Becca.

Edit and sync iCal calendars on multiple machines

openModeller is a static spatial distribution modelling library. It provides a uniform method for modelling distribution patterns using a variety of algorithms. It can be used to predict species distribution (fundamental niche) based on a set of georeferenced occurrence points and a set of environmental layers.

Posted by Kurt | Permalink

01.26.2005 14:54

MySQL would have been easier

Myche pointed out that MySQL has all sorts of cool importing stuff:



I'm starting to use Hibernate (http://www.hibernate.org/) to do
automated persistence.  I'm also thinking of using HSQLDB
(http://hsqldb.sourceforge.net/) in certain applications.  I have a
need to use an in-memory database and HSQLDB allows such a thing with
SQL92 (well, a subset).  this means I can build my quick app and if it
grows move it to MySQL without changing the Java code.

redet is on my list of things to check out and possibly add to fink. "Regular Expression Development and Execution Tool"

Posted by Kurt | Permalink

01.26.2005 13:00

AMS database craziness

All I can say is that there has to be a better way! I created this script today: build-db.bash that creates the database of cores and k15/s7 data for each paleomag cube. This is just out and out ugly. Probably 80% of the problem is that I am using bash. Python or perl would make this less painful. The rest of the problem comes from that this is not a straight import. Data for each sample takes 4 lines: 1 line of header and 3 lines of 5 measurements each. In that header is a lot of information. My sample names look like this:
   |    |||  | |   |- Optional cm decimal place
   |    |||  | |- Depth in cm
   |    |||  |- Section number.  1 is the top section (water/bottom interface)
   |    |||- (w)orking or (a)rchive half
   |    ||- (g)ravity, (p)iston, (t)rigger
   |    |- Core number
   |- Cruise abbreviation -> BPSIO 2004 == bp04
And then there is the trouble of making an SQL date out of "1/3/05 4:33 PM." It all adds up to a whacked out script. Suggestions?

Posted by Kurt | Permalink

01.26.2005 10:26

Fisher stats

Dingler's link: Tsunami Visualization
Dear Paleomagnetist/Geomagnetist

I thought you would be interested in knowing that people at the Barr Smith Library at the University of Adelaide, Australia, have put together a remarkable website on R. A. Fisher: Fisher. The Australian statistician John Henry Bennet has been a prime mover. Fisher spent his last years at the University of Adelaide where he was much appreciated.

Those of us privileged to know Fisher, admired his critical and capacious mind, his elegant and concise writing style. To say that he was a distinctive presence would be a gross understatement. His "Dispersion on a Sphere" formulated in 1951 (altho not published until 1953) placed paleomagnetic and studies of the prehistory of the geomagnetic field on a quantitative basis. I wouldn't be suprised if his paper is the most widely cited reference in our discipline.

Fisher regarded himself as a biologist: statistics the tool, biology the objective. But during his lifetime, his reputation as a statistician probably out-shone his reputation as a biologist or geneticist. In the last few decades, his work has come to be increasingly admired by biologist, particulary evolutionary biologists and Darwinists generally. As I understand from Henry Bennet, this was one of the motivations for creating this excellent website about one of the great luminaries of the last century. Ted Irving 21 Jan 2005
There are a lot of lessons about ideas and invention. ME 101 at stanford pummeled us with this: Ideas are not rare. They are everywhere. The ability to communicate an idea and get it implemented is rare. The idea may be great, but if people are judging your presentation they will never get to the content.

Some random ideas I had while standing at the Kappabridge for a few hours (the mind will wander):
  • I need to make a sample storage container for cubes. It should hold them all in a nice foam padding and in order. Ziplocks suck for this.
  • What open source software (or other types of) projects do you try to emulate or look up to? Especially in the documenation department. My best attempt to date (density) is really not that good in terms of docs. Yes, there are man pages and web doxygen, but they are WEAK. Who does the best job? I frequetly learn my emulation. I useually snag a tar ball of some GNU project like grep or one of the SIM projects to look inside for how they did things like command line args, man pages, or doxygen.
Workshop Announcement - Antarctic Geological Drilling (ANDRILL)

For further information, please go to: andrill.org
The ANDRILL - United States Steering Committee invites applications for
participation in the United States ANDRILL Workshop to be held at the
Embassy Suites Hotel, Denver International Airport, 1-2 April 2005.

ANDRILL is an international research program designed to investigate Antarctica's role in Cenozoic global environmental change through stratigraphic drilling and numerical modeling. ANDRILL will obtain stratigraphic records of key intervals and events in Earth's history from locations proximal to the Antarctic cryosphere, using sea-ice and glacial-ice as drilling platforms. ...
Jeff Ota says that he teaches his Santa Clara classes with a tablet computer. Jeff, how about some details of how it works. Good and bad?

Played with pysqlite from fink yesterday. Eh? Sounds like it is kind of like connection pooling/batching type ideas. Batch up a bunch of changes and do them all at once. More complicated than I am interested in. Maybe someday soon.

Myche mentioned an alternative to OpenOffice on the Mac since they nixxed the Aqua port: NeoOffice/J. It is Java based.

Warren's link to Mac OSX software by Marc Liyanage in Switzerland entropy

UCSD's ask a librarian live. Anyone tried this yet?

Arg. I have not followed up with the Harvard CFA about ADSABS RSS feeds.

That was a quick dump of interesting things sitting in my inbox.

Posted by Kurt | Permalink

01.26.2005 06:50

McAfee Visual Firewall patent

The patent system is so broken it hurts. I just saw that McAffee was granted this patent for a firewall that tracks the incoming packets and geographically shows where they are coming from. There have been algorithms that attempt to show the physical location of IP addresses for MANY years. One example is NeoGeo/CAIDE This is just combining that with the age old etherman that only ran on SGI's that used coff libraries (IRIX up to 5.3).

There was also interman or something like that. These programs made a graphical representation of the location of all traffic seen by a host. During the Marsokhod Desert 1996 field deployment to Arizona, I had an old Indigo R3K machine sitting down in front of the projection screen running etherman. I could then see traffic to all the science team computers, our file server, and traffic out from the desert dropping in from our private satellite link to the field. I could then see Marsokhod sending traffic back in along with Dan and Eric on the SGI Indy logging back into the Intelligent Mechanisms Group.

Ethereal claims to be an etherman clone, but I did not see any of the standard nodes with links graphics that were the mainstain of etherman.

I found through google the name of the family of programs (they were binary only): etherman, interman, packetman, loadman. These were all a part of the Austrialian netman package.
         (C) Copyright 1993
    Curtin University of Technology
       Perth Western Australia

Found solaris binaries here: netman

Code Complete Rev 2 by McConnell might be worth a look the next time I am in borders.

plans is another web calendaring system, this time in perl. I still think phpcalendar would be better for the Driscoll group.

MPCA - Principle component analysis on 100 MB data sets. GPL'ed.

Ugh. Just heard President George W. tell a reporter that the reporter was acting like a senior citizen. That is not a stateman like act.

Posted by Kurt | Permalink

01.25.2005 21:18

Random links

How to Record a Podcast. This covers how to snag audio from phone calls to do interviews using skype for your phone calls.

Bjork's new video with a cat.

Maybe elog is what I want over a blog. It is an electronic log book with a web publishing side.

SciAm article on Seeking Better Web Searches.

flatgallery produces photo galleries by just dropping images into the directory on the web server. Uses php and no database. Might be nice.

VCG Metro computes distance between meshes. Might be good for mesh registration/alignment projects.

Posted by Kurt | Permalink

01.25.2005 16:03

sqlite and sql rock!

So now that I am working up a script to inject my data into a little sqlite database, things are getting exciting.
   open `sqlite bpsio04.db 'select sheetURL from \
   CoreLoc where corenum=1;'`
The database finds the URL for time image and Mac OSX pops up Safari with the right URL:

I believe this is where Alex gets to say "Why did you not listen to me about how useful/easy/cool SQL is back in 2000?"

Posted by Kurt | Permalink

01.25.2005 11:26


My first try with sqlite is showing success. I am using sqlite 2.8.5 from fink. The goal is to create tables for the BPSIO / Gaviota / Goleta / Santa Barbara cruise from August 2004. If I can get used to this whole sql thing (better than with MER), I think this may be a better way to work with core data. Here are the first few commands that I ran from the shell to get started:

  1. sqlite gaviota.db " create table CoreLoc (id INTEGER PRIMARY KEY, cruise VARCHAR(20), corenum INTEGER, coretype varchar(1), datecollected TIMESTAMP, lat REAL, lon REAL, depth REAL, sheetURL TEXT);"
  2. sqlite gaviota.db " insert into CoreLoc (cruise, corenum, coretype, datecollected, lat, lon, depth, sheetURL) values (' BPSIO04' ,1,' g' ,' 2004-08-04 21:09:15' , 34.36116666, 120.108, 480, ' http://schwehr.org/Gaviota/bpsio-Aug04/bpsio-log/images/bpsio-log-35.jpg' );"
  3. sqlite gaviota.db " select * from CoreLoc"
Random links:

bib2html for BibTeX

1-159268954-6 - Magnetic dust on mars

Posted by Kurt | Permalink

01.25.2005 08:30

Mac Video

I bet the whole net has slowed down today due to the 1984 mac video. I snagged a copy from this mirror. The original release site.

One feature that fink commander really needs is a "Pause Build Function." I have been trying to build kdebase3-ssl for about 4 days now, and everytime I have had to kill the build after a couple hours because I'm mobile somewhere and need to save battery power. Today, while I work, I am going to give it another try starting at 7:46 AM.

Saw an announcement about ANT and was confused thinking that it was a java build tool with RSS. This is something different: antnottv - video blogging. Hmmm... I already watch too much TV. More ways to clog my DSL link.

Update: kdsbase-ssl 3.3.2-21 finished building at 10:10 AM. 2 hours and 20 minutes.

Freshmeat just changed their RSS feed. Yikes! Now it comes with the screen shot and more description. Great improvement.

Posted by Kurt | Permalink

01.24.2005 15:05

UCSD Ship costs

UCSD Ship rates

  Ship Rates   Technician Rates  
  (Provisional)   Fixed Projected
  2004 2005 2004 2005
R/V Roger Revelle $19,368 $20,336 $3,000 $3,150
R/V Melville 17,976 18,874 2,814 2,955
R/V New Horizon 11,757 12,345 1,239 1,301
R/V Robert G. Sproul 7,438 7,810 929 975

Do not forget to add 13% overhead.

Posted by Kurt | Permalink

01.24.2005 12:05


Here is a program I would like to see. I call it "ScienceBlog." This would be a weblog with some different features. MER for example, each entry would be a raw, color, or mosaic image. Then for comments, people could define a region of discussion in the image. They could draw a box around an image and say, "Check the layering in this rock" or "This rock is weird" or "Is this a dust storm in the distance". Then you could have scientific discussions right on the image. You could do the same for graphs like the APXS and Mossbauer graphs. There it would be, "This peak indicates Jarosite, but we did not see it pre-RAT". The site could have groups such there are people who can see certain posts and others who are not able to see them. You could then have a time release too. For example, there could be a mode where a comment is automatically released in 6 months and another mode where the author gets bugged in 6 months about release and then if they do not answer in 1 month, it goes public. Would this help capture the scientific process? Scientists could blank out the public or groups to stay focused.

Posted by Kurt | Permalink

01.24.2005 07:52

X-Core lives!

There is a title. X-Core has been re-animated. This is a great lesson to myself. I wrote this code as fast as I could to win the first SIO Visualization Competition (which I did win), so I cut as many corners as possible. When I code fast, I do what I am used to, so I added some asserts to arguments passed in and made as many things const as possible. Where I was lacking what in return code checks and DOCUMENTATION. I did not know doxygen at that point so I had not really knocked the need for documentation completely into my brain. After density, I know what is required for minimal documentation. Return code checking is easy to let slip, especially when no easy to get at documentation is available. I still have never written great documentation to date. So the lesson today is:
  • Use doxygen and document all functions, methods, parameters and especially return codes.
  • Take that doxygen output and make sure that you check all return codes!
But the upside is that I ran processCore using an input file generated with g3data and some hand tweaking. It worked! I got some surprises. I forgot that it generates a meter stick image file along side. Very handy, but the mind slips. processCore makes each section an image and then a total core image. Woo hoo! Now I need to use g3data on the rest of the sections. And this time, i will be checking the description files into cvs. Right now, I am using the "XCore-v1.0" format. Why did I put a "v" in there? Here is the extension I want to use:
  xpd - Xcore Photo Description
Now that I think about it, making a color image will be easy. All I have to is make separate grayscale images for each channel of red, green, and blue.

Ever wanted a keyboard shortcut to zoom in and out of your powerpoint like in photoshop and ill? Check out this hint.

After using Mac OSX for almost 4 years, I would not want a linux box as my desktop anymore, BUT they make fantastic servers of all kinds (including graphics rendering, file serving, compute farms, and embedded). Looking at the nForce Professional makes me drool. My ultimate work setup would be a 15 inch Mac laptop, highest end Mac desktop, and a dual opteron running 64bit only linux (it is time to trash 32 bit with linux). Anandtech review

Interesting links:

cfg - OSSP cfg for C/C++ style config files. Maybe good to have in addition to gengetopt?

tsep is yet another website search engine. May require mySQL?

cw is a text colorizer for command line programs. An alternative to the that I maintain the info file for.

01.23.2005 19:13

Medical Geology

Medical Geology is a new one to me. 4.4K hits in google.

Been working on my old xcore stuff. I have processCore building on the Mac, which was pretty easy, but it will take a bit to make the input files with g3data. The C++ code inside processCore is not good. Hardly and documentation and some not so valid code. Whoops. I was trying lots of new things back in 2002.

Posted by Kurt | Permalink

01.23.2005 15:11

msnbot please go away

I just added a robots.txt file to my website. The msnbot is the highest traffic load and I have no interest in supporting Microsoft. Here is my robots.txt file:
  User-agent: msnbot
  Disallow: /
I did a search on www.msn.com for msnbot to get info about the bots name and robot.txt preferences.

Here are the links that I used to get some info on the robots.txt: Last but not least, some info on the msnbot:
Non-authoritative answer:       name = msnbot.msn.com.

Authoritative answers can be found from: 46.207.in-addr.arpa nameserver = ns3.msft.net. 46.207.in-addr.arpa nameserver = ns4.msft.net. 46.207.in-addr.arpa nameserver = ns5.msft.net. 46.207.in-addr.arpa nameserver = ns1.msft.net. 46.207.in-addr.arpa nameserver = ns2.msft.net. ns1.msft.net internet address = ns2.msft.net internet address = ns3.msft.net internet address = ns4.msft.net internet address = ns5.msft.net internet address =
msnbot has been 17% of the hits this month, 17% of the files, and 327MB! Thanks microsoft. Would you rather I just sent you a DVD of the whole site?

Posted by Kurt | Permalink

01.23.2005 10:15

NSF Visualization Contest and Exploring Yellowstone

Entries for NSF Visualization contest for this year are due May 31st.

How do I tell Safari to hang onto the Yahoo https cert? I sign onto yahoo mail using ssl and am starting to get annoyed by the popup to accept the cert as ok. The same goes for ssl pop to ucsd in Mail.app. I should be able to save the cert and have neon lights come on if it changes! That would be security 101.

A Yellowstone Adventure - Deep in the Norris Geyser Basin

You may want this map of Norris Geyser Basin up while reading on. Unfortunately, most of what is below is off the map to the North.

Driving in the car this morning, I went by some place with this smell that triggered a very vivid memory that I have from working in Yellowstone National Park. That strong sulfur smell. I think it was May of 1995, and I went with Deena and Don L. to help out working on hot spring research. Deena was focusing on silica sinter growth. We met up with one of the parks research rangers and headed up to the Norris geyser basin for an in depth tour. At the top of the stairs before you go down, we went right along the side and straight down past the railings to the bottom of the basin. We entered into another planet from my perspective. To kneel down at the edge and see all the crazy structures of the very first pool was like landing on another planet. I believe this was the Porcelain Basin. We then followed the ranger around to the right side of the basin to the north eastern edge. If memory serves me, this is close to where the original hotel was. We then walked west onto this dome area and mist and snow started to come in. We sat down on this dome which was a nice 80°F and listened and felt the thumping of the geyser basin enveloped us. The air was freezing and we could not see more than 50 feet, but we were warm and safe sitting on this spot. This is where one of the key reasons for this guided tour became apparent. The ranger was instilling all of his knowledge about safe travel through geyser fields. The basic rules are that you should never enter a geyser field if you are not a professional, and even then you must be extremely cautious. There are hidden traps and what was safe yesterday is not necessarily safe today.

After sitting on this other planet for about 20 minutes, the snow flurry subsided turning into a light drizzle and we headed into the geyser field. I can only imaging what the tourists thought of us all in our official outfits out there in this geyser basin. This place is so huge that we surely looked like ants. We traversed through many places I do not remember out to the boiling sulfur pool. This pool is covered by black spheres of sulfur so completely that the surface looked to be solid ground. Yet, slide the beads aside and there was sub-critical water, just waiting to burn you. We heard the story of when they put a sample bucket down into the brown water quite a ways and returned with near molten sulfur. Think back, I wonder what the temperature is that sulfur becomes like the taffy that was described.

We then crossed some braided streams out to Realgar Spring where arsenic minerals are sitting right there on the surface with that distinctive realgar color.

This got us to not even noon that day. The number of adventures in just one day was phenomenal. That whole trip was, for me, just like be an astronaut landing on the surface of another planet.

And now, back to working on mud of the Santa Barbara Basin and my thesis.

Posted by Kurt | Permalink

01.22.2005 19:48

THX and JAPANESE cooking

Just got a book called JAPANESE cooking by Emi Kazuko and the movie THX 1138. The cook book looks nice with some interesting ideas like using spinach instead of seaweed for wrapping sushi. It has excellent color photos and was on sale at Borders for $6. I can't wait to sit down and watch THX 1138 again, plus it has his original student film that I have never seen before.

I looked in the Google Hacks and Google Pocket Guide. I still have not found the conversions trick that I used before and now can not remember. I did write down a list of things.
  -"foo"   (exclude a term)
  "foo * bar"
  site:schwehr.org inventor
But then I looked up the conversions and found this conversion of convert 10 inches to centimeters.

Search google for geology news

I have been thinking of quick writing up all my ideas on what are the ways you can do model construction for images. I may site down one morning and pound that out. This topic needs to leave my head as I don't have time to think about this stuff!

The Free Software Magazine. Will this be any good?

RMS has a blog. This should be interesting. And I quote: "About two months ago I heard about a project at the MIT Media Lab to develop a web browser for parrots. The smarter kinds of parrots can get very bored if they are home for hours with nobody to play with, so the idea makes sense." Need I say more? I really should become a FSF member.

I need to summarize the traffic about installing fink across large number of machines. Definitely some interesting stuff, but nothing yet that I would want to hang my hat on. I got one comment about deploying MS Windows instead. Err... no thanks. We already had multiple major failures outside of flight operations from viruses romping through the JPL professionally managed Windows boxes. That was not fun.

datamaster is another Dia UML sketch to code tool.

Posted by Kurt | Permalink

01.22.2005 14:27

Misc stuff

Delcom has a little USB development kit for $50. Now I need an excuse to get and play with one of these. They specifically mention Mac OSX, which is great. But why not Linux? Anyone have any good sites describing little projects with these? How about a wheel encoder for the winches on the ship? That would be a big help for those that don't have one! I never did use that little HC11 that I bought like 10 years ago. Maybe the Linux support just comes with the kernel?

I just saw this book while browsing last weekend: Mars: The NASA Mission Reports Vol 2. Funny thing is that I saw a very large number of pictures that I created. Lots of MER pictures. There were stills of Viz and RSVP that I made. In addition, there were tons of stuff created by Aurelio, Shigeru, Steve, and Zareh. It looks like they went through the photojournal for most of their images.

Someday, I would like to setup cvsweb for density and xcore packages. But that will not be until after I finish my thesis.

The Open 3D Visualization Toolkit is here: ltc.smm.org/visualize/toolkit. The site has been slashdotted so I haven't been able to look at it yet.

itk - NLM Insight; Segmentation and Registration Toolkit to support the Visual Human project.

openqvis - Open source volume rendering. Has some interesting datasets that would be easy to import and it is based on Coin! But not Voleon?

gcx - CCD photometry tools.

As for everyone who is upset about the Hubble, I understand, but my opinion is that we should spend that money for another space telescope and launch it with something other than the shuttle. I know there there is another space telescope set for 2011, but why not make it two? These projects are so valuable. Why waste time, energy, and possibly lives when we could build something for probably around the price of repair? Ok, so politics makes it easier to repair than get something new which has a high likelihood of being canceled by congress.

mathomatic - This one is already in fink. Symbolic math program. Wow. That built and installed in under a minute. Funky green screen. Now what can I do with this?

On my todo list is the type exactly this CVS tutorial for Jeromy B.

New fink version in cvs: 0.23.5 - "Fink: The Mojito of computing."

Posted by Kurt | Permalink

01.21.2005 19:30

Tim Burton's Corpse Bride

Tim Burton's Corpse Bride trailer. Can't wait!

Posted by Kurt | Permalink

01.21.2005 10:39

TG with Jerry Bode

From all-at-sio. This evening:
Jerry Bode is retiring from the Deep Sea Drilling Project and has
graciously donated the beer for tonight's TG.  So, come on down and
enjoy the yummy muchies, good beer, as well as the beautiful weather.

2P159388343EFFA2HPP2514R1M1 - Check out the textures in this pancam image from Spirit. It is hard to tell what I am looing at, but that might be some layering. Spacecraft 2 is spirit, right?
  MER A  ==  2  ==  Spirit
  MER B  ==  1  ==  Opportunity

I have been reading copies of the magazine Vision Systems Design while on the bus. (www.vision-systems.com) There is interesting stuff in here, but it mostly reads like a big sales brochure. I had never heard of Camera Link before. In their May 2004 issue, they tried to compare camera link with GigEther (why?), FireWire (only 400, not 800. Why?), and some other things. They got a lots of their bits versus bytes all messed up. Now I am reading about calibrating noise in CMOS imagers using a highly controllable sphere of light that you put over the camera. I always think that calling a light by a "Color Tempurature" is so strange.

I really get annoyed by the AOL messages about being connected in multiple locations. It usually happens with wireless hickups. Plus the messages do not tell me from WHERE?!?!

Posted by Kurt | Permalink

01.21.2005 08:18


Welcome to Friday morning.

Yesterday a few interesting things came up. First, Catherine showed me a teaching tool that seems to have huge promise for larger classes. Each student gets a remote that is keyed to their student ID. Then in class there are receivers. You can ask questions in class and get immediate responses in terms of a bar graph. This is like the Pittsburgh Science Center for Nomad, but you get to know who answers what. At the PSC, each chair could vote on of three ways to control the Nomad rover as it went through the Atacama desert. You could use this kind of tool to get students to collaborate. Answer first on your own, then discuss, and see if as a group you can improve your answer. This is a device with a lot of potential. What exactly was this thing called? Google would not tell me.

Yesterday, I took a look at sugarbush, our Linux seismic server, that runs Mandrake 8.1. Now I can not remember for sure when 8.1 was the current version, but it must have been 2001. So this machine needs an update no matter what, but at the moment it would boot under its own power. It comes up into the boot process and makes it two past the interactive selection and hangs as it says it is starting DevFS. I tried to do fsck.xfs, since it uses SGI's XFS for the root partitions. However, I later discovered from the man page that fsck.xfs does nothing and returns success (EXIT_SUCCESS in C). Thanks. Then I looked up the xfs_repair man page and here is what I had to do while booted in knoppix:
  sudo -s
  umount /mnt/sdb1
  mount -w /mnt/sdb1
  umount /mnt/sdb1
  xfs_repair -v /mnt/sdb1
Guess what... Still did not work. I tried removing devfs from the lilo boot line, but got nervous about this since knoppix does all sorts of renaming of what is sda and sdb, so maybe I would push the system over the edge. It was really strange looking at the logs. syslog stops in November and the messages log ends on Jan 10th. Time for a total wipe of the / and /usr partitions and an install of the latest Mandrake version.

I am currently debating putting my AMS K15 and s[7] matrix data into sqlite for working on the rest of my thesis. There is a nice python module for working with sqlite. This might make it much easier to select portions of the data for doing things like a windowed bootstrap that I want to try out. The idea is that I would set a window of N samples, where N > 6. I would move this window along the core and bootstrap a group. Then slide the window my one sample and bootstrap again. The idea is to bootstrap the smallest region of the core so that samples are as genetically related as possible. For the samples in BPSIO Core 1 Sec 2, there is a sample every 3 cm. That would give me a minimum window of 18cm. This method would suffer the same problems as U-channeling does, but I might be able to find coherent deformation.

For the todo list: look up and explain what exactly a statistical null hypothesis is.

Here are some interesting links on industrial control and instrument data acquisition from Linux Weekly News. ProcessViewBrowser looks like it has had much more development. Exciting projects! Qt is still on my list of things to learn more about, but which will not happen until some time way in the future. By the way, SCADA stands for Supervisory Control And Data Acquisition.

A geologic article on yahoo: Volcanic Warming Eyed in 'Great Dying'. They talk about massive volcanic flows in Siberia. The technical term would be Large Igneous Province (LIP), right? Like the Ontong Java plateau or the Decan Traps in India.

I have a little story from doing my research yesterday. I was working away measuring paleomag cubes on the Kappabridge when I realized that a contract lens was bugging me. After about an hour of this, I was getting kind of annoyed. It is about time to go for a new pair, so I figured I would just throw out that one and work some more. It would only be a couple hours before I would get home. I was mostly okay, but it was a not the most fun looking at the computer screen. I finally had to just go home. That was when the trouble started. I decided to read a computer vision magazine I had with me while I was on the bus. I wanted to bury myself away from the wet cough lady behind me who was insisting on clinging to the back of my seat for dear life. Bus drivers on the 34 route are not all alike. Some are good drivers, but other treat the breaks like a binary switch so everyone is banging back and forth for the whole ride. Yesterday was a binary switch kind of guy. Between the driver, the reading with one contact, and the nasty bus area, I was pretty nauseous by the time I got off that bus.

I added MulleSite and gcam to my MacOSX 10.3 install notes doc yesterday. I still have not got around to working on turning my two talks into a movie.

Methane rain on Titan [bbc.com]

From the "Tao of Mac" is a list of essential osx software

cvstraq uses sqlite.

SimplybibTeX - PHP BibTeX web program with RSS and more. Any good?

Posted by Kurt | Permalink

01.20.2005 23:56

Core 1 AMS

I finished measuring the AMS for the Santa Barbara Core number one for the working half, section two. My first impression is wow. The samples almost all have distinct maximum and intermediate eigen vectors.

pywrat - Tools for using ssh and scp from in python scripts.

This evening I went to a secret showing at AMC La Jolla. Ridley Scott was there along with the head of Fox studios. That was cool.

Wow. I am actually getting feed back on my blog. This comes from Myche:
I thought I'd relay that I've been using blojsom
(http://blojsom.sf.net/).  Though this app has a rather odd name, it
is implemented in a way that I can both understand and appreciate.
Namely, it uses servlets and page templates to render your blog
entries (which are still kept as flat files).  It doesn't have to
"precompile" the pages like nanoblogger.  Dunno yet what implications
this has versus using a database to store the entries, but I like it
quite a bit.

Blojsom also has the obligatory plugin interface. I use a few of these (like weather) and had an interesting time learning how to install and configure them. Also, I'm using the "kubrick" theme, which is not the default.

Posted by Kurt | Permalink

01.20.2005 07:22

Land and Mud slides

An article about the CDC guide to land and mud slides.

VisIt - Scientific Visualization software caught my eye in freshmeat today.

Posted by Kurt | Permalink

01.19.2005 16:34

Visited NASA Ames

This morning I went over to the Intelligent Robotics Group (IRG) at NASA Ames and talked to bunch of folks about a new type of autonomous vehicle. It was cool to see how similar our thoughts are even though we both started separately without knowing the others intentions.

Before that I got a demo of Anne and Randy of their GameBoy based bot-ball device. Holly-smokes! This thing is awesome!

I now have to cleanup my disk from SPIE, Electronic Imaging, Visualization and Data Analysis, or what ever it was called. My disk is a mess at 97% full after two days of rendering frames. I will try to get a movie file with overlaid with audio. I was able to up the gain in Audacity and write out audio that does not force me to crank the volume.

I started writing my first article for my blog call "About." Talk about surprising. It is an attempt to explain why I am blogging, what I intend to use it for and what will and will not be included. For example, I do not have a lot of the actual analysis from my thesis in here. That is for my thesis.

Talking to Anne and Randy crystallized my feeling that at some point I need to move past Nannoblogger to a more powerful tool. The biggest gripe right now is the time it takes to update the html from a change to the data files. This makes the write, update, test cycle a major pain when you have an html tag that is wonked.

I just gave iMovie a try with exported frames from PowerPoint. Ugh. It chopped the side edges off the frame. Then when I resized the frame it seems the video freaks out about 1/3 of the way through that frame. What is the official NTSC frame size? I always forget. Is it 640x480?

One little time about the new terminal in San Jose Airport is that if you go down to the higher numbered gates, it is quieter and there outlets on the columns. Not the most comfortable place to sit, but better to have power, which is not available anywhere at gates A1-8.

What is the best way to edit php code with emacs? Dropping into html mode does not seem all the helpful.

I hear that Navtech is the engine for all the mapping web applications. Who are these guys?

Posted by Kurt | Permalink

01.19.2005 07:47


Couple random links and notes:

Read an article in the newspaper talking about how OYO Geospace (Houston, TX) is trying to get their technology used in tsunami warning systems. What really makes up a tsunami warning system?

Marie - a LGPL'ed robotic design tool.

VIPS - IPS is an image processing suite designed for extremely large images and colorimetry.

What is VTK CISG Registration toolkit

Posted by Kurt | Permalink

01.18.2005 15:25

Talks are done

That is it for my talks. I got one question: Are you using this kind of thing to manage AUVs in real time. Yes, but I have not had a chance yet. I think the audio recorded okay, but I have not tried to listen to it yet.

How copyright could be killing culture

A quote from this BBC news article:
Ms Rice retorted: "I have to say that I have never, ever, lost respect
for the truth in the service of anything."

Posted by Kurt | Permalink

01.18.2005 12:51

Tues midday at IS&T/SPIE

BBC on the Shackleton hut to be resurrected.

Got a chance to talk to Justin Maki at lunch time. He is the only person I have talked to here that I knew from before.

Titan RAW frames from Huygens released by the ESA.

MER Cad images - I am not sure where this came from and if is an okay release, but it is neat to take a look at.

All of my twisted python work accept cvstoys and build bot is now in the fink cvs tree and I am not the maintainer. Woohoo!

Posted by Kurt | Permalink

01.18.2005 10:08

Justin Maki's talk

Justin Maki's "20 Cameras on Mars" talk was excellent. This was definitely the best overview of the imagers that I have seen to date. I recorded Justin's talk for myself, so I will try to write some highlights at a later date. Also, he presented 30 polarized stereo pairs. I was worried that since the projectors were way at the pack of the room onto two 15foot tall screens that getting the horizontal and vertical match might be a problem, but just the opposite. These were the best static stereo images I have seen to day. Now keep in mind that I have not been back to JPL in 9 months. The synthetic stereo pairs from the microscopic imager were rockin!

SPIE could do with some better support. It is almost as bad as the Chicago airport. Power outlets are few and full of people using them. Wireless works and is free, but it takes about 2 minutes for an ssh connection to be established. Once I got a connection, I was getting 359KB/s, but yesterday, I thought they were blocking ssh for the first couple hours I was here. I can not seem to get pop-ssl to get out to my mail, so it is back to pine.

I think I am the only person here with a Mac laptop. Crazy! I have seen two or three with linux desktops, but this place is Windows XP crazy. I am a fish out of water.

Looking at Webalizer, I had a huge usage spike! I was averaging about 2800 hits a day, then yesterday it hit 5200. What happened? this can not be from my SPIE talk. There were only 30 people there, and I do not think that it was my, since my laptop was barely on the net. Maybe it was the msnbot which is now up to 8499 hits (15.35%).

Fire has plugins? Where do I get them and what do they do? Does Fire allow logging of sessions? I'd like to log some and not others. How?

Morning links:

cms made simple? Seems to use static files too?

Durability of Usability Guidelines

Posted by Kurt | Permalink

01.18.2005 08:30

Tuesday at SPIE Electronic Imaging

8 AM Tuesday morning - I am sitting in the big ball room before Justin Maki's talk. Should be a great one. They have polarized stereo images on the screens. Last night, my talk went well. The room the VDA section is in has about 65 chairs and it was half full. I hit record on my iPod with iTalk as the speaker before me was doing his conclusions. The one drawback of the iTalk is that it obscures the lock button. I was worried that the pause button would get hit during my talk. But checking when I got hope last night, it sounds like I have a good audio file. I need to boost the level, but the sound quality is good. Is there a way to change the record gains a little on the iTalk?

I now have used the dlo Jam Jacket 20 (that thing that wraps my ipod 20GB) for a week now. Overall I like it. The iTalk fits and I never feel like the iPod is going to slip out of my hands anymore. There are a few drawbacks. First off, this this is the great lint attractor. I think that my dress pant pockets must be devoid of all pocket link because it is plastered to my ipod case. My other worry is that it does no cover the back top edge of the ipod, so if I drop the ipod just so, it will hit metal on the ground instead of the rubber case.

While I am waiting for 15 minutes, I will see how much of yesterday I can capture. The most exciting talk in my view was the invited paper by C. Chen. It was a little frightening when we all saw his powerpoint had 76 slides. Yikes. He ended up blasting through may of his slides. But more to the point he talked about looking at citation relationships between research articles. The basic unit that he works with is that if two journal articles site an older paper, those two new articles and the older article are strongly linked. The main drawback of this is that technique is that the new cutting edge papers that are the hot topic get left out until the citations get farther on in time. There is an advancing front that this system tracks. He then talked about pruning algorithms to prune all but the most important links. He then looked for seminal papers: those that link clouds or flurries of research. He showed the prion research and the mad cow disease research as two clouds that linked through an early key paper that introduces the link between the two. This could be an excellent tool for those new to a field to discover some older key research on a topic that they need to review first. This method tends to highlight the summary or survey papers since they tend to be strong linkers.

I quickly jumped on the idea of McGreevy's word relationship work in the 90's where he looked at what words relate to other key words in a paper. McGreevy's work could be used to test links. If the word associations between two linked papers are totally dissimilar, then maybe they are not related papers. Check out McGreevy's work here:

McGreevy's work

Posted by Kurt | Permalink

01.17.2005 11:24

Monday's talks

10:15 AM I'm not in the San Carlos room for the Visualization and Data Analysis section. This section is right by the stairs in the Marriot. This is a very small room about the size of Vaughn Hall 300.

Things of interest for Monday:

2:50 Recovery of a missing color component in stereo images (or helping NASA find little green Martians)


8:30 20 Cameras on Mars: The Mars Exploration Rover Imaging System
11:00 Low-cost photogrammetry for real-time 3D measurements
11:20 Photogrammetry for geological applications: automatic retrieval of discontinuity orientation in rock slopes
2:50 Three-dimensional scene reconstruction using multiview images and depth camera
3:10 Smoothing region boundaries in variable depth mapping for real time stereoscopic images
5:30 Videometrics Posters
5:30 Color Imaging X: Processing, Hardcopy, & Apps - Posters

WARNING: unstructured dump follows...
Making python programs into apps: py2app

The Marriot hotel's wifi settings sent my laptop into spasams.

  • Self organizing maps - projecting higher dimensions to 2D:

    With neural networks, what is "Extremal Energies?"

    Unwrapping a tessellated sphere. Efficient representation, but some nodes are duplicated. This saves neighborhood pointers. Uses O(n) space, rather than O(n^2)

    By mapping geo-political data onto a sphere this gets rid of the political issue of everyone wanting to be in the center of the map.
  • Haptics of Fluid Dynamics - Anders Ynnerman - Sweden

    CFD data from SAAB of Unmanned Aerial Vehicles - Case study on how to deploy volume haptics.

    Visual Rending - surface rendering, volume rendering (pressure field, curl for opacity, stream ribbons)

    Used the ReachIn Desktop Display

    Proxies for haptics - passive representations - keep pen stable. Apply spring equation as you go through the surface. Extending this to volumes. Dragging the proxy with you as you force the pen along. Jumping the proxies within the streamlines. Feels like you are trapped in a cylinder.

    For volume rendering, how can you know that you have fully explored the volume data set? How do you know that you have the right transfer function and have not missed a tumor in the body?

  • Illustrative City Models

    Photorealistic obscures data changes and thematic information. Also leads to high expectations from the user. Hard to keep up changes.

    This guy would get in trouble with the dept of homeland security if he was using US cities.

    This is the kind of data that Peter Selkin got out of UCSD with special permission for his GIS/surveying class.

    Shading by face with small number of different hues using N-tone shading along with depth cueing by decreasing color saturation with depth. Then add shadow casting. Add facade texturing by say using light and dark windows on buildings. Then add variable edges to buildings and window LOD.

Lunch Break

Invited talk by C. Chen - Measuring the movement of a research paradigm.

Unit of Analysis = a knowledge domain

Pioneers: kuhn (1962), Englebart, science mapping: h small

Posted by Kurt | Permalink

01.17.2005 07:34

Off to SPIE

I am heading in to SPIE. My talk first talk will be this afternoon at what looks like 5:00PM. This will be the SIO Mars talk. Tomorrow afternoon at 2:20 will be the visualization architecture talk.

9:45AM I am in the conference hall. This place is very quiet. Off to the San Carlos room. If you are lost, the conference is hidden upstairs.

Posted by Kurt | Permalink

01.16.2005 21:14


OpendTect Student Competition and free licenses for this OpenInventor based seismic tool. This is OpenSource, but what license?

Posted by Kurt | Permalink

01.16.2005 08:20

A patent on panoramas?

The comments are starting to flow in about fink deployment and Mac OSX for Mars operations. Keep them coming and I will summarize later this week.

There was an assertion that pro graphics cards are basically the same as the high end gaming cards. I have intentionally ignored the details of graphics cards for the last 4 or 5 years since that knowledge goes out of date year. But, with statements like that, what really is the difference between the $500 cards and the $2-5K cards? Is the texture ram different? Is the quality of graphics different? Quality of rendering used to be massively different. There was less effort spent making sure the graphics are 100% correct in the consumer cards when I looked a few years back. Can the cheaper cards do LCD shutter glasses in a window? Can they blend two neighboring screens for Panorams?

What is this SIFT algorithm for making control point panoramas? Is this a valid patent with JPL's Vicar software being around for a while? autopano-sift which needs hugin

sMArTH - My hope for MathML never ends.

In addition to version tracker:

xwinmanimp is like xtermcontrol but even more low level. Control anything. Note that I have not tried this one yet.

Great news! Looks like Daniel Henninger is going to take over and submit all of my twisted related fink info files. Thanks Daniel!

Posted by Kurt | Permalink

01.15.2005 20:23

Phoenix - Mars Lander 2007

Check out the Phoenix - Mars Lander 2007 web site. The PI is Peter Smith at the University of Arizona, Tucson. There are a lot of familiar faces on the web page from previous Mars missions.

Posted by Kurt | Permalink

01.15.2005 18:59

Flying to SJO

Dia is an open source diagram package. Maybe with uml2php5 I might actually use the UML training from 1998. Not likely, but it is still a possibility. The Dia homepage. There is also umbrello in fink.

How does wikindx stack up to Endnote? Arg. Another application that uses mysql. Would be nice if it were self contained.

While waiting to board the plain to SJO, I opened the SPIE Electronic Imaging list of talks and the first thing I see is Justin Maki's picture. Excellent. That will go well with my two talks. The first one on Monday is an overview type talk where I will cover general concepts that are important to the scientific end users visualization tools. There will be movies made from both Fledermaus and OpenInventor/density. The imagery will focus on the Spirit rover and Gusev crater.

My second talk is more a systems overview. This will be a look at the flexibility require for working on visualization system for a wide variety of crafter from Mars surface rovers (what if Hugyens had been a rover with and RTG?) back to peopled ships on Earth and underwater vehicles. The key is being able to throw out or tweak any part that is necessary. I will start with the Viz architecture and show how it lets you layer on top of plain old ivview. People watching Viz 1.0 did not always realize that they were seeing more than ivview until a 3D model appears out of nowhere :)
I probably have to cut the second talk down substantially. Well see how the rehearsals go! 20 slides in 15 minutes is not going to happen! So the disclamer is as aways, all talks subject to change until I actually give them!

Posted by Kurt | Permalink

01.15.2005 11:13


Myche just let me know that the Titan images are up on the Titan PhotoJournal Page It looks like the folks at the ESA did not project the images MIPL style, so their images so the mosaics are a bit strange. These are amazing images!

Sounds from Titan. Worth a listen. The radar is very amuzing.

Posted by Kurt | Permalink

01.15.2005 07:35

Mac OSX for Mars lander missions

It looks like my previous entry about setting up fink for a large number of machines got snagged by the spam filters. Before I write to the fink lists again, I want to briefly add to those questions with a list of the things that I believe it will take to make Mac OSX the dominate OS for the next Mars lander mission. For MER, Linux (RedHat 7.2) was the OS for about 80% of the machines inside the flight firewall. There rest were almost all SunOS 5.9 (Solaris 9) boxes. The primary file server was a two refrigerator sized Sun (I'll save the performance comments for another day - summary: "Not Good"). I also saw two full sized rack SGIs. No one I talked to knew what they were for. I believe they were Origins running IRIX. They lived not in the server room, but in the FAX room. There were a number of G5 OSX 10.3 machines, but they were all outside the firewall, not as banded as Windows and the foreign nationals, but not in for the main party. What will Mac OSX need to be able to do to dominate next time?
  • Stabilize 10.4 quickly. The OS tends to be locked down a year or two before landing to avoid surprises both in security and stability.
  • Better graphics card support. We need to be able to use the top of the line pro graphics cards for 3D models. If Linux has better support, then that is what will be used.
  • Be able to drive 3D stereo shutter in a window. Every other OS does. And be able to do this in both X windows and Aqua.
  • Get SGI to port Performer to OSX. RSVP uses Performer on Linux right now.
  • Make sure Java 3D is fully supported and fast. The science interface (formerly WITS) uses Java 3D
  • Push fink even farther in terms of packages supported and improved testing. There needs to be guidelines and documentation of how fink has been deployed across 200+ machines at one site.
  • Being able to make sure that all relevant fink packages are able to be build optimized if possible. Having fink on a G5 with code not taking full advantage of the compiler optimizations will frustrate the team.
  • Improved support for clustering, especially for render farms. It should not require Xserves.
  • Get OpenAFS shipping with OSX (and not just osx server). Many flight people count on AFS up to the transition to the flight line activities. Then they have to drop back to NFS for flight ops.
  • Go for more CPUs in one box... how about 4?
  • Really push 64bit support into more of the os and get way beyond 8GB of RAM per machine. We want 32GB and more soon.
  • Support the big Panoram displays.
  • Start shipping OSX with fugu preinstalled and do everything to make secure systems easy on the admins and users.
  • Make sure that there is good support for the VxWorks tools on OSX. I have not touched VxWorks since 2000, but since this is what is on the flight vehicles, all the cross compilers and tools need to work well on OSX. That will make testing easier. The Windows boxes that many people use for VxWorks are not allowed inside the flight systems.
Those are the main things I can think of right now. I really would like to see OSX and fink taking a bit part in the next missions.

Additional notes:

Forgot to mention that these are my personal opinions and ideas. These posts in no way represent the ideas of any institution! Nor do I have any say about the makeup of the flight computers.

Some more things OSX could do:

  • Stop the distinction between OSX Server and standard OSX.
  • Make sharing home directories between disparate Macs super easy. I have seen too many weird hacks to make this work right.
  • Provide something like SGI's FAM. This would make programs like the Collaborative Information Portal (CIP) easier to implement and easier of the servers. The idea is a filesystem change notification system. Some programs need to watch for changes in large/deep filesystems. Stat'ing large parts of a tree is totally unacceptable.
  • Make deploying servers with raid arrays in the 100's of terrabytes easy. 5.6 TB is not really that much storage for a mission these days. That's scratch space for just a couple of people processing stereo 3D data. Plus make backup easy.

Posted by Kurt | Permalink

01.14.2005 15:23

Jerry has retired

Just stopped by the ODP West Coast Repository building to wish Jerry Bode well for his retirement. He was serving up some insane cheese cake, strong coffee, and cookies. There was some serious story swapping going on including the Glomar Challenger being taken over by the Argentinian Navy way back. I had not heard about that one before. Right up there with the Ewing (sp?) being attacked by pirates over on the East side of Africa in Aug/Sep 2001. At least in Argentina there were no bullets fired.

ODP West Coast Repository

Quicksilver - What exactly is this Mac OSX app? Something about global hot keys?

A little applescript to set the osx desktop background from this article:
tell application "Finder"
  set currentPicture to the desktop picture as string
  set the desktop picture to 
   ":Library:Desktop Pictures:Solid Colors:Solid Aqua Blue.png"
  set the desktop picture to currentPicture
end tell
BTW, I donated $40 to tsunami relief last weekend.

The painters are still working on our building. Some guy is sanding away next door to the north. Will they ever do a 2nd coat on our railings? They never did sand the old paint before the 1st coat.

I have been working on movies now for SPIE, but I am desperate to get back to the AMS data. The Santa Barbara data looks so exciting at this point. I have got to measure the rest ASAP.

3:30 PM PST - Just saw my first picture from Huygens on CNN. Give me more! Nothing on the Saturn Photojournal. Give me more! I can not wait! The first picture I saw had a dendritic pattern. Rivers of methane with erosion? That would be amazing! Aha. Here is the Huygens Descent page. Landing on another planet or moon for the first time is a HUGE deal.

MathWorld article on
simulating the tsunami wave. Scroll down for a graphic of the simulation.

How the heck do I listen to NPR sound smil files on Mac OSX. Do I have to install Windows Media player? I do not really feel like doing that!

hidden doc preferences

Forgot to mention that today I got a dlo Jam Jacket 20 (crazy name, eh?) for my new 20GB iPod. It got the clearish one. Some of the colors were a little crazy. Seems to fit with my iTalk. My iPod feels less like it is going to slip out of my hands or pocket now. I'll try to remember to write what I think of it after a week or two.

I just tried to search my blog using google. It is not getting indexed very often. I need to fixup swish-e and the keyword/phrase index. I was hoping to find the ITConversations blog, but no luck. I have now listened to all 6 talks that I downloaded. Today I went through 2 on the bus: Phil Zimmermann and Rasmus Lerdorf. Both were pretty good. Now here is what I would like. I would got for 2 general catagories. The first and most important would be in geology and geophysics. How about interviewing people who are really good at a technique. Have them explain what it is and what it is good for. Go grab someone who does ground penetrating radar and grill them. What kinds of things can be studied with it? How do I go about getting a unit? How do I process the data? What is coming in the future? What about people recording their own conference talks. Most would really suck, but if we could get a stream of good ones going, I think there would be a huge impact on the community. It is easy to learn about computers... go to any bookstore or news stand and you will find tons. But the easy access to geology and geophysics just is not there. geology.it is better than nothing, but it is not /. Hmmm... I think I wrote about this before, but since it is hard to search my blog, I can't tell you in which entry and just how much of a broken record/scratched CD I am. :)
I was thinking of going for iPodderX to get pod casts from ITConversations, but I did not realized that it costs. No thanks. Into the trashcan it goes. I think I'll go for free software. What about iPodder? I installed that back in November, but have not touched it since. I'll give the big lemon on my dock a try for a bit. I see Adam Curry's name in the credits for the idea. I can say that he is a visionary, but I listened to two of his podcasts and that was enough.

So now that iPodder downloaded some audio files, did it do anything with them?

Posted by Kurt | Permalink

01.13.2005 18:45

PHP iCalendar, Galerie, and WordPress

PHP iCalendar looks like it might be what we need for a calendar program. Have to give this one a try. And I did. Took about 15 minutes to setup everything since I am already running apache2 on my laptop. Here is what I did:
  • Download phpicalendar from http://sourceforge.net/projects/phpicalendar - I got 2.0beta from Oct 2004.
  • Edit your /sw/etc/apache2/httpd.conf and add this for php4:
      AddHandler php-script php
      AddType text/html       php
      # PHP Syntax Coloring
      # (optional but useful for reading PHP source for debugging):
      AddType application/x-httpd-php-source phps
  • Uncomment the php module in httpd.conf
    LoadModule php4_module        /sw/lib/apache2/modules/libphp4.so
  • Change the httpd.conf index list:
      DirectoryIndex index.html index.php
  • cd ~/Sites
  • tar xfz ~/Desktop/phpicalendar-2.0b.tgz
  • cd phpicalendar-2.0b/calendars
  • cp ~/Library/Calendars/Work.ics Kurt.ics
  • sudo /sw/sbin/apachectl restart
  • open http://localhost/~$USER/phpicalendar-2.0b/index.php
  • That's it! Now you can go figure out how to configure other stuff
I used this blog entry to get the apache2 configuration tips.

Apple's iCalendar parsing page

Galarie looks handy for generating galleries of images. Ok! This is really nice. I made a test web page call GalTest from some pictures in iPhoto. It noticed my selected pictures.

Maybe Lisa would want to try PyBlosxom? Are comments build into it? Seems poorly maintained. Very messy and broken web pages.

Actually WordPress is GPL and sounds pretty good. wordpress.org If I was not already using NanoBlogger, I might try it out myself.

Posted by Kurt | Permalink

01.13.2005 08:46

density-0.18 released

Just out: A new 0.18 release fixes up density for SPIE Electronic Imaging. First, I added the headlight option to render. The more important change in density-0.18 is that the render and simpleview command scripts actually should be able to find the associated _bin binaries.

I should check out bashlib for writing CGI scripts easier with bash.

MIT has another tutorial on OpenInventor ASCII scene graphs. The InventorMentor is online here. I need to buy a new copy of this book!

Just sent out a request for ideas on deploying fink across some place like IGPP. Here is the text of the message:

This is something that I am not directly involved, but I would be interested in hearing what other think. One of the departments at my school is looking to deploy fink across about 120 machines which is about 1/2 desktops that are aways on the network and 1/2 laptops that will come and go from the wired and wireless networks along with going to sleep. The admins feel they need to maintain a global fink that they maintain that all of the users can count on. They also want to let laptop users have their own fink tree in say /sw.

So the questions are:
  • Has any group done this before and what lessons are you willing to share?
  • How would you go about this if you had the freedom to do it anyway you want?
  • If you kept a network fink tree how would you export it to machines? Appleshare? NFS? AFS? SMB :)What about those writable directories? If you have an apache server from fink, it wants to write into /finkdir/var... If you did NFS, you would probably want to do a readonly export right?
  • I know that all the binaries are supposed to be the same across platforms, but would a big roll out effectively require separate G5 and G4 trees (supporting the old G3's around here would not be worth it)
  • Would you just have machines periodically rsync in the debs and list of packages from a machine and have only local fink installs?
  • What else does this bring to mind?
Just found a new version of fink with selfupdate-cvs: "0.23.5 | tastes better than tripe!"

Thought I had a bug with Fire 1.5. I could not figure out how to send text, but I just discovered that you have to resize the composition window from the bottom. It is minimized to totally hidden by default and there is just a little dot down there you have to drag up.

I really need to learn svn. But when am I going to do that? Someone needs to tell me that emacs checkins will be just the same.

GeekTool and other programs on the person's site look interesting.

I have been watching my webalizer graphs and tables for about 2 months now. I would like to say thanks for Microsoft for bombarding my machine. For this month their bot (msnbot.msn.com), they have 6163 hits (15%), 4782 files(15%), 131MB (7%), and 80 visits (2%). Gee thanks. Maybe I should create a robots.txt file that tells Microsoft to take a hike. I've never created a robots.txt file before.

Myche shared this about streaming from an iSight using Java. Streaming QuickTime with Java

Understanding Audio Compression

TWIG does not cut it for a calendar program. Bummer.

Posted by Kurt | Permalink

01.12.2005 16:13

Good lookin data

I measured not as many samples as I wanted today for core 1GW, but the results are pretty nice. There is definitely clustering of the maximum and intermediate eigenvectors. I have a file that tracks the sample weight of the cubes called weights.dat. Check the CVS Id field to see how recent the file is. That also has some info on the naming conventions. One nice trick I have been using to catch labeling problems is to gnuplot the sample names. It should form a straight line for each core as I weigh the samples in order. That has caught a number of little problems early while I could still correct them.

Check out this news feed in NewsFire or Safari for the latest spirit images. Much easier that always going to the JPL MarsRovers (sorry Lance, but this is much easier!)

Ugh. We need a group calendar of events for the driscoll lab. Some quick research shows some promiss. Hopefully I can give this an hour or two when I get back from SPIE. Did not find anything in fink. Strange.
Anyone want to offer their opinions of these?

Posted by Kurt | Permalink

01.12.2005 07:44

Another day of sampling

Jerry is retiring after this week. Congrats! But that is a big loss for us :(
This article describes using DokuWiki. I would have ignored it accept for the fact that it apparently uses static files rather than a database. I am getting a bit tired of everything and its cousin requiring a database daemon be running. I like simple stuff! There is a time and a place for oracle, postgres, or mysql, but also a time/place for sqlite, db, or straight text files!

Wow only 30 spam messages this morning.

From Bob Ippolito - turn on readline support in OSX default python:
  python `python -c "import pimp; print pimp.__file__"` -i readline

Posted by Kurt | Permalink

01.11.2005 19:55

gmtplus continued

Right after that last entry, I sat down at Steve Cande's computer and tried to install gmtplus via fink with gmt 3.4.4 already installed in fink. I ran into trouble right away with GMT_LEVEL not existing and not having netcdf in the BuildDepend line. That netcdf thing caught me off guard. I would have expected that in gmt-dev. I had to update gmtplus to not try to use GMT_LEVEL for GMT 3, so gmtplus-3.4.1 was quickly born. That seems to have installed fine and now Steve should be testing it.

Steve provided some sample data for people to use to try out gmtplus. The byte order is the same between these Sparc Blade100 machines and Darwin, so the data should just work. These are some older Lamont cruises:

conrad eltanin vema
The new mac mini Looks pretty nice. Should fit anywhere. This could be an awesome little server if you could just attach a couple little (physical size, not skimpy on the GB) firewire drives. Great for clustering, no?

Has anyone tried using Mac OSX Thunderbird instead of the Apple Mail.app? Would be nice to use something that works across all platforms. Today was the first time I have seen someone actually using MS Entourage for email (on a Mac). What is the idea of having both Outlook and Entourage? Is Entourage the Mac version? Here is a faq entry about how to import mail from Mail.app to Thunderbird.

I talked to Lisa about content management systems for her paleomag class. Here are some quick notes:

The grad daddy is slash or slashcode which is based on http://slashdot.org it might be kinda hard to install... some stuff on OSX is here: Apple topic

Another that might be easier is phpnuke PHP nuke howto

A lot of people use movabletype which is about $40 for one class.

Here is the catagory list from sourceforge.

Maybe one of the wiki tools would be easy like moinmoin which Aurelio setup, which is based on python.

Posted by Kurt | Permalink

01.11.2005 15:39

gmtplus lives

Just ported to mac osx the gmtplus that JJ, Ed and Dru have hacked on for the last couple of years. And I made a fink info file. I am calling this version gmtplus-3.4.0. It should work with gmt 3.4.x, but I did not try it. Hah! I actually built things with gmt 4.0 and had to add some #if macros to get around the API changing for ps_plotinit and GMT_2D_to_3D for gmtplustrack and pluswiggle respectively. Any program that uses the decode_record calls is suspect with bad pointer mojo. BEWARE! Backup your brown book files and such before using this. This comes without data files. You will have to someone like Cande or Sandwell for data.

sudo -s mkdir -p /sw/fink/10.3/local/main/finkinfo/sci cd /sw/fink/10.3/local/main/finkinfo/sci sudo apt-get install wget wget http://schwehr.org/software/gmtplus/gmtplus.info less gmtplus.info fink install gmtplus

Posted by Kurt | Permalink

01.11.2005 09:46

Today's pmag paper - Carlut04

For the paleomagnetic meeting this morning, Julie has picked this reference:
Timing of volcanism along the northern East Pacific Rise based on
paleointensity experiments on basaltic glasses.
Calut, Cormier, Kent, Donnelly, and Langmeir
JGR, 109-B04104
First off, I have to say that not doing the volume/issue/page number thing is idiotic. Thanks AGU. Grr. Some notes from the paper:

  • EPR spreading rates are 80-150mm/yr implying dike injected every 7-13 years if they are about 1m thick or it maybe less frequent events with many dikes
  • Lava flows may extended several kilometers from the vent. Wow. That is incredible that it can go that distance with cold water right up against it.
  • Used the Periodic Table X software to lookup what Po is: 84 polonium 210. Was mentioned in the paper for Po-Pb dating of fresh lavas.
  • Using sedimentation over a lava flow is bad in that the sedimentation rate is dependent on hydrothermal and tectonic activity - why tectonic activity?
  • Paleointensity from fresh MORB eruptions give zero age test determination on the order of 1µT. That would be very useful if the results are of this quality. This is from averaging many samples from one flow.
  • Found that the centers of the ridge segments tend to have younger flows than the ends.
  • This study does not actually do transects of glass samples.
  • Study area was chosen due to a big seismic experiment here to map layer 2A.
  • Goal to comparare north and south segments. The northern segment is an unusual inflated zone with a wide central valley.

ASC Scientific is the place that I get paleomag cubes from. It is the 8cc sample boxes that I use for my fabric research.

This morning was one of many umbrella inversions trying to get here.

I am going to meet with Steve and Dru about getting mggplus/gmtplus (or whatever it is called) turned into a fink package. Steve will hopefully be giving it a run through on the RVIB Palmer for the next month. So when he gets back, maybe I can commit it to the fink tree.

Trying out the upgrade to Fire 1.5. We will see if I like the new tabbed window approach. Not that I often talk to more than one person at a time. What is the last column for? It is blank.

Posted by Kurt | Permalink

01.11.2005 06:34

Sending events to the Coin viewer and ...

Oh wow, on 60 spams made it past SpamAssasin this morning and yet another "your account is over quota." Wee. :(
Here is a bit of code to keep in mind if I get back to working on my drivers for the Space Mouse. This snipped by Peder sends a mouse button up event to the viewer:
  SoMouseButtonEvent event;

I forgot to mention, but yesterday I released a new fink info file for SimVoleon 2.0.1. According to SIM, 2.0.1 fixes the offscreen rendering trouble with Voleon data. Unfortunetly, it is now missing the doxygen documentation.

Another package that would be good for fink: radartools for processing SAR images. I have not done SAR processing before, I but I know a lot of folks at SIO do. Never mind! I looked at the page for it and it requires IDL. That totally violates fink policies.

Could Cinelerra be ported to Mac OSX?

Things for the iPod: ITConversations. Now that The Linux Show (TLS) is gone, I need something for the bus!

I can now say that the About Geology feed is officially annoying. It keeps republishing old articles from weeks ago. I already know there was a 9.0 in Sumatra. Thanks. If you lived under a rock in the back of a cave, you would know by now.

Every home has got to have an MP3/ogg stream, right? Well I do not have one, but if I did, it might be this one: GNU MP3 Daemon. Jose, what do you use for your ogg stream that I listed to with xmms during the fires a year ago?

NASA's B15 page shows an interesting view. It was over to the left about half way across the water when we when by it two years ago. This succcer is huge: Photos from NBP0209 of B15. Plus NASA put out a cute little movie.

Apple now has an new web site: IT Pro

Posted by Kurt | Permalink

01.10.2005 15:36

Canyon and Channels Eos article

Looks like I am going crazy with title capitalization. Oh well. This is the first entry in the Science category. This is for anything non-Mars specific and hopefully lots of information relevant to my thesis! Here is another interesting Eos article:
  Cap Timiris Canyon: A Newly Discovered Channel System offshore of
  S. Krastel el al
  Eos Transactions, 85-42, 19-Oct-2004, P 417,423
For starters, I did not know where Mauritania is situated. It is off the west coast of Africa, south of the Canary Islands. These authors mapped parts of a huge channel system that has no large present day river system at its head. They show a core and a small seismic line through the area. I have not read the article in detail to comment, but it looks promising. Could use a more in depth journal article with more and larger seismic lines and a much bigger multibeam figure!

Posted by Kurt | Permalink | Categories: Science

01.10.2005 14:44

Mars analog - Argentinian volcano

This is my first try at categories. I just created the Mars category like this:
  nb -t "Mars" -c new -a
That worked and created a file called data/cat_2.db.

Here is the article reference that caused me to start creating categories and this is really why I wanted to use a tool like Nanoblogger. I wanted to be able to comment on things I see and then be able to get back to them easily in the future.
  Capahue Volcano: A modern Terrestrial Analog for the Opportunity
    Landing Site
  Varekamp, J.C.
  Eos Transactions, AGU
  P. 401, 407
The article talks about the unusual mineralogic and compositional findings from Opportunity being rare on Earth but can be found in in alteration zones with fumarolic activity in some volcanic craters. Sulfur rich gasses create acidic fluids that leach rocks creating bleached areas rick in hematite cristobalite, gypsum, alunite, or jarosite. In the Copahue Volcano, he says that the majority of the water that is leaching material is coming from the magma. What would happen if there was a large amount of this kind of activity over a large period of time? Then you block the outlet, build up pressure, and get a plinean eruptions? What would this say about the possibility for life? The spring that the author studies is coming out at a pH of 1.3 and 80 °C. Ouch!

There is also a great picture of the author standing over an outflow with red hematite precipitates along the side. I did not see any discussion of what is leaving in these waters. That would be interesting to compare to what is in Yellowstone where I have looked at things like gray and pink streamers, phormidium, and calathrax. I forget the name of the super high temperature yellow organism. I took these pictures back in 1995 when working for Deena Braunstein and Don Lowe. These links may break in time when I fix these sites.
  Phormidium Picture
  Calithrax Picture
Not Mars stuff, but here are two links to UCSD repository of documents, PhD Theses, etc.
  SIO Pubs
  SIO eScholarship Repository

Posted by Kurt | Permalink | Categories: Mars

01.10.2005 11:18

density bug

Just ran back into a bug in density that I forgot was there. So here is the work around until I get it fixed. First the problem:
  simpleview body.wrl 
  /sw/bin/simpleview: line 9: ./simpleview_bin: No such file or directory
The simpleview script is supposed to set the DYLD_LIBRARY_PATH for the simage library and then call simpleview_bin with all the same args as usual. Looks like I forgot to really handle finding simpleview_bin in the scripts. So for now, you need to set the DYLD_LIBRARY_PATH and use simpleview_bin directly. Sorry about that.
  export DYLD_LIBRARY_PATH=/sw/lib
This will mess up a few things like man page formating. To undo this change when done, just type this:
When I get this fixed, I will release density 0.18. Do not expect this anytime soon. An alternate fix is to edit /sw/bin/simpleview and /sw/bin/render and change "./" to be "/sw/bin"

Posted by Kurt | Permalink

01.10.2005 09:41

Harvard adsabs

I have been talking to the Harvard ADS people about RSS feeds and it is starting to happen. I just tested an earlier version and this has serious potential for helping all of us scientists keep up to date. I really hate email notification. Soon I will have a NewsFire channel with the latest abstracts. From the test run this morning, I got this article: The Tharsis Region of Mars: New Insights from Magnetic Field Observations by Johnson and Phillips. And then I got the pdf. Hats off to Alberto and anyone else working on this!

Posted by Kurt | Permalink

01.10.2005 08:41

Wanted - FinkDesigner

Here is something that would make small project for someone (I wish I had the time to do this!): FinkDesigner

This program would be a GUI design tool for fink info packages. I would like a tool that guides people through the development of the info package providing guidance and documentation while at the same time enforcing the fink packaging rules. It would let you select the type of package, say normal, perl, python, with/without ssl, and then guide you from there. Does the package use autoconf? Where is the source. Then grab the source, fill in the md5, and walk you through build and install. It would be great to have a graphical tree of what is build and be about to say, no this file needs to be renamed and moved here. Select which files will be a part of the DocFiles.

The program could do all sorts of niceness along the way. It can detect in shared libraries and help implement the shared library policy by doing such things as otool and figuring out the library version numbers. It could provide a pallet of description fields that let you drag and drop fields in. Simple check boxes that let you override fink policy if you must make an exception. It should also be able to handle both developers with CVS access and submitting through the fink tracker. It should help you upload and track versions of the info file and responses from others in the talk backs.

It could maybe even have a custom watcher models that looks in a couple places for newer versions of the source files: the Source location, freshmeat, rpm/deb sites, etc.

Note to SGI: you guys should really port fink to IRIX. IRIX is just clinging to life. That would really help keep IRIX relevant. The freeware CD's were always a nice idea, but never worked very well. Without the tools in fink, I hate using IRIX. I spend forever downloading and compiling software by hand to make systems barely usable.

Over at techbooksforfree.com, they have Introduction to Data Communications. Looks like it covers a wide variety of topics! Quality? I do not know. I have yet to read any of it. Looks like a good introduction to topics if you are a beginner. There are some other interesting books there about LISP, Parsing, PostScript, Subversion, Numerical Recipes in F90, Emacs Lisp. Here are some things that I would like to read when I get a chance: A little code from the Coin mailing list. This is a nice example by Peder of having a dragger control the size of a sphere with an engine.
#Inventor V2.1 ascii
Separator {
  Separator {
    # scale dragger
    Scale { scaleFactor 0.1 0.1 0.1 }
    # consider translating the dragger to the corner of the Cube
    # Translate { translation 0.05 0.05 0.05 }
    DEF dragger DragPointDragger { }
  Translation { translation = Calculator {
    A = USE dragger . translation
    expression "oA = vec3f(A[0] * 0.1, A[1] * 0.1, A[2] * 0.1);"
    } . oA
  Cube { width 0.05 height 0.05 depth 0.05 }

Posted by Kurt | Permalink

01.09.2005 17:13

swish-e never ends

I am getting closer to getting swish-e working for my blog. I have been working on my laptop to get this going. It was not supposed to take this long. It was supposed to be a time saver. Doh! Here is a bit of what I have done so far on Mac OSX 10.3.7.

I now have it so that searches work but are not pretty. I have the three info files installed: swish-e.info, swish-e-pm.info, and html-fillinform-pm.info. For the two perl ones, I am using pm584. For testing, I have BLOG_DIR=/Users/schwehr/Sites/blog which gives a URL of:
Then I made a directory ${BLOG_DIR}/cgi-bin and then added these lines to end of /sw/etc/apache2/httpd.conf:

        Options ExecCgi
	SetHandler cgi-script

Then restart apache2 from fink like this: sudo /sw/sbin/apachectl restart

Now copy in the original search file: cp /sw/lib/swish-e/search.cgi $BLOG_DIR/cgi-bin. You then need to change the perl being used on the first list of search.cgi in your cgi-bin dir. The first line of mine looks like this:
If you do not do this, you will get an internal server error.

I then added some html for the search text box on the right to the template/main_index.htm file right before the contact entry:
<div class="sidetitle">Search</div>
<div class="side">
<form action="cgi-bin/search.cgi" method="get"><input type="text"
name="query" class="input" /> <input type="submit" value="Search"
class="input" /></form>
In the cgi-bin dir (this is not the best place), I greated the swish-e index:
  cd ${BLOG_DIR}/cgi-bin
  cat swish-e.conf
    IndexOnly .htm .html
    DefaultContents HTML*
    StoreDescription HTML* 
    metanames swishtitle swishdocpath
  swish-e -c swish-e.conf -i ../archive ../articles
  open http://localhost/~schwehr/blog/
Now I need to constumize search.cgi since it comes up whacky buy reading cd BLOG_DIR/cgi-bin && perldoc search.cgi

Posted by Kurt | Permalink

01.09.2005 14:38

Core 1 sample photos and movie

Core 1GW-Sec2 has been sampled. I altered the grab frames script to only delay one second after each frame is processed and it now talks if MulleSight stops working. I still should make the convert happen in the background to get closer to one second. Here are the links for the photos and movies: I can't wait to get the results from these 40 samples.

Posted by Kurt | Permalink

01.09.2005 08:58

Sampling core 1GW-Sec2

This is the big tamale core: 1GW-Sec2. This core is in the toe of the slide. So far, I have described the core and setup the iSight. I have it taped up on the filing cabinet looking down at the top of the core. I tried taping it to the vent pipe, but it fell 5 feet down onto a table. The pipe is wet. At least the iSight did not break. It looks like the maintenance people did not fix the leak from the roof vent. I have the core on the table positioned that the overhead vent will not leak on the core, just next to it on the table.

Posted by Kurt | Permalink

01.08.2005 21:13

More swish-e

Finally search.cgi runs. I had to create a total of three info files to get everything working. Drafts are available in my fink scratch area. These files are currently far from complete.

Posted by Kurt | Permalink

01.08.2005 17:47

iTalk and microphone

I just bought a clip on microphone for my iPod and iTalk. It is a Sony ECM-C10 for $40. It works well. The only thing is that the record level is a little low, but I am sure I can fix that with Audacity or some other audio tool. I am now totally equipped to record myself while I make my presentation (and chase my cat around the house). I talked to Debi earlier this week about releasing the talk as a power point exported movie with voice over. Very exciting stuff.

For the time lapse video of core sampling, Christie had all kinds of ideas when looking at the video. she suggested a voice over with explanations of what is going on and adding labels that point things out. Sounds like a good idea!

Just got "How To Wow - Photoshop For Photography" by Davis and Willmore. We will see what I think of it. It has a CD in the back with sample files. Have to give it a try to see what I think. The topics are certainly right, nice color images show what is going on, and the price was good (half off at Borders) at $24.

Posted by Kurt | Permalink

01.08.2005 07:46

Current rainfall

At 7:40 AM PST I measure 2.5 inches of rain with a highly scientific device: a plastic tub of butter container outside my front door. One minor detail is that there are a bunch of plants in the tub.

I got some feedback this morning from my twisted python email to their list. Thanks to Philipp Schmidt, I will double check my info files and commit them. He even created an info file for akaDAV. These python programs with a setup.py seem to work really well! I was terrible trying to setup new modules in the python 1.5 days. There was no standard way of coping.

Thanks to B10m for his post on how to setup Swish-E. Looks like it is not to bad and it gets even better: 2.4.3 now using the GNU autoconf system so it should work nicely in fink! That will be nice. My blog is getting long enough that I am starting to have trouble finding entries that I need to get back to for whatever reason.

Just found a website called MARS GEOlogy: marsgeo.com. I have no idea on the quality of information, but the author seems to have interesting posts on a mars discussion group. I have not yet found a bio on the author, Aldo.

Info on how to turn a pdf into a text file on your pdf can by found on macdevcenter - build an eDoc Reader

Does gengetopt now have specific c++ support? Strange log messages on the fink commits mailing list. Doh! The info file for gengetopt does not seem to get installed into the info directory?

Posted by Aldo:
  ... describes the geological history of the Gusev area. It outlines
  paleolake boundaries south of Gusev, describes how Ma'adim Vallis was
  carved, how Gusev was flooded, at what flow rates, what volumes of
  water would've been required, how long the flooding event might've
  lasted and so on.

Rossman Irwin III, Alan Howard, Ted Maxwell: Geomorphology of Ma'adim Vallis, Mars, and Associated Paleolake Basins. Journal of Geophysical Research, December 30, 2004
Started working on a draft swish-e.info file. It still needs split-offs for -dev and -shlibs.
  otool -L /sw/lib/libswish-e.2.0.0.dylib 
        /sw/lib/libswish-e.2.dylib (compat ver 3.0.0, cur ver 3.0.0)

I spent a while trying to figure out how to setup swish-e. It works fine from the command line, but I can not use the -e option to make it conserve ram. It writes out too many tmp files and runs out of file descriptors. Then I tried to get the cgi-bin search to work. That does not work since I can not control where the perl files go. They end up not in the PERL5LIB path. Perhaps I have to make a separate perl module info file that goes into swish-e-2.4.3/perl and does the whole perl versioning thing. Here are a bunch of notes:
cat swish-e.config
IndexOnly .htm .html .shtml
DefaultContents HTML*
StoreDescription HTML* 
metanames swishtitle swishdocpath

swish-e -c swish-e.config -i ../Viz04

swish-e -c swish-e.config -w "Santa Barbara"

# SWISH format: 2.4.3 # Search words: Santa Barbara # Removed stopwords: # Number of hits: 5 # Search time: 0.018 seconds # Run time: 0.061 seconds 1000 ../Viz04/resume_print.html "Kurt Schwehr | Resume" 9756 1000 ../Viz04/resume.html "Kurt Schwehr | Resume" 10729 633 ../Viz04/pop1.shtml "Kurt Schwehr | Presentation" 2818 633 ../Viz04/index_print.html "Kurt Schwehr | Home" 5960 633 ../Viz04/index.html "Kurt Schwehr | Home" 6099

<Directory "/Users/schwehr/Sites/cgi-bin"> Options ExecCgi Includes MultiViews Indexes SetHandler cgi-script </Directory>

./search.cgi Can't locate SWISH/API.pm in @INC
Back to writing the thesis...

Posted by Kurt | Permalink

01.07.2005 18:56


I finally got my little script for MulleSight on line in my scripts directory. Check out the code2html version:

Posted by Kurt | Permalink

01.07.2005 14:39

Sampling movies

Yesterday I sampled with the iSight running MulleSite. So here are the movies I made after pruning with the remove-dup-frames-0.1.bash script. Check out the search function added to nanoblogger over at B10m. That would be nice to add!

Posted by Kurt | Permalink

01.07.2005 11:46


Alright! pmag-1.8.3 is now in the fink unstable tree. Time for everyone to do a fink selfupdate-rsync. Please let me know one way or the other if it works. Both positive and negative feedback are critical. Generally with fink packages, I get no feedback. Remember that it takes some time for the sf.net mirrors to pick up the tar ball, so select use source URL fink complains that the source is not on your default mirror (usually option 5)

The rain has started!

Just posted to the twisted python mailing list. Unless I hear from someone (anyone!), I am going to drop getting the info files for twisted, buildbot, cvstoys, etc info fink. I never figured out how to get buildbot running. It looks cool, but it takes more time than I have to figure out how to set it all up for my really basic situation. I really need some document that holds my hand and tells me exactly what to do. With all that power comes some complexity which can leave out users like my that are too casual to really dive into it and invest some time.

For my Mac trick of the moment, check out this. First quit x11. Then open a terminal and run the following command. open-x11 makes sure that x11 if first running and then runs the command. This is good for wrapping x11 programs in a double clickable Aqua app.
  open-x11 xterm

For example: marine-geo-fink-v2.dmg.bz2
Another idea for nanoblogger is to be able to work from an RCS or CVS file. I could then maintain a text file with my todo items in it and nanoblogger could check the revision history for new entries and allow folks to see the list of time in addition to the latest entry. Maybe not that useful, but maybe someone would find if very handy.

I really would like to know how to pull all the vcalendar entries from iCal with Applescript. It has to be pretty easy to do, right? vCalendar entries get a .ics extension. So now that I look in ~/Library/Calendars, I find that there are raw vCalendar files for each catagory of calendar. That was easy. Now I just need to have an easy to use package that understands vCalendars or write some grepping type stuff.

Check out this ical blog. It has been dead for two years, but there are interesting tips. Doing things like using iCal to schedule things like cron or anacron. Interesting!

A weblog to iCal script is discussed on Orielly's Weblog Developer page here. I know, it's from 2002. I now want something that goes the other direction.

Maybe this python package is the way to go: PDI. But there is nothing to download?!?!

You may want to check out Apple's free list of dates that you can import from the ical library

libical is now a part of Free Association, but there are no updates since Nov 1002. The last official release of libical is 0.23

vobject 0.1 is a python vCard and vCalendar project that actually has a recent date (Dec 2004).

That is enough for now on calendar stuff!

BTW, if my server suddenly disappears in the near future it might be due to the leaking fire sprinkler right over my desk. This yellow goo just missed the iMac's LCD screen by about half an inch. What a thing to discover when I got back from Christmas break. Yuck.

Posted by Kurt | Permalink

01.07.2005 08:04

Processing yesterday's data

The news this morning is dreading the coming of the pineapple express that is supposed to come in and dump 7 inches in the next couple days. Look out on the beaches for the San Diego sewer system. It has not started here yet. Just damp for the morning air. They are sawing this could be the biggest storm to hit San Diego in recorded history. I am sure the rest of the country will laugh at the amount of rain as trivial.

Now I have to get through all the stuff generated yesterday. Last night I forgot to save an exit my blog entry, so this morning I had to hand the data entry which has 07 Jan in its name. Whoops.

I have yet to write about the demise of The Linux Show (TLS). I was just starting to get into the show and realized that Jeff is a really good host. I always like hearing Eric Raymond's opinions. I know it cost them a lot for bandwidth, the pbx, and phone services in addition to take up a lot of their time that they could be billing out. It was a great show and I will miss listening to it on the bus going to and from SIO.

Yesterday, I upgraded to NewsFire 0.62. What is different? The changelog/release notes do not indicate exactly what is different.

Spam is getting out of control on ucsd's email services. I now get a message every couple days that I have gone over quota. I am getting 60-80 spams a day that make it past their spam assassin installation.

IMP from blursoft.com might be a good thing to use to generate thumbnails rather than my little bash script. Doh! Looking down the lists of dependencies is Mono. I have yet to install or try Mono.

Wow an interest geology article made it past my RSS feeds. This article talks about halloysite clay at the Dragon Mine in Utah. It goes into the uses of the microtubes serving as carriers for a wide variety of nano applications. Clays are crazy materials.

The same company that build the Canada arm for the space shuttle got the contract to build the robotic servicing mission for Hubble at $154M. Article (space.com)

CoinDesigner - This may be cool. I have not had time to check it out. It's under the GPL license.

How to make a graphviz view of the dependencies in fink from Richard Cobbe:
    apt-cache dotty > dependencies.dot
    dot -Tps -o dependencies.ps dependencies.dot
I am now trying to process the photos from yesterday when I sampled core 5 from BPSIO-04. Why doesn't Shift-Apple-R not work to rotate? Apple-R does work for the other direction. What gives? Oh, that was Option-Apple-R. The images for sampling core 5GW photos were taken with an Olympus Camedia C-3000 Zoom with 3.3 MPixels. Look at the photos here. Here is the script that I use on the webserver to generate thumbnails. This one uses the netpbm tools. I have new ones that use ImageMagick, but I found this one first.
I also moved the nb index/phrase/keyword too and generated a little html view using code2html.
This is what SGI should give to all users of RAID array systems. This is what I asked Ed to find for stowe (which he still has not found anything).
Perl script for checking software RAID status
I really need to try GNU ddrescue on my old Jaz Disks sometime soon! It would be nice to recover all the Haughton data. I think I got all the Yellowstone data on CDs. I should make 1 DVD for each project. It is a shame to not have that stereo data up on the web for all to use. Plus to run stereo on all of it too now that stereo can select the other eye to avoid the yellow eye problem I had in yellowstone with a broken wire making images yellow.

Posted by Kurt | Permalink

01.06.2005 20:49

Sampled another core

I spent this morning up at the Core Locker/DSDP West. Sampling the working half, section 2 of core BPSIO-2004 went well. Took from about 7 AM until 1 PM. Takes a while to do things right. I tried a new thing today. I brought my iSight along and taped it on the wall. During the day, I ran a script to grab frames every to sections to record the "action." So sampling cores is not the most action packed, but the hope is to illustrate the process. I had a lot of troubles with the recording, but it basically worked. MulleSite seemed to stall a bunch and then just stopped grabbing frames. I recorded a thousand copies of the same image. I had to stop and restart bot the script and MulleSite.

Lisa and I worked on pmag today. We got out a release of pmag numbered 1.8.3. This one is the first release with a new make system. It needs a bit of work still on the makefiles, especially for building the cygwin binaries automatically. I hope to get the fink info file out by tomorrow.

Posted by Kurt | Permalink

01.05.2005 19:07

Using mbsystem on the mac

Since this keeps coming up, here are some notes on how to install mbsystem on mac osx. This note only is good until mbsystem 5.0.6 comes out.

First off install fink from the binary installer if you have not yet done that. Note that if you are using Mac OSX 10.2 it is time to upgrade to 10.3. I won't even try to support 10.2. You will not be able to install it from fink! Do not forget to put the fink commander application in your /Applications folder (and I recommend the Dock too).

Then start up Fink Commander. Type "gmt" into the search window in the top right. You will then install gmt and gmt-dev 3.4.5 with the blue install button. Do NOT install GMT 4.x until mbsystem 5.0.6 is released!!!

At this point, make sure you have installed the Xcode developer CD and the X11SDK installer also on that disk.

Now type mbsystem into the search field on the type right. If there is no mbsystem, you will need to go to FinkCommander->Preferences->Fink and select unstable packages. Close the preference window and wait for it to rebuild the list of packages. If you still do not see mbsystem, do Source->Selfupdate-rsync (or if you already use the cvs version of updating, you know what to do).

Once you select mbsystem and hit install, it will take about 10 minutes to install mbsystem 5.0.4.

I hope that helps!

I tried to bake a leek tonight for dinner. Never done that before. I got very mixed results. The center was fantastic, but the outer leaves were tough. I just slice off the top and then halved it down to almost the root. Did the get the dirt out in a cup of water. Then shoved some butter inside and cooked it in the oven at 350°F for 15 minutes. Next time I should try this recipe: Baked Leeks

Posted by Kurt | Permalink

01.05.2005 14:37

Converting to unix formatted text

This keeps coming up. Our lab instruments save data in a format that causes the unix tools to not be able to read them. First off, you should open a window with man ascii in it. First I opened the original data file in emacs. Emacs puts a (Mac) in the bottom left corner to let us know that this file in not a unix text file. The other thing you might see would be a (DOS). There will be nothing there for a unix style file. Then put the emacs window into hexl mode like this: M-x hexl-mode.
emacs for mac format file:
  000000c0: 310d 6270 3034 2d36 6777 2d31 3333 2e30  1.bp04-6gw-133.0
              ^^                                      ^
man ascii:
  00 nul   01 soh   02 stx   03 etx   04 eot   05 enq   06 ack   07 bel
  08 bs    09 ht    0a nl    0b vt    0c np    0d cr    0e so    0f si

emacs for unix text file: 000000c0: 310a 6270 3034 2d36 6777 2d31 3333 2e30 1.bp04-6gw-133.0 ^^ ^
So the new line for mac is 0x0d in hex and we need to get these files to 0x0a. Which from man ascii, we see that this is changing cl to nl. If you have GNU tr installed (with fink you need to fink install textutils to get gnu tr), GNU tr allows you to use some standard escape sequences to make things easier to remember. The command to convert mac to unix then becomes:
  tr '\r' '\n' < rawdat.k15 > myunixdat.k15
This also helps if you keep your data in a CVS tree. If you do not convert the data file to unix, CVS will consider the entire file changed since it sees one line that is changing.

A total change of topic, but it is nice to be able to track airline flight status on line using flightview

Posted by Kurt | Permalink

01.05.2005 08:03

NewsFire 0.61

I just installed NewsFire 0.61 and it still does not have the mouse over summary available. On further search, I did find the Item List Appearance selector under NewsFire->Preferences->Appearance. Then I selected Headline and summary style. Awesome. This was what I was looking for. Now I can zip through freshmeat in 15 seconds flat and know that I did not miss anything good.

The day before yesterday I was complaining about not having a show desktop button. As timing would have it, I walk talking to dkilb and she did just that. "How the heck did you do that?" With her response being "Just hit F11". Exposé at work.

Last night I started looking at the blogging awards called The Best of Blogs Awards for 2004. Besides the fact that there were no technical or scientific categories, I do not think I would have got any votes if I had been nomination :)Best Knitting blog. Even though Joanne Stocked showed me that it is a good thing to take along on long research cruises when you have to watch stand for 12 hours a day.

I had forgotten about Cooking for Engineers. Time to give it a read for a while to see what I think.

Mars Surface Laboratory on space.com

Another OpenInventor file format tutorial Studierstube scripting mentor

After using NewsFire 0.61 for a bit, I still would rather have the hover mode for the descriptions.

I had another idea for a nanoblogger feature in addition to the keyword/phrase index and the URLs by most recent and most frequent:
  • A calendar of events. Perhaps even using a little applescript to extract certain types of iCal events
  • If someone wanted a currently listening to type feature, that could be easily done with applescript.
  • Have the archive index also have the description in addition to the title.
  • Chronologically, for each article show the keywords that occur. This would be the super compression mode. The idea is to give the gist of what is going on in each log entry.
Check out this little bash script called itunes that gives a shell interface to iTunes. Here is applescript to get the current song being played:
osascript -e 'tell application "iTunes"' -e 'set theSong to name of \
current track' -e 'set theSinger to artist of current track' -e 'set \
theTune to theSong & " - " & theSinger' -e 'return theTune' -e 'end tell'
Sometimes I find things in my notes that I have no clue what they were for. How about this little bash thing that sends something to applescript. Why was I cutting off the first 4 characters?
  echo $(cat "$FILE" | cut -c 5-) | osascript
ffproxy - Perhaps this is what Aurelio and I were wishing for last year.

Posted by Kurt | Permalink

01.04.2005 20:54

A day of measuring and NOVA Mer

I spent a good bit of today measuring Anisotropy of Magnetic Susceptibility (AMS). Just noticed that the Wikipedia has an entry for paleomagnetism, but nothing for AMS. I can get pretty close with Magnetic Susceptibility

I am just now playing with the emacs html-mode to see if that makes it easier to catch mismatched tags. It is a little strange in that indenting is really bizarre.

It sounds like it is raining pretty good out again. It's nice hearing the rain on the roof. Not as nice as described by WWdN.

NOVA - Random comments to follow

NOVA's Welcome to Mars is just starting up. I can't remember if I talked to these guys or not. I remember talking to NHK. There was some talk of getting a 3D model Lightwave and in for something, but I never really asked what it was for.

They started with Justin's vicar pasting of all the images on the projector.

Opportunity driving up on driving up on the big crater. I haven't seen that video before. Is that really 3D or tricks with a color mosaic?

The Spirit trouble. Ugh. Why doesn't vxworks provide a real filesystem? The dos filesystem. Why do they have to focus first on this trouble? I know it's exciting, but ugh. I'm glad that I wasn't one of the critical path people for debugging that kind of thing.

Those guys in the sandbox really have some patience. That is more than I could do. That were running continuous tests all day everyday and they go SLOW!

I am having flashbacks to watching the CIP (Collaborative Information Portal) displays checking for when the next downlink is going to occur.

Hear comes the water-life discussion.

"Bedrock is the holy-grail of geologist"

Hey, there is Nathalie Cabrol talking about Gusev possibly being a lake ala Haughton Crater.

And there is Peter Smith.

Lots of 3D movies that I don't know. Someone has been busy. Who? They use the Dan Maas model of MER.

There are scenes with CIP and RSVP.

Another Opportunity movie in the sunset that I do not know. Cool.

Ugh. There is that darn ice cream freezer.

There was the top of MWM's head.

Lots of video that looks like it was done by Shigeru. Yellow boxed insets with clean pans.

Blueberries -
  • tektites: impact beads - should all be in one layer
  • they prefer the concretions theory

Hey, we have concretions coming out of bedrock at SIO up by the soccer field at the head of the canyon. I wonder what has happened to them with all the rain?

They are saying that this has to be an evaporite (salt flats) due to the high content of sulfate rocks. Steve says that is impossible to explain without water. Then there is the ripple evidence. Ask me what I really think.

Oh boy, here is the DC salty sea presentation where they used hardly any graphics. Steve talked for over 5 minutes without any images. CNN dropped random mer images over him talking until they cut away. Sigh. That was hard to get quality materials transfered back to DC last minute. Looks much better now on NOVA. Thanks!

Goethite - "Can only for in the presence of water"

Das ist alles. I did not see anyone from SSV, but hey, we were the folks behind the scenes helping to glue it all together for the world to see. The fact that we were not seen says that we were getting th job done :)
That was a fun show to watch. Thanks NOVA!

Posted by Kurt | Permalink

01.04.2005 17:13

MER on PBS tonight

The Mars Rover Blog came through today... Nova is airing "Welcome to Mars" for the one year of MER on Mars tonight at 8PM. Maybe I'll show up on the footage in the background. If they show the huge panorama printout carried in but about 8 people, my knees will be the first on in (on the very left) and I am the only one wearing shorts. Perhaps you can see an orange hat too. Next to me was Aurelio I think.

Posted by Kurt | Permalink

01.04.2005 11:41

Paleomag meeting

Today we are discussing the MagIC database with Lisa leading the discussion. MagIC == Magnetics Information Consortium. Lots of text interface python scripts.

Posted by Kurt | Permalink

01.04.2005 07:28

It is seriously raining

Well it is seriously raining by San Diego standards. 9000 gallons of sewage spilled yesterday at Tourmaline Beach. Yummy. You have got to love the San Diego sewer system.

I am rebuilding the HEADER.html for the core 6 paleomag samples so that when you click on a thumbnail you actually get the picture.

An interesting article in the New Hampshire news paper: UNH scientists not surprised by catastrophe. It quotes Larry Mayer and Jim Gardner of UNH. I still can not believe that the Indian Ocean area does not have any sort of tsunami warning system.

And then there is an article from Germany (Germany?) that quotes Chris Goldfinger from OSU. 'Clusters' of earthquakes yield and ominous scenario

I have yet to revert NewsFire to 0.5, but I have got to do it soon. 0.6 is driving me nuts. It is much slower to see which articles are worth reading.

Back to sampling. Just in case you were wondering (and I sure you were not), the core 6 sampling is under request number 750 from the SIO Geological Collections.

Posted by Kurt | Permalink

01.03.2005 08:24

Rainy Monday morning

Is that thunder in the distance? Yup. Saw some lightning too. Outside the plants are drowning.

Why is there no button on the Mac to drop all open windows to the Dock? Even Windows has this feature. I want my desktop back, but there are too many windows open.

Just upgraded to NewsFire 0.6. I only saw the new version on versiontracker while grabbing a new copy of VLC to view a Windows Media file. It is strange that NewsFire 0.5 would not tell me there is a new version. Maybe 0.6 will fix my problem with it. The program stops refreshing after running for a couple days. If I quite and restart then it will start working like normal again. Doh! NewsFire 0.6 does not show the sumary when the mouse is over the article. I use that mostly with freshmeat entries.

I spent 5 hours in the DSDP West yesterday sampling core 6 from the BPSIO cruise.Sampling the top 35 cm was no fun. It is so soupy and sticky. It takes a lot of time to do each sample right. This was my first attempt to take as many samples from a core as possible. It tries 2 side by side samples, but I don't like that very much. I worry about getting close to the edge of the liner. The SIO King Kong really is not that big. Here are photos from the sampling. The iSight movie shows that taking a movie of your core is not a useful thing to do :( Sampling photos iSight movie Excel Spreadsheet with core description but no photo A quick tip from LinuxGazette
  convert file.jpg eps2:file.eps
The eps2 specifies the postscript level. According to Kapil, the jpg is just wrapped in some header goo for anything higher than level 1. Nice! Accept I don't really use postscript anymore. I used editable pdfs for my last paper with LaTeX.

Search the whois database

Posted by Kurt | Permalink

01.02.2005 12:37

iSight frame grab

I still have not found the simple write a frame to a ppm or rgb data in memory program for the iSight or other firewire cameras on the Mac but I'm going to give versiontracker another chance to see what I can get for saving frame from my iSight. I am going to go into the lab and sample some cores today. Taking pictures with real film gets expensive quick and then I have to scan them or pay more to have them scanned. With digital, I get instant feed back to know if they are good enough. Here are the programs I found on version tracker that are free: gcam takes movies without any troubles, at least that is when you remember to twist the front of the iSight to the on position. However, I don't know how to take just one image.

Ok! MulleSight is where it's at! I down loaded it and 5 minutes later, I know how to grab frames using a little applescript ditty that came in the dmg. This rocks! Now I need a 2nd iSight and a pantilt unit and I'll be in business with 3D stereo. Woo hoo. But for now I can a super easy way to grab frames with the iSight. Perfect! A big "thank you" to znek. My search has ended for now. I must get over to the lab.

  osascript MulleSightAsWebcam.applescript 
Did not try "Grab photos," but it also comes with source and I'm sure it works nicely. Oh yeah, I had though about a little command line program to search google for whatever. Here are some simple examples. It could not be easier, but do I really need %3A for the ":"?
  open http://www.google.com/search?q=schwehr
  open http://www.google.com/search?q=schwehr+kurt
  open http://www.google.com/search?q=site%3Aschwehr.org+google
  open http://www.google.com/search?q=site:schwehr.org+google

Posted by Kurt | Permalink

01.02.2005 09:04

nb phrase index

I like the idea of the index of phrases more and more. However, I think I left too many words in the index for it to be useful. At this point, it is highly unlikely that I will spend any more time on this until a much later date. This is the main reason for me to have a web log: I want to be able to go back and see what and how I did something. Plus, if as I look at peoples web logs, how do I know what they are interested in? Should I subscribe to their feed? Will I gain info about my interests? This is especially frustrating in my search for decent geology web logs. I wish there was a real slashdot for geology, geophysics, and planetary science. space.com and geology.about.com are just not the same! The rate that I learn new things from slashdot is so much higher (ignoring the comments which are mostly cruft). As such, it is time for a quick list of features that might be good to implement in the future:

  • Allow selection of the link to be an article number, the article name, article filename, or the article date field.
  • Separate web pages that sort alphanumerically and by most frequent. Most frequent will help people see what the blogger is writing about most often.
  • Better tmp file handling. Use /tmp instead of the local directory.
  • Allow caching of old articles so speed up big blogs.
  • Rewrite this with a real database like sqlite.
  • Make some GUI tools to help with keyword generation
  • Tool to create a list of proposed new words. Have a file of words definitely excluded.
  • Provide a kick start list of words to ignore.
  • Separate word lists: one for case sensitive - the other case insensitive
  • Handle interspersed line breaks and html coding.
  • Can ht://Dig do this kind of thing?
  • Create files that extract the web links in the log by position (time wise) in the log entries and by most often occurring.
  • What else would make a nice statistical summary?
  • How about a graph of how the top N words are connected by articles? An excuse to play with graphviz?

More Sumatra earthquake info: USGS first motion plot.

I made a quick demo file of the available smilies in nb. I sure you would all hate it if I over used the animated smilies like the role: :role:

Posted by Kurt | Permalink

01.01.2005 22:35

Phrase index

I have been working on my index of phrases for nanoblogger this evening. Thanks to Anne for being a sounding board for ideas. The version number may say 1.2, but this is still just a hack :)
  Source:  http://schwehr.org/software/nb-build-index-1.2.bash
  Index:   http://schwehr.org/blog/archives/phrase-index.html
  Phrases: http://schwehr.org/blog/data/phrases

I noticed that the RSS feed for my blog has drastically changed this evening when I switched to nanoblogger 3.1. You now get when whole entry in the feed rather than just the first paragraph. That is a nice bonus I hope. I definitely prefer having the option of reading the feed offline. The only bummer is that I got double on each article in NewsFire. That will go away as they role off the RSS feed file.

Posted by Kurt | Permalink

01.01.2005 18:24

Moods working?

This should be a smiley

Wow. It actually works. I have been trying to get these things to work for a long time! I copied moods and configure blog.conf like this

  cp -r /sw/share/nanoblogger/moods $BLOG_DIR

MOODS_URL="http://schwehr.org/blog/moods" AUTO_FORMAT="1"

Then it magically worked. Now I do not need to put in paragraph tags either.

I updated the nanoblogger.info file so that nb --manual now works if you have set BROWSER in ${BLOG_DIR}/blog.conf to something like open

Posted by Kurt | Permalink

01.01.2005 09:12

Remote Sensing Class

Just ran into this online class on remote sensing by Nicholas Short at Goddard. Looks like it might be very nice. I stumbled on it when looking at google images.


I found this useful list of smtp servers by ISP, which gives pacbell as mail.pacbell.net. It seems to work great.


Just discovered Knowing Knoppix. I have not had a chance to look at what is in this book, but might be good for those new to Linux.

While waiting for data to download, I tried out CocoaDialog. Very nice. I need to keep this handy to pop up dialog boxes from shell scripts.

I did a quick update of my cvs tree to 2005 and at the same time ran tidy on my Mac 10.3 install guide. The new version drops the micro number. What do you call the 3rd number after major.minor? Is micro right? Probably not.

I have been trying to think how to add more than just camera animation to the simpleview application in my density package. There are too many options. I could use an Inventor Engine in the IV file or maybe a VRML keyframe thing, but I can never remember how to do that stuff. It would be nice to have something like the camera flight path tool. Have time markers and set the world to that position. But how do I cleanly do this. The other idea is similar to Viz without the network interface. Have a simple file that specifies tuples something like this:

  TimeOffset NodeName ParameterToModify ParameterArgList ...

e.g. Have a lake level rise in Gusev Crater, Mars:

0.0 LakeLevel translation 0. 0. 0. 0.1 LakeLevel translation 0. 0. 0.1 0.2 LakeLevel translation 0. 0. 0.2 0.3 LakeLevel translation 0. 0. 0.3 0.4 LakeLevel translation 0. 0. 0.4

I am not sure. My main goal is something simple I can explain in a couple minutes that can handle at list a little bit of work with multiple moving objects. That is why I named simpleview as simple. I really do not want it to end up like Maya or blender. Those take a long time to learn. I would like to get to know those, but do not have the time right now. Currently I am leaning towards blender because it is open even though I do have access to a float license to Maya.

To go along with the MER stuff I have been looking at today, I found the first real interesting thing from RSS feeds on Mars. PBS has a presentation by Steve Squires Steve is not is normal energetic self and sounds a bit like he has a could which gives the talk a nicer pace. Definitely worth the 12 minutes!

Posted by Kurt | Permalink

01.01.2005 00:05

Happy New Year

Just in case you had not noticed, it is 2005. Happy New Year. I just noticed that they had a News Years event at the Hyatt hotel just a couple blocks from here. I should have walked over there or gone down to surf side to see the La Jolla cover fireworks.

I have been wanting a C++ serialization system for quite a while. Something like python's pickle or marshal. Maybe s11n is that tool?

Posted by Kurt | Permalink