02.28.2005 14:16

Oceanside Slide

Oceanside Slide pics by Liz Johnson on 6 Feb 2005.

Today the Sed Group is heading to La Jolla to view more recent slides.

Posted by Kurt | Permalink

02.27.2005 20:01

The super demag poster

I just finished up the first draft of a "super AF demag poster" generator. This little python program uses pysqlite to go into my Santa Barbara paleomag database and find all of the AF demagnetizations that Jason and I have been doing. It then uses Lisa's plotdmag to make a postscript plot of each one. Then a tangled web of netpbm and pil (Python Imaging Library) creates a png image of each plot. I then use pdflib_py to layout all the plots in a 3ft x 5ft poster form. Tomorrow I will know if this sucker plots out okay on the large format HP printer in the lab.

In the process, I wrote a type_pkg variant info file for pil_py so that I could have pil_py24. Sent that off to the package maintainer... Jeff the super packager. I really like the new variant mechanism. Makes life much easier for python in fink.

It has been a busy day! I did just over 4 hours in the lab this morning doing more AF demag. Without further delay, here is the code in all its wartiness.

Posted by Kurt | Permalink

02.27.2005 13:21

Opportunity/Spirit traverse maps

From the OSU crew:

Posted by Kurt | Permalink

02.26.2005 16:55

python sets

Wrote a quick little application that uses python sets. Very handy. What I want is a list of samples that are missing between different tables in my thesis database. There are a couple reasons for missing samples. First, they may not have been measured for all the critical properties. Maybe I have weighed them, but not measured an NRM on the magnetometer. The second reason is a type in the sample name. Sometimes you make a typo when entering a sample name. It happens. The earlier these are caught, the better! In this program, I load each tables sample name field into a set and then use symmetric_difference to find out what samples are in one or the other table, but not both. Here is the code...

Posted by Kurt | Permalink

02.26.2005 15:36

pdflib-py fink variants package

I did it! That was not too bad. I now have a python variants package for pdflib. I will be checking this into CVS within the next hour after I give it a lookover. Already, I caught that I forgot the BuildDepend on pdflib. Doh! I need a spare mac that I can wipe and reinstall a clean fink to test the build of each package. I threw a draft of the info file online:
Quick sed tip to cutting out lines from the middle of a file...
    wc -l foo.txt
    sed '5,45 d' foo.txt

Posted by Kurt | Permalink

02.26.2005 12:59

IBM Cloudscape and PDFLib lite

Alex suggested IBM's CloudScape/Derby database. Maybe I should consider Eclipse in the future too. It might actually stick around unlike many IDE's that I've used (except emacs).


Writing the Global Digital Elevation Model Viewer Application in Python: A Pythonic odyssey by Tom Babbitt


I am going to try to setup PDFLib python interface in fink. Let's see if I can do this in a couple hours...

Posted by Kurt | Permalink

02.25.2005 19:33

pdflib lite rocks for af demag

Panda is definitely out. The bus errors sent me off to try PDFlib Lite. It was not long before I had a sample pdf with test af demag plots and discovered that they show up nicely in Illustrator. Rockin! Here is my test code. Only the C API is in fink. It feels weird and clunky to go back to a C only API. Managing allocating and destroying objects is much easier in C++. I also want to see if I can get the python interface going.
// Using the c api
#include <stdlib.h>
#include <stdio.h>
#include <pdflib.h>
#include <iostream>
using namespace std;

int main (int argc,char *argv[]) { PDF *p; int font; if (!(p=PDF_new())) { cerr << "Could not create PDFlib obj\n"; return (EXIT_FAILURE); } PDF_TRY(p) { if (-1==PDF_open_file(p,"afdemag-poster.pdf")) {cerr<< "Error:"<<PDF_get_errmsg(p) << endl;return (EXIT_FAILURE);} PDF_set_info(p,"Creator","X-Core"); PDF_set_info(p,"Author","Kurt Schwehr"); PDF_set_info(p,"Title","BPSIO-04 AF Demag"); //PDF_begin_page(p,a4_width,a4_height); PDF_begin_page(p,2592,5184); font = PDF_load_font(p,"Helvetica-Bold",0,"host",""); PDF_setfont(p,font,42); PDF_set_text_pos(p,800,4900);

PDF_show(p,"BPSIO 2004 AF Demagnetization Curves");

int image; for (int x = 0; x<2300 ; x += 250) { for (int y = 0; y < 4500; y+=200) { cout << x << " " << y << endl; if (-1 == (image = PDF_load_image(p,"jpeg","1.jpg", 0, ""))) {cerr << "crap" << endl;} else { PDF_fit_image(p,image, x, y, "scale 0.4"); PDF_close_image(p,image); } } }


PDF_close(p); } PDF_CATCH(p) { cerr << "Doh!\n"; return(EXIT_FAILURE); }

PDF_delete(p); return(EXIT_SUCCESS); }
Here is the sample that generates. Be warned that Safari wants to download this as afdemag-poster.pdf.bz2. You will need to change the extension to .pdf.bz2 on the desktop for things to work right. Weird.

Posted by Kurt | Permalink

02.25.2005 13:37

panda and backups

Cooked up a quick script to put in the crontab for sorcerer. This generates a 400 MB compress tar.bz2 archive.
cd ~pmag
TAR=~/Desktop/rawdat-`date "+%C%y%m%d"`.tar
echo "Create a tar of RAWDAT"
tar cf $TAR RAWDAT
echo "Compressing the tar"
cd ~/Desktop
bzip2 -9 $TAR
echo "Done.  Pleae write the tar on your desktop to a CD and take it home!"
To work with the bziped archive:
  View:  tar tfjvv rawdat-20050225.tar.bz2
  Extract: tar xfjvv rawdat-20050225.tar.bz2
I started working on a fink info file for panda. This one does not yet conform to the shared lib policy. I will do that some other day. It does fix the libtool symbolic links that the 0.5.4 distribution has. My draft fink packages. Basically, I want to autogenerate a large poster of AF Demagnetization plots for my core samples.

Update on panda: This does not seem to be the most stable of libraries. I do not have time to fight with a program that is so touch with what it can and cannot read. Even one of the demo programs gives a bus error. My set program fails when I give it my tif versus working with theirs.
Welcome to the Panda 0.5 sample application...
TIFFClientOpen: One of the client procedures are NULL pointer.
Bus error

Posted by Kurt | Permalink

02.25.2005 08:12

New Firefox - 1.0.1, Mayer in the news

If you are using Firefox, go download the new version right now!

There is a new of numarray-py. Woohoo! Now I have to start using it and start seeing what there is available for eigenvalues and random number generation.

Technological Revolutions In Sensors, Robotics, And Telecommunications Allow New Views Of Ocean - Larry Mayer of UNH in ScienceDaily.

A non JPL publications... Mars Exploration Rover Technical Data by Ruper Scammell. Who is he?

ssh tunneling hint

xtermcontrol hint - I put xtermcontrol in fink :)

Posted by Kurt | Permalink

02.24.2005 13:25

gnuplot index feature

Today is the first time that I have played with the gnuplot index feature. This helps to reduce the number of loose data files in a directory. Here is a quick sample from the section boundary markers in my Santa Barbara data. The key thing is to have TWO blank lines after each group.

cat << EOF > dec-sec-bound.dat
# There is no core 0
0 0

# Core num: 1 37.5 0 37.5 360

# Core num: 2 - no values 0 0

# Core num: 3 39.5 0 39.5 360

# Core num: 4 - no values 0 0

# Core num: 5 44.0 0 44.0 360

# Core num: 6 - no values 0 0 EOF
Then here are the gnuplot commands to print this. Core 0 is there so core 1 is refered to as the 1 index.
  set xrange [-10:370]
  set yrange [0:200] reverse # to plot with zero on top
  plot 'dec-sec-boundary.dat' index 1 using 2:1 with lines \
       , 'dec-sec-boundary.dat' index 3 using 2:1 with lines

Posted by Kurt | Permalink

02.24.2005 12:01

ODP database, GroupDAV

I hope Apple adds this kind of thing to iCal and Address Book. GroupDAV. I haven't looked closely, but this is the right direction. It is time that the world catch up to what Solaris could do in the mid 90's.

I got a really nice response from David Flacker at IODP. When I get back to the odp data in a day or two, I will try this out. I am currently trying to rotate my cores bases on the nrm direction in a way that is easy to tweak. I am aiming for a zero centered -180 to 180 graph.

Yes, you can use tools like wget:

wget http://iodp.tamu.edu/janusweb/physprops/msldat.cgi?leg=146&site=893&hole=a&outputtype=5&depthmin=5&depthmax=10

There is no formal publication of the input parameters accepted, but the variable names and the variable range and form can generally be gleaned from the "name" parameters found when viewing the source of our .shtml pages.

outputtype=5 is equivalent to the Excel Integration delimiter preference on the Report Options popup and should result in fairly clean, non-html comma-separated vector files

depthmax=10 when only upper limit is specified depths <= limit are returned, vice versa for depthmin parameter; when both a depthmin and depthmax are specified, values between and including the range end-points are returned ( depthmin <= x <= depthmax)

All of the reports you can run are browseable in the janusweb URL tree: http://iodp.tamu.edu/janusweb or are otherwise summarized in the links here: http://iodp.tamu.edu/janusweb/links/links_all.shtml


Apologies for the aggravations. While it serves the bulk of data request needs without additional human intervention, it is very possible to generate queries that will not run to completion within the timeouts of the typical web-browser. The query system in place generates reports directly from the underlying database. We do not provide any kind of pre-query analysis to offer automated estimates on how long a query will take. The amount of time is highly variable with data type, data model structure, the exact specifics of the query inputs, and the amount of data being searched/sorted.

Best practice suggestions:

Look at the Database Overview page (http://iodp.tamu.edu/janusweb/general/dbtable.cgi) to see what data is available on-line.

Be as specific as possible about what you want: e.g. if you know the site, specify the expedition with it--often the database can bring a data index into play that result in orders of magnitude improvements on turn-around time. The problem from the server perspective is not how much data is returned, but how much data has to be sifted through to figure out what should be returned.

Avoid doing queries by depth ranges where possible. Depth is a calculated value, not an absolute stored value. A lot of overhead is incurred by the database server in the way these depth filters are currently implemented. If you need it, though, it will work, just set your timeout limits for wget high (quarters of hours rather than minutes).

Experience notes:

The CGI-based query system does use cookies to remember expedition, site, hole, and report options settings throughout a browser session. The values remembered are prefilled in forms. Enable cookies for the iodp.tamu.edu site if that fits your usage patterns at all.

The query system uses popups for the Report Options dialog and for query result errors and warnings. E.g. default query behaviour when no data is found is to generate a popup window and a page-back event. When popups are disabled by the browser, all you see is the page-back event--no time to look at the empty result page.

Regards, Fackler

Posted by Kurt | Permalink

02.23.2005 09:16

Posted by Kurt | Permalink

02.22.2005 20:51

XCore in the FSF directory!

I never noticed this before, but XCore is in the FSF software directory!
And hey, this is a picture of me working on Devon Island in the Haughton Crater. This is the first I have seen this one. The picture is by Peter Essick who shot pictures for a National Geographic article.
  MarsOnEarth img 931

Posted by Kurt | Permalink

02.22.2005 20:17

You are now a Ninja

Posted by Kurt | Permalink

02.22.2005 19:00

ODP Janus Web

I managed to get the 893a MSL (mag susc) data a week or two ago, but had trouble getting any other data.

http://iodp.tamu.edu/janusweb/physprops/msldat.cgi with site 893 hole a gives:
Error - 504
Gateway Timeout: (80)
Reason: Timed out while waiting for data from the remote host.
Things like http://iodp.tamu.edu/janusweb/paleomag/tensdat.cgi with site 893 hole A just flash up something and jump back to the page. When I hit forward, there is no data on the page. Is there no data available or am I doing something wrong? Is there a way to tell what is actually available for 893A and B?

Also, is there a way to put a query in a URL? I'd like to be able to do something like:
  wget http://iodp.tamu.edu/janusweb/paleomag/tensdat.cgi?site=893&hole=A
You get the idea. Then it would be easy to snag the data from any shell or python script.

Posted by Kurt | Permalink

02.22.2005 14:42

Landslides at SIO

Today, we had what turned into a mini lab meeting for part of the sed group. Really, it was just who we ran into in the hallway. But with all the rain, landslides are all over San Diego County. So we had to go out and see what is happening with the NOAA Fisheries building at the other end of the SIO campus. We walked down the beach and found what looked like an extremely fresh slide. We grabbed a life guard and found out that it happened last night. The hill came down on dike rock! Some bits and pieces of metal came down too. The fresh debris exposed some really great sedimentary geology. I got 70 pictures of it before the ocean erodes away all this great stuff. Check it out!
  Galerie photo show of Slides at SIO

Posted by Kurt | Permalink

02.22.2005 12:51

Lisa's and Roi's computers are back to life

Woo hoo! Much progress in the land of working computers. Here are the fixes:

Roi's flailing gcc... Go into /Library/Receipts and remove each receipt for the developer tools one by one and install the package again from the developer dmg.

Lisa's stubborn X11SDK... Do NOT enable Virex. Virus software should not run all the time on Mac OSX. It just causes problems. Scanning programs are fine. One 10.2, Norton AntiVirus destroyed several machines to the point of having to reinstall everything. Once Virex was disabled from active scanning, X11SDK installed right away on Lisa's G5.

Posted by Kurt | Permalink

02.22.2005 07:03


Doh! Got a bug report that weka failed to build on 10.3.8 with fink. I confirmed it:

init_tests: [mkdir] Created dir: /sw/src/weka-3.4.3-1/weka-3.4.3/build/testcases [mkdir] Created dir: /sw/src/weka-3.4.3-1/weka-3.4.3/reports

init_compile: [mkdir] Created dir: /sw/src/weka-3.4.3-1/weka-3.4.3/build/classes

compile: [javac] Compiling 635 source files to /sw/src/weka-3.4.3-1/weka-3.4.3/build/ classes [javac] Note: Some input files use or override a deprecated API. [javac] Note: Recompile with -deprecation for details. [copy] Copying 167 files to /sw/src/weka-3.4.3-1/weka-3.4.3/build/classes [rmic] RMI Compiling 1 class to /sw/src/weka-3.4.3-1/weka-3.4.3/build/class es


run_tests: [junit] Class not found "weka.AllTests"

BUILD FAILED /sw/src/weka-3.4.3-1/weka-3.4.3/weka/build.xml:139: Java returned: 1

Microsoft is trying to patent a pointer comparison operator. You have to be kidding.

Posted by Kurt | Permalink

02.21.2005 18:43


If you get bored with reading my blog, perhaps you should gizoogle it. WARNING: strong language - Blame the language on gizoogle and Snoop Dog, not me.

Posted by Kurt | Permalink

02.21.2005 17:18

Marine sediments magnetic susceptibility paper

Rey et al have a very interesting paper in the 25 March 2005 Marine Geology:
Early diagenesis of magnetic minerals in marine transitional
environments: geochemical signatures of hydrodynamic forcing.
I will be reading this one right after Roi's Tarduno02 paper.

Posted by Kurt | Permalink

02.21.2005 17:12

Full python database creation

Just finished up the full database creation python scripts. The bash script was getting to be 5-10 minutes long and I have a long ways to go before I am done. Right now I am at 7-8 seconds with the python script. There were just so many calls to awk, cut, and sqlite. Now with the python code, the script is much simpler and tons faster. It uses the string split method extensively and I have some classes that handle converting k15 data into hext and s formats. Check out the two scripts:
ams.py is my module of helper functions that includes the classes Hext, S_Matrix, DepthLookup and SampleName. I really like SampleName. I give it a string from one of my sample names and it pulls apart all the fields. It also has a __str__ method so that if I turn it into a string, it goes back to the original.

build_db.py is a little more complicated. It reads each of the text data files off of disk, creates the table, and inserts all the values. Having really string processing in python makes the core location and core section tables so much simpler and easy to read.

Posted by Kurt | Permalink

02.21.2005 12:48

Apple's X11SDK installer blows chunks

With help from Martin Costabel I started replacing missing files on Lisa's G5 that the X11SDK did not install. I am starting to thing that Apple's package manager is a piece of junk beyond just basic usage. There is (easy) way to query, uninstall, reinstall, verify, packages especially when logged in remotely. On Lisa's machine with 11SDK, it missed all but a couple headers, all the .a's, and /etc/font... and that is what I have found so far. I finally just copied these from another machine. That got a lot back on track, but not all.

Posted by Kurt | Permalink

02.21.2005 08:57

mbsystem 5.0.6 is in fink

Last night mbsytem-5.0.6 was released to the wild. I just finished putting it in the fink unstable tree. This now forces you to upgrade to gmt 4.0. You should see a speedup of the mb edit commands since Dave and Dale enabled optimization with the compiler. Please email me about any troubles from the fink install.

Posted by Kurt | Permalink

02.21.2005 07:26

MBSystem 5.0.6 released

Just found an email in my inbox from Dave Caress talking about the new features in the just released MB-System 5.0.6. I am about to set forth on the task of getting it into fink. In the meantime here is the email from Dave:

  One new feature that is not strongly highlighted in the release
  notice is the inclusion of MBgrdviz, an OpenGL based tool that
  provides some simple visualization capabilities for GMT grid
  files. Since MBgrdviz is not yet well documented, this note is
  to provide some extra hints on building and using this new
  program. MBgrdviz is essentially a demo and test program for a
  visualization library called MBview that will be central to
  several new MB-System tools now under development. These new
  tools will include interactive project management, improved
  bathymetry editing, and patch test tools in the relatively near
  Since the visualization API depends on OpenGL as well as X11
  and Motif, the install_makefiles script now includes an
  OPENGLLIBS parameter. If you do not define (or comment out)
  OPENGLLIBS, then install_makefiles will construct the Makefiles
  structure so that MBview and MBgrdviz are not built. In this
  case, the rest of the MB-System tools will build and install as
  before. Thus, to build MBgrdviz, you must define the OPENGLLIBS
  parameter properly for your system. The comments in the
  install_makefiles script include appropriate settings for
  MacOSX and IRIX.
  Since there is no real documentation yet, early users of
  MBgrdviz will be determining if the interface is sufficiently
  obvious to be used without documentation. Feedback on
  ease-of-use will be appreciated.
  Here are some hints for using MBgrdviz:
  1. MBgrdviz can display up to ten instances, each potentially
  displaying different grids. The program will work with grids
  that are in geographic (longitude latitude) coordinates or in
  projected coordinates systems (e.g. UTM eastings and
  northings). However, the program will only recognize projected
  coordinate system grids that have been made by mbgrid, as
  mbgrid embeds the projection used in the grid file comments.
  2. Each grid is displayed in an MBview window with its own view
  control widgets.
  3. The basic controls for display type are in the View
  menu. Options include 2D or 3D display, setting the primary
  data displayed in color (e.g. the topography in the primary
  grid, the topographic slope, or an overly grid), shading style
  (none, shading by illumination, shading by slope, shading using
  an overlay grid such as backscatter), and turning topography
  contours on or off.
  4. Interaction with the data is controlled by a number of mouse
  modes; the mouse mode is set through the options in the Mouse
  menu. Options include Pan & Zoom, Rotate Model, Rotate View,
  Shading, Pick Area, Edit Sites, Edit Routes, and Pick Nav. If
  the MBview window is resized large enough, a set of buttons
  appears on the lower left that duplicates the Mouse menu
  options, making it easier to switch modes. The top status
  display box on the left always displays the mouse button
  behaviour for the current mouse mode. So, in Pan & Zoom mode,
  this box shows that the left button does pick operations, the
  middle button does pan operations, and the right button does
  zoom operations.
  5. The Rotate Model and Rotate View mouse modes are only
  available when the data is displayed in 3D.
  6. The basic pick operation will pick a single point on the
  topography, display a red X at that location, and show
  information about the pick in the middle status box on the
  left. If the user drags the cursor while the right button is
  down, two points will be picked (the point where the button was
  pressed and the point where the button was released), and the
  two points will be displayed as red X's connected by a line
  segment between the points. In 3D views this line will be
  draped along the topography. The status window will show the
  positions of both picks and the distance between them.
  7. The Pick Area mode allows regions and areas to be picked
  interactively. In MBview, a region is rectangular with sides
  that are oriented N-S and E-W in the grid coordinates, is
  picked with the left button in Pick Area mode, and is displayed
  as a yellow box. An area is an arbitrarily oriented rectangle
  defined by dragging the middle button in Pick Area mode, and is
  also displayed in yellow. The aspect ratio, and therefore the
  width, of an active area can be altered by dragging the right
  button up or down in Pick Area mode. In 3D views both region
  and area bounds are draped on the topography.
  8. MBgrdviz also supports display of Sites, Routes, and
  Navigation.  Sites and Routes may be created and altered
  interactively in MBgrdviz, and can also be exported to files
  and imported from files. Site and Route files created by
  MBgrdviz include comments documenting the format. The
  file-naming convention for Site files uses an *.ste suffix, and
  an *.rte suffix for route files. Navigation is imported by
  reading datalists of swath files supported by MB-System.
  9. The Edit Site mouse mode allows the user to select and move
  sites with the left button, create new sites using the middle
  button, and delete already selected sites using the right
  10. The Edit Route mouse mode allows the user to select and
  move route waypoints using the left button, create new route
  waypoints using the middle button, and delete selected route
  waypoints using the right button.
  11. Navigation may imported in two fashions from the Action
  menu. The Open Navigation menu item allows a datalist to be
  selected and opened; the program then reads the navigation in
  each swath file referenced by the datalist and displays it as a
  simple black navigation line. The Open Swath menu item also
  reads navigation data from a datalist, but it causes the
  program to also read and store the acrosstrack bounds of the
  swath bathymetry associated with each navigation point, with
  implications covered in point (12). In 3D views, the navigation
  can be displayed using it's real vertical position (the
  navigation tracks will usually display somewhere above the
  topography), or it can displayed draped along the topography,
  or both.
  12. The Pick Nav mouse mode allows the user to pick individual
  nav points and to select and deselect arbitrary sections of the
  navigation.  If the navigation has been imported using the Open
  Navigation menu item, then the selected navigation will simply
  be displayed in red, and information about the selected
  navigation will appear in the middle status window on the
  left. If the navigation has been imported using the Open Swath
  menu item, then the selected navigation will also have the
  bounds of the associated swath bathymetry displayed in
  yellow. In 3D views, the swath bathymetry bounds will be draped
  on the topography.
  13. MBgrdviz displays the topography in three
  resolutions. Whenever an operation modifies the display, the
  data will first be redrawn in low resolution, and then in a
  high resolution. During operations involving mouse dragging
  (e.g. panning, rotating, zooming, or two-point picking), the
  display will be refreshed rapidly in low resolution, with the
  high resolution display appearing once the operation is ended
  by releasing the mouse. For large grids, the high resolution
  display may not be at the full resolution of the grid. A full
  resolution display can be obtained by clicking the Full button
  below the middle status box on the left. The lower status
  window on the left shows the current status of drawing the
  display - typically users only see a "Done" status except
  during full resolution drawing.
  14. The Reset button below the middle status box on the left
  resets the display to the original scale and orientation.
  15. The Clear button below the middle status box on the left
  clears any picks and selections.
  There are some known bugs in this first release version of
  1. The Rotate View function simply doesn't work right. I don't
  have the order of the OpenGL rotations set properly so that
  Rotate View operations actually rotate the 3D display around
  the user's viewpoint.
  2. There is a memory leak that produces error messages when
  MBview instances are dismissed.
  3. In 3D displays the program tries to figure out what part of
  the grid is actually in the display window so that it can skip
  drawing areas that cannot be seen. MBgrdviz also then adjusts
  the resolution used during low and high resolution drawing
  accordingly. This allows the effective resolution to increase
  as one zooms in on a particular area.  However, the current
  algorithm frequently underestimates the viewable area, with the
  result that visible parts of the grid are left undrawn.

Posted by Kurt | Permalink

02.20.2005 20:43

Suse Linux makes EAL4+

So Mac OSX makes level 3, so Suse Linux goes and makes CC EAL 4+. For some reason, I have never tried Suse.

Test-Driven Development in Python


I created the original twisted fink packages, but guess what?!?! I don't know how to use twisted. I know NOTHING! Here is a twisted intro so maybe I will now know something.

Posted by Kurt | Permalink

02.20.2005 06:11


RealClimate.org was mentioned in this months Scientific American. Here is what the site is in their own words:
  RealClimate is a commentary site on climate science by working
  climate scientists for the interested public and journalists. We aim
  to provide a quick response to developing stories and provide the
  context sometimes missing in mainstream commentary. The discussion
  here is restricted to scientific topics and will not get involved in
  any political or economic implications of the science.

pybliographer and bibtex2html are both in fink, but pybliographer does not currently build. Maybe I can use bibtex2html to do a better job of my papers web page. num-utils also looks cool. Maybe after I install gnome-python-py23 I can get pybliographer to build. I finally created a temporary gnome-python-py24 and that got me to a pybliographer that built and installed. I have not tested it.

Posted by Kurt | Permalink

02.19.2005 12:06

k15_hext from python getting better

If I had more time, it sounds like what I need is Pexpect. Bummer that it is not in fink.

I have not been able to get k15_hext to hang out and let me talk to it. I figured that it would be easy to pass it 4 lines at a time, but no. There is no time to really figure out how to do this, so I am on to the next thing. Here is the code that works for cat, but not k15_hext:
#!/usr/bin/env python
import os,string
fin.write("bp04-6gw-s1-141.2	schwehr	1/3/05	4:33 PM	 1716	 -4.75\n")
fin.write("9.3800E+1	9.2950E+1	9.3550E+1	9.3550E+1	9.2950E+1\n")
fin.write("8.9150E+1	8.9100E+1	9.2900E+1	8.9150E+1	8.8900E+1\n")
fin.write("8.9950E+1	8.8750E+1	8.5100E+1	8.9950E+1	8.8800E+1\n")
fin.flush()  # <---- this works for cat, but not k15_hext!!!????
print fout.readline(),
print fout.readline(),
print fout.readline(),
print fout.readline(),
Here is what I have so far which starts and kills a k15_hext instance each time. Not efficient, but it works.
#!/usr/bin/env python
import popen2,string

class K15HextEngine: """ Doc for K15. Uses the k15_hext from Lisa's pmag to convert k15 Kappabridge measurements to hext statistics. """ def __init__(self): self.totCount=0 print "finished init"

def getHext (self, line1, line2, line3, line4): """Unable to keep a k15_hext open and write to it over time. Why not?""" # FIX: check if new line at end, if not, add one blah=2 fin,fout=popen2.popen2("k15_hext") fout.write(line1) fout.write(line2) fout.write(line3) fout.write(line4) fout.close() hline1 = fin.readline() hline2 = fin.readline() hline3 = fin.readline() hline4 = fin.readline() hline5 = fin.readline() self.totCount += 1 print "total = ",self.totCount return hline1,hline2,hline3,hline4,hline5

# Test code if __name__ == '__main__': print K15HextEngine.__doc__ k15_hext = K15HextEngine()

l1 = "bp04-6gw-s1-141.2 schwehr 1/3/05 4:33 PM 1716 -4.75\n" l2 = "9.3800E+1 9.2950E+1 9.3550E+1 9.3550E+1 9.2950E+1\n" l3 = "8.9150E+1 8.9100E+1 9.2900E+1 8.9150E+1 8.8900E+1\n" l4 = "8.9950E+1 8.8750E+1 8.5100E+1 8.9950E+1 8.8800E+1\n" h1,h2,h3,h4,h5 = k15_hext.getHext(l1,l2,l3,l4) h1,h2,h3,h4,h5 = k15_hext.getHext(l1,l2,l3,l4) h1,h2,h3,h4,h5 = k15_hext.getHext(l1,l2,l3,l4) print " RESULTS: " print h1,h2,h3,h4 del k15_hext
Now I just need to make this a nice module that lets me get some sort of hext object.

Posted by Kurt | Permalink

02.19.2005 10:03

Goodbye NewsFire, hello NetNewsWire Lite

My beta of NewsFire expired this week. That was enough motivation to really go start using NetNewsWire Lite. I am starting to really like this new world. Not quite as clean and simple as NewsFire, but it does have things like letting you know when it is checking each news feed. I still have not played with the Firefox RSS capabilities which might be pretty cool.

Posted by Kurt | Permalink

02.19.2005 09:43

Python k15

Onward with the python k15 program. Here is my first working example. After this, I need to turn it into a class, so I can have an object that takes k15 strings and spits back the k15_hext stats. The other thing it has to handle is that after 100 samples, it has to close the k15_hext process and start a new one since that is the maximum number of samples setup in that fortran program. Here is my working example:
#!/usr/bin/env python
import popen2,string
fout.write("bp04-6gw-s1-141.2	schwehr	1/3/05	4:33 PM	 1716	 -4.75\n")
fout.write("9.3800E+1	9.2950E+1	9.3550E+1	9.3550E+1	9.2950E+1\n")
fout.write("8.9150E+1	8.9100E+1	9.2900E+1	8.9150E+1	8.8900E+1\n")
fout.write("8.9950E+1	8.8750E+1	8.5100E+1	8.9950E+1	8.8800E+1\n")
print fin.readline(),
print fin.readline(),
print fin.readline(),
print fin.readline(),
Running ./k15.py returns:
   bp04-6gw-s1-141.2    bulk susceptibility =   90.5700073
  F =      2486.69  F12 =        63.74  F23 =      4088.38
  0.34519  156.2    3.4    0.8   10.3   86.0    7.4  246.3    2.3
  0.34164  246.3    2.3    0.9   10.3   86.0    7.4  156.2    3.4

Posted by Kurt | Permalink

02.19.2005 09:31

Setting up webalizer

This is take two on setting up webalizer and I did it this time where it is not a problem to give all the details of what I did. So here is what I did! This is using the Apple supplied Apache web server (not the fink Apache - that would use different paths to all the files).
  fink install webalizer
  cd /Library/WebServer/Documents
  mkdir webalizer
  export EDITOR=emacs
  crontab -e
Now when crontab pops up emacs to edit your crontab file, enter this, but put everything starting with "30" and ending with /dev/null all on one line:
30 2 * * * /sw/bin/webalizer -p -N 2 -D \
/Library/WebServer/Documents/webalizer/dnscache -n sorcerer.ucsd.edu \
-o /Library/WebServer/Documents/webalizer \
/private/var/log/httpd/access_log > /dev/null
You can change -n sorcerer.ucsd.edu to your websever. You can kick things off right away by running webalizer with those options. Otherwise you will have to wait until after 2:30 AM for this cron job to get run. To see the results, take a look at:
I also added a robots.txt to the webserver to tell microsoft to bugger off.
User-agent: msnbot
Disallow: /

Posted by Kurt | Permalink

02.18.2005 20:00

NASA Mars life update from MSNBC

Check out Alan Boyle's Cosmic Log for more info on the mars life craziness. Thanks to Kiley for the link!

Posted by Kurt | Permalink

02.18.2005 19:44

NASA Statement on False Claim of Evidence of Life on Mars

From the NASA HQ press release:
Dolores Beasley/Gretchen Cook-Anderson 
Headquarters, Washington 

Feb. 18, 2005

RELEASE: 05-052

NASA Statement on False Claim of Evidence of Life on Mars

News reports on February 16, 2005, that NASA scientists from Ames Research Center, Moffett Field, Calif., have found strong evidence that life may exist on Mars are incorrect.

NASA does not have any observational data from any current Mars missions that supports this claim. The work by the scientists mentioned in the reports cannot be used to directly infer anything about life on Mars, but may help formulate the strategy for how to search for martian life. Their research concerns extreme environments on Earth as analogs of possible environments on Mars. No research paper has been submitted by them to any scientific journal asserting martian life.
It was nice to know that I was right in warning that this was not a good thing to cover. Now I can stop worrying about ideas of how methane variations could be explained by non-biological means. I was starting to think about methane hydrates either in the near surface in the non-polar regions or on the ice in the polar regions. I do not know that the stability fields look like at the low pressures of Mars, but it might be possible to have methane hydrates in the cold parts of the planets, no? Anyone else want to do the stability calculations? I am too lazy.

On another note, I tried to fix Roi's laptop today without success. We removed the /Library/Receipt file for DeveloperTools.pkg and were able to install the package again, but got the same ncurses bomb. Then we tried gcc3.3.pkg, but the installer said that we could not install gcc3.3 on that machine. Strange! GUI's that don't tell you what they really are doing or thinking drive me nuts.

Today I measured AMS for core 2GW, so the main AMS data has all been collected. I need to fixup my database creation code and start crunching and evaluating this stuff!

CrappyApps - Documents what is wrong with lots of software

Lisa showed me an AMAZING program today that one of her sons discovered: Processing 68 The demos are amazing and the code is C-like yet fantastically simple.
Processing is a programming language and environment built for the
electronic arts and visual design communities. It is created to teach
fundamentals of computer programming within a visual context and to
serve as a software sketchbook. It is used by students, artists,
designers, architects, and researchers for learning, prototyping, and

Posted by Kurt | Permalink

02.18.2005 08:38

Possible X11 solution

Need to try this:
Q9.13: I'm still having problems with X11 and Fink.

A: If the hints in the Fink tries to install XFree86 or X.org or X11 and upgrade from 10.2 entries don't help, or aren't applicable to your situation, you may need to flush out your X11 installation and remove any old placeholders and partially/fully installed X11-related packages:

sudo dpkg -r --force-all system-xfree86 system-xfree86-42 system-xfree86-43 \ xorg xorg-shlibs xfree86 xfree86-shlibs \ xfree86-base xfree86-base-shlibs xfree86-rootless xfree86-rootless-shlibs \ xfree86-base-threaded xfree86-base-threaded-shlibs \ xfree86-rootless-threaded xfree86-rootless-threaded-shlibs rm -rf /Library/Receipts/X11SDK.pkg /Library/Receipts/X11User.pkg fink selfupdate; fink index

(the first line may give you warnings about trying to remove nonexistent packages). Then, reinstall Apple's X11 and the X11SDK.

Posted by Kurt | Permalink

02.18.2005 08:08

OSX Common Criteria Cert

OSX 10.3.6 makes EAL 3. Not too exciting for my day-to-day use of OSX, but this has implications for working on things like spacecraft operations.

Posted by Kurt | Permalink

02.18.2005 07:21

fink x11 troubles

I have been looking at how to pass text via stdin to programs called from python. os.system can not do this. Looks like os.popen, os.popen2, or os.popen3. I have yet to find a nice working example on the web. Maybe one of these modules would do the trick: pipes or commands. Hey wait, I found something that works!
#!/usr/bin/env python
import popen2,string
print fin.readline()
print fin.readline()

Oceans on Mars: LPI 1741

More random notes for fink problems. Lisa has the forever disappearing Apple X11. Perhaps this command will help debug the problem.
  fink list system-xfree86
  fink-virtual-pkgs --debug
Here is what "fink-virtual-pkgs --debug 2&>1 foo" gives on a happy machine:
Package: broken-gcc
Status: purge ok not-installed
Version: 3.3-1
builddependsonly: true
description: [virtual package representing a broken gcc compiler]
homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: cctools Status: install ok installed Version: 498-1 builddependsonly: true description: [virtual package representing the developer tools] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: cctools-single-module Status: install ok installed Version: 1.0-1 builddependsonly: true description: [virtual package, your dev tools support -single_module] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: darwin Status: install ok installed Version: 7.8.0-1 description: [virtual package representing the kernel] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: gcc3.1 Status: install ok installed Version: 3.1-1 builddependsonly: true description: [virtual package representing the gcc 3.1 compiler] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: gcc3.3 Status: install ok installed Version: 3.3-1 builddependsonly: true description: [virtual package representing the gcc 3.3 compiler] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: gimp-print-shlibs Status: install ok installed Version: 4.2.5-1 description: [virtual package representing Apple's install of Gimp Print] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: macosx Status: install ok installed Version: 10.3.8-1 description: [virtual package representing the system] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-java Status: install ok installed Version: 1.4.2-1 description: [virtual package representing Java 1.4.2]

Package: system-java-dev Status: install ok installed Version: 1.4.2-1 description: [virtual package representing Java SDK 1.4.2]

Package: system-java12 Status: purge ok not-installed Version: 1.2-1 provides: system-java description: [virtual package representing Java 1.2] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-java13 Status: install ok installed Version: 1.3.1-1 provides: system-java description: [virtual package representing Java 1.3.1] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-java13-dev Status: install ok installed Version: 1.3.1-1 description: [virtual package representing Java 1.3.1 development headers] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-java14 Status: install ok installed Version: 1.4.2-1 provides: system-java, jdbc, jdbc2, jdbc3, jdbc-optional description: [virtual package representing Java 1.4.2] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-java14-dev Status: install ok installed Version: 1.4.2-1 description: [virtual package representing Java 1.4.2 development headers] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-java15 Status: purge ok not-installed Version: 1.5-1 provides: system-java, jdbc, jdbc2, jdbc3, jdbc-optional description: [virtual package representing Java 1.5] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-java3d Status: purge ok not-installed Version: 0-0 description: [virtual package representing Java3D] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-javaai Status: purge ok not-installed Version: 0-0 description: [virtual package representing Java Advanced Imaging] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-perl Status: install ok installed Version: 5.8.1-1 provides: perl581-core, system-perl581, attribute-handlers-pm581, cgi-pm581, digest-md5-pm581, file-spec-pm581, file-temp-pm581, filter-simple-pm581, filter-util-pm581, getopt-long-pm581, i18n-langtags-pm581, libnet-pm581, locale-maketext-pm581, memoize-pm581, mime-base64-pm581, scalar-list-utils-pm581, test-harness-pm581, test-simple-pm581, time-hires-pm581 description: [virtual package representing perl] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-xfree86 Status: install ok installed Version: 2:4.3-2 provides: xserver, x11, libgl, xft2, fontconfig1, rman, xfree86-base-threaded description: [placeholder for user installed x11] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-xfree86-dev Status: install ok installed Version: 2:4.3-2 builddependsonly: true provides: x11-dev, libgl-dev, xft2-dev, fontconfig1-dev description: [placeholder for user installed x11 development tools] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-xfree86-shlibs Status: install ok installed Version: 2:4.3-2 provides: x11-shlibs, libgl-shlibs, xft1-shlibs, xft2-shlibs, fontconfig1-shlibs, xfree86-base-threaded-shlibs description: [placeholder for user installed x11 shared libraries] homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage
Here is the cranky machine:
fink list system-xfree86

i system-xfree86 2:4.3-2 [placeholder for user installed x11] i system-xfree... 2:4.3-2 [placeholder for user installed x11 devel... i system-xfree... 2:4.3-2 [placeholder for user installed x11 share...

fink-virtual-pkgs --debug


- found gimp-print-shlibs 4.2.5-1 - checking for library libX11.6.dylib... found in /usr/X11R6/lib - checking for X servers... Xquartz - system-xfree86-shlibs provides x11-shlibs - found an X server, system-xfree86 provides xserver and x11 - checking for header X11/Xlib.h... missing - checking for library libGL.1.dylib... found in /usr/X11R6/lib - system-xfree86-shlibs provides libgl-shlibs - system-xfree86 provides libgl - checking for header GL/gl.h... missing - checking for library libXft.1.dylib... found in /usr/X11R6/lib - system-xfree86-shlibs provides xft1-shlibs - checking for library libXft.2.dylib... found in /usr/X11R6/lib - system-xfree86-shlibs provides xft2-shlibs - checking for library libXft.dylib... found in /usr/X11R6/lib - libXft points to Xft2 - system-xfree86-dev provides xft2-dev - system-xfree86 provides xft2 - checking for library libfontconfig.1.dylib... found in /usr/X11R6/lib - system-xfree86-shlibs provides fontconfig1-shlibs - checking for library libfontconfig.dylib... found in /usr/X11R6/lib - checking for header fontconfig/fontconfig.h... missing - checking for rman... found, system-xfree86 provides rman - checking for threaded libXt... threaded - system-xfree86-shlibs provides xfree86-base-threaded-shlibs - system-xfree86 provides xfree86-base-threaded


Package: system-xfree86 Status: install ok installed Version: 2:4.3-2 description: [placeholder for user installed x11] provides: xserver, x11, libgl, xft2, rman, xfree86-base-threaded homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage

Package: system-xfree86-dev Status: install ok installed Version: 2:4.3-2 description: [placeholder for user installed x11 development tools] provides: xft2-dev homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage builddependsonly: true

Package: system-xfree86-shlibs Status: install ok installed Version: 2:4.3-2 description: [placeholder for user installed x11 shared libraries] provides: x11-shlibs, libgl-shlibs, xft1-shlibs, xft2-shlibs, fontconfig1-shlibs , xfree86-base-threaded-shlibs homepage: http://fink.sourceforge.net/faq/usage-general.php#virtpackage
So what the heck is wrong??? This looks okay to me. Now I try installing something from X11 and get:
fink install g3data

fink needs help picking an alternative to satisfy a virtual dependency. The candidates:

(1) xfree86: Free X11 implementation for Darwin and Mac OS X (2) xorg: Free X11 implementation for Darwin and Mac OS X
Trying a fink selfupdate (which is rsync). Got a new version of fink. Damn! The 2.5GHz G5 with 4GB of ram is errr.... FAST!!! Then did a scanpackages. Still no go! Grr. fink index. Here is the build depend list for g3data...

BuildDepends: pkgconfig, gtk+2 (>= 2.2-1), gtk+2-dev (>= 2.2-1), imlib, atk1, pango1-dev, glib2, glib2-dev, x11-dev

fink list xfree86 xorg x11- Information about 4669 packages read in 1 seconds. i system-xfree86 2:4.3-2 placeholder for user installed x11] i system-xfree86-dev 2:4.3-2 [placeholder for user installed x11 development tools] i system-xfree86-shlibs 2:4.3-2 [placeholder for user installed x11 shared libraries] x11-app 1.0-2 Apple's X11.app X server x11-dev [virtual package] x11-shlibs [virtual package] x11-ssh-askpass X11-based SSH passphrase dialog xfree86 4.4.0-14 Free X11 implementation for Darwin and Mac OS X xfree86-base-threaded [virtual package] xfree86-base-threaded-... [virtual package] xfree86-shlibs 4.4.0-14 Free X11 implementation for Darwin and Mac OS X xorg 6.8.1-12 Free X11 implementation for Darwin and Mac OS X xorg-shlibs 6.8.1-12 Free X11 implementation for Darwin and Mac OS X
Finally, here is where fink is at on this machine:
  fink --version
  Package manager version: 0.23.6
  Distribution version: 0.7.1.rsync
Roi's problem with the developer tool kit I have no idea on.

Jason has a G3 laptop in the lab that has gcc crashing in ncurses...
Please submit a full bug report,
with preprocessed source if appropriate.
See  for instructions.
make[1]: *** [../objects/read_entry.o] Error 1
make: *** [all] Error 2
### execution of /var/tmp/tmp.2.jS01rl failed, exit code 2
Failed: compiling ncurses-5.4-20041023-4 failed

Posted by Kurt | Permalink

02.17.2005 20:59

All major working sections sampled

Today I measured AMS on 3GW-s2 and took pmag cubes from 2GW-s1. This core comes from the evacuated zone of the Gaviota slide. That was the last major working half to sample.

Posted by Kurt | Permalink

02.16.2005 12:04

Sampled core 4

This morning I sampled pmag cubes out of core BPSIO-2004-4GW-Sec1. The photos are here

Sample, measure, sample, measure ....

Posted by Kurt | Permalink

02.15.2005 08:06

pmag paper of the week

Today is the paper that I picked for the weekly paleomag meetings. Here is what I chose:
  Kawamura, Ogawa, 2004. "Progressive change of pelagic clay microstructure
  during burial process: examples from piston cores and ODP cores,"
  Marine Geology, 207, 131-144.
First some definitions to help out:
  ped - A "large" aggregate separated by voids or sufaces of weakness.
        By large they mean like 10 μm

From macosxhints:
  Luckily, MySQL 4.1 has a feature to preserve backward compatibility by
  using the "old" encryption for your passwords -- users with such
  password are able to connect via CocoaMySQL without any problem at
  all. Creating an "old" password is simple in the Terminal, using the
  MySQL command line interface:

SET PASSWORD FOR 'mysql_username' = OLD_PASSWORD('password');

quicksilver - Extends the concept of Apple-Tab. The official site: quicksilver

Posted by Kurt | Permalink

02.15.2005 06:05

YahooMaps from AddressBook

Use Yahoo! maps with Address Book. I haven't had a chance to try it yet, but this would be really nice since the Google one did not work with Safari.

The annodex plugin looks cool for Firefox. If only I used firefox.

Ocean wave energey off of Scottland. What is the environmental impacts of this kind of thing? A pdf report about wave power in the US. Why is the US not going for more of these alternative energy resources?

voodoopadlite sounds interesting. Too bad the lite version does not export to html.

02.14.2005 19:42

Roi has fink troubles

I took a l look Roi's fink installation today. Fink seems to be fine but what happened to his developer tools? First I tried helloworld.c. No crt2.o. What? So I copied that from my machine to his. helloworld works fine. Then I tried building something in fink and ncurses still will not build. I also got rid of the UseBinaryDist flag in his /sw/etc/fink. Yikes. That is not a good thing to have. Here is where I am now with fink failing build anything. autoconf dies with the c pre processor...
checking whether we are using the GNU C compiler... yes
checking whether gcc accepts -g... yes
checking version of gcc... 3.3
checking how to run the C preprocessor... /lib/cpp
configure: error: C preprocessor "/lib/cpp" fails sanity check
### execution of /var/tmp/tmp.2.UHpHJC failed, exit code 1
Failed: compiling ncurses-5.4-20041023-4 failed
I tried to reinstall the developer tools but they say there is a newer version on the machine and refuse to install. Doh.

Posted by Kurt | Permalink

02.13.2005 19:06

Sampled core 3

I sampled core 3 today. I am really liking the Galerie created web pages.
Some interesting web sites:
postsecret - Very interesting.  And a bit scary.

Posted by Kurt | Permalink

02.12.2005 21:55

Posted by Kurt | Permalink

02.11.2005 15:19

Baumgartner 1992

I took Figure 1 from Baumgartner et al 1992 and used g3data to extract the locations of the cores in the figures. I then made a text file that looks like this:
  -120.05847761  34.2642265575   "241"
  -120.042416247  34.2650120181  "224"
  -120.035680911  34.2654127726  "230"
  -120.020617309  34.2290362724  "239"
  -120.02320158  34.2228575909   "214"
  -119.983294404  34.2124333872  "P-2"
  -120.042857811  34.1911209481  "265"
I then ran the fledermaus command line program to generate an sd file:
  /Applications/IVS\ Fledermaus/tools/cmdop mkpoint3d -type ascii\
  -format "x y" -in baumgartner92-cores.dat -out\
And that was all there was to it. I can now see these cores in the scene file

Posted by Kurt | Permalink

02.11.2005 09:59

New core sampling photos

I am now trying out using Galerie to generate the web pages for core sampling photos. The generally look pretty good accept 2 of the photos have their aspect ratios munched on the first batch. Here are the parameters that I used:
Main page title: Sampling 1GW Sec 1
Link:  BPSIO04 
Link addr: http://schwehr.org/Gaviota/bpsio-Aug04/core-sampling/
  Max size: 160
  Cols 4
  Rows 2
  Add title
  Add date
Template: Black
  Max: 600
  Add titel, Add date, Add EXIF, Add IPTC
  Add full-size orig
Quality 80
Create separate folders
Check out the results here:
The tops of the cores are pretty ratty since the cores are so soupy up at the top.

SQLite 3 is now in fink. Woohoo! Not that I will notice any difference since I hardly use any of SQL's features.

The is the first time that I have hit a lock with a fink cvs update. People, you need to stop using Mac OSX 10.2. 10.3 is so much better in many ways plus having fink maintainers worrying about 10.2 is wasted effort. I picked a busy time to do a cvs update...
  cvs update: [17:43:30] waiting for planetmirror's lock 
      in /cvsroot/fink/dists/10.3/stable/main/finkinfo/libs
Here is multibeam data from the Brits. No links to the actual data.
  ... British Royal Navy, Wednesday Feb. 9, 2005, shows a slide from a
  presentation that took place Wednesday at the United Kingdom
  Hydrographic Office in Taunton, Somerset, England. The picture shows
  the coastline off Sumatra in south east Asia at the epicentre of the
  earthquake that caused the tsunami in south east Asia on December 26,
  2004. The deep blue shows the deepest parts of seabed. (AP Photo/PA,
  Royal Navy)

Posted by Kurt | Permalink

02.11.2005 08:43

Larry Smarr at JPL

Larry Smarr of CalIt2 visited JPL last week. You will find Schwehr et al referenced at slide 16 with MER data visualization. The panoramas were created by JPL/MIPL with MICA.

2.04.05 Invited Talk to the NASA Jet Propulsion Laboratory
"LambdaGrids--Earth and Planetary Sciences Driving High Performance Networks and High Resolution Visualizations" Pasadena, CA

Versions: [html] [ppt]

Posted by Kurt | Permalink

02.11.2005 08:04

DICT protocol lookups

This is great! You have got to check this out if you are a fink user.

fink install dict
  dict ferrimagnetism

1 definition found

From WordNet (r) 2.0 [wn]:

ferrimagnetism n : a phenomenon in ferrites where there can be incomplete cancellation of antiferromagnetic arranged spins giving a net magnetic moment
  dict yacc

2 definitions found

From The Free On-line Dictionary of Computing (27 SEP 03) [foldoc]: yacc {Yet Another Compiler Compiler} From Virtual Entity of Relevant Acronyms (Version 1.9, June 2002) [vera]: YACC Yet Another Compiler Compiler (Unix)
Now I need to pass this through acoc to colorize the responses!

Also... Map Address Book addresses via Google Maps - Uses Firefox since Safari is not supported by google maps.

Posted by Kurt | Permalink

02.11.2005 07:50

Science web links

A good bit of rain today in San Diego today. It is nice to have a rainy day to offset all the sun.

Icebreakers Clear Channel Into McMurdo Station - There is the RVIB Palmer in the background. Glad they made it after there was all that trouble with the B15 iceberg.

Ultra-low Cost Well Monitoring Could Save Thousands Of Marginal Oil Wells - Cheap wireless devices let the operators shut down a system temporarily when the system is not producing enough flow to be profitable. These things are already out there, so better to keep them going that shut them down.

3-D Seismic Technology Locates Natural Gas In Fractured Reservoirs - The diffractions in fractured systems must make it really hard to see what is really going on down there.

ssh-tweaks - ssh tunneling. I should check out SSH Tunnel Manager (stm). I should really tunnel my email in to ucsd so that I stop getting these annoying ssl cert popups all the time.

Why does the iTunes Party Shuffle insist on playing lectures and speeches? I don't really want to hear myself talk while working.

Posted by Kurt | Permalink

02.11.2005 07:22

DMV's should not be tighly linked

H.R. 418 is supposed to stop terrorists. It is only going to punish regular citizens. It would force all state DMV's to share driving record information. Sorry, but I'm all for sharing felony convictions, but not this kind of stuff!

A friend and I got one ticket each in the span of 1/2 an hour outside of Dallas. I'm sorry, but they were just making money, not helping the public good. Or how about the cop hiding in the orange grove in the central valley of Calif. Thank goodness I was driving a federal vehicle so I could prove to the insurance company that I was driving for work. The cop was sure that I had stolen it. Now why would I steal a US Government minivan? How stupid do I look?

Yet every day here in SD, I see people really dangerous maneuvers (right turn from the left hand land, and visa-versa) right in front of cop cars and NOTHING?

That's enough of a rant for now. This law would seriously suck.

Here is the contact info for the senators from CA.

Posted by Kurt | Permalink

02.10.2005 19:45

Playing with emacs lisp

Here is a couple minutes of playing with emacs lisp evaluation. Emacs lisp must be the most widely installed version of lisp/scheme. I tried starting with the emacs list intro, but was not able to do anything so here is what I did:
  emacs test.el
  M-x lisp-interaction-mode
  (+ 2 2) C-j
  ; 4 gets printed to the screen when you hit the C-j
  ; Wow, a caculator
  '(yada yada "blah blah") C-j
  fill-column C-j
  ; Emacs says that the fill-column is 70
  C-u 6 6 C-x f
  fill-column C-j
  ; Emacs says that the fill-column is 66
  (set 'fill-column 10) C-j
  fill-column C-j
  ; Emacs says the fill column is 10... fun fun
  (concat "abc" "def") C-j
  ; Emacs says "abcdef"
  (substring "The one true editor is emacs uber alles" 23 28) C-J
  ; "emacs"
  (+ fill-column 60) C-j
  ; add 60 and fill-column.  Very important math
  (concat "The " (number-to-string 2) " two" ) C-j
  ; emacs says "The 2 two"
  (+ 1 2 3 4) C-j
  ; 10
I am now having flash backs to Nick Parante's CS107 class in 1993. That was the only time I really wrote any lisp other than when I tried to create an emacs mode for ESRI's Arc Macro Language (AML). The AML mode actually kinda worked, but I forgot to copy the code from sierra.wr.usgs.gov before I lost my account. Back to work!

Posted by Kurt | Permalink

02.10.2005 17:37

Visualization Class video stream plus svn emacs mode

psvn.el is an emacs subversion(svn) interface. The Red-Bean SVN book

If you are at UCSD and looking for system support for you computer, consider contacting the Chemistry and Biochemistry Computer Facility

Joe Keefe setup recording one of my talks at the VizCenter 09-July-2003. Check it out here

Or jump straight to the RealPlayer stream: http://chemcomp1.ucsd.edu:8080/ramgen/ks-vislect2.rm. I had to open RealPlayer and then File->OpenLocation and paste in the URL.

Posted by Kurt | Permalink

02.10.2005 13:56

PDF of Santa Barbara varve data to gnuplot

Attached to Wolf Berger et al (2004) is a data file of varve thickness, TOC, and fish scales that I wanted to take a look at. Here how I did it. First get the data and make it text
  wget ftp://rock.geosociety.org/pub/reposit/2004/2004055.pdf
  pdf2text 2004055.pdf
  emacs 2004055.txt
Use your text editor to save the varve years to one file and the varve thicknesses to another. It must be one value per line. e.g. M-x replace-string space C-q C-j. Then use the unix paste command (thanks Julie!) to bring the two columns together:
  paste mm.dat year.dat > varves-mm-year.dat
Now use this python script to convert the mm for each year into a depth of each year:
#!/usr/bin/env python

import os # os.remove() import sys # sys.exit() import string

inFile=open("varves-mm-year.dat") depth=0. # cm print "# depth in cm, calendar year" for line in inFile.xreadlines(): mm,year=line.split() depth+=float(mm)/10. print depth,year

Then fire up gnuplot:
  set yrange [] reverse
  set xrange [] reverse
  set xlabel "year"
  set ylabel "depth(cm)"
  plot 'varve-depthcm-year.dat' using 2:1
And now you have a nice graph showing the varve thickness decreases with depth and that compaction happens. Woowee.

Posted by Kurt | Permalink

02.10.2005 09:22

crontab for updatedb

Here is the crontab for root that updates man pages and the locate commands database. Do "sudo crontab -e"
5 0 * * * /sw/bin/updatedb
30 0 * * * /usr/libexec/locate.updatedb
15 0 * * * /usr/libexec/makewhatis /sw/share/man
To view your crontabs you can do:
  crontab -l
  sudo crontab -l

Posted by Kurt | Permalink

02.10.2005 09:06

Spring AGU - New Orleans

Here is my abstract for AGU New Orleans. I am only part way through the lab work and data processing so there will be more results in the meeting than what is discussed here:

                Slumps and sediment deformation:
              AMS of the Gaviota and Goleta Slides

During the last year, we have collected six gravity cores from the
northern side of the Santa Barbara Basin for paleomagnetic
deformation tests. The first two cores were collected on the
smaller Gaviota Slide on the western side of the basin.  These
cores provide the slump end member, since one core comes from the
evacuated zone while the other comes from the toe or accumulation
zone.  Two additional cores were collected on the toes of the much
older Goleta Slides and were designed to provide an undeformed
reference fabric; it was our hope that they would be mostly
pelagic drape but also penetrate into the top of the deformed
material.  One core appears to be an excellent reference core with
clear laminations, while the other has distorted laminations that
show the sediment has undergone some type of deformation.  In
between the two slides, we collected two cores on either side of
the crack that runs between the two slides to investigate creep
and slope stability.  All cores have been sampled at a 3 cm
interval with 8cc cubes to measure anisotropy of magnetic
susceptibility (AMS) and AF demagnetiziation.  The Gaviota slide
sediments show a dominantly triaxial fabric with some alignment of
the maximum eigen values, while the drape materials from the
Goleta Slide show a typical oblate undeformed fabric.  There are
hints of deformation at major clay layers, which we continue to
investigate.  Several of these cores appear to show a biogentic
component to the AMS signal that fades with depth.  This component
has a tendency to mask the deformation signal in the top sections
of the cores.  The NRM inclinations of the deformed core regions
tend to be shallow and erratic showing what look at first glance
to be excursional features.


Experimental Radar Provides 3-D Forest View - Using SAR to measure the biomass between the forest canopy and the ground.

Updating my Macs to 10.3.8. Seems that things are ok so far.

More todo item:
  • Add gplate.org to fink
  • Add CoinDesigner to fink
  • Add the SGI Inventor programs to the fink/Coin/SoXt system (needs a new release of Coin past 2.3.0)

Posted by Kurt | Permalink

02.09.2005 18:57

Resurecting data

Sounds like the meeting between CalIt2 and JPL went well last week with posibilities for future collaboration.

Today I resurrected some paleomag data from last weekend. It had been quite a while since I had used flow with cubes, so I had put the cubes in with the arrow point down into flo rather ran out of the side of flow/up at the ceiling. Lisa's pmag library has a nice little program call di_geo to help that I wrapped into a little script:
rm -f better.mag
cat bad.mag | while read line; do
    decinc_bad=`echo $line | awk '{print $5,$6,0,-90}'`
    decinc_good=`echo $decinc_bad | di_geo`
    first=`echo $line | awk '{print $1,$2,$3,$4}'`
    last=`echo $line | awk '{print $7,$8}'`
    echo "$first    $decinc_good $last" >> better.mag
By the way, I used code2html to make html out of the script.

Posted by Kurt | Permalink

02.09.2005 13:52

Apollo Quicktime VR's

Check this out when then slashdot traffic has died down:

Panoramas.dk - This site even beat us to the MER quicktime VR's since we were so overloaded early in the mission.

Posted by Kurt | Permalink

02.08.2005 17:50

AMQ and ActiveMQ

Is AMQ real? It is supposed to be a Wall street version of network. Supposedly like ActiveMQ? What a terrible press release. I was pretty confused by the end. Why would Sun put this in their kernel? Is this a replacement for Tibco's software? How does this compair to the standard the RTI helped make with NDDS? Or Ensemble?

Then there is Spread which has its own open source license. Arg. Another license.

Or corba. Or yada yada yada...

Posted by Kurt | Permalink

02.08.2005 10:28

Random Links

Handheld Transmitters Connect Students And Teachers In Class

print from the command line using the Print Center's default printer like this:

open -a Print\ Center /path/to/file_to_print.pdf

This morning, I got fooled by the UCSD network folks. It looked like a virus bombarding my webserver with IIS attacks. Turns out they were just doing a sweep for vulnerable machines.

Posted by Kurt | Permalink

02.08.2005 06:39

Pro nuclear power argument

I was reading an article the other day that was pro-nuclear power. I have been very against nuclear power for a long time, but I came up with a twisted reason for why I should be pro-nuke.

As practitioner of marine geology and geophysics, I am always on the lookout for environmental markers in my study areas. These markers can be ash layers, certain fossils, storm events, 14C datable material, Pb210, Cs, major climate shifts, etc. Note that Cesium is in that list. Most of these we, as humans, do not create. But sometimes, scientists are able to do "labeled releases." For example, you could coat a pile of sand gains with something unusual like a special paint and then release a couple kilograms of marked sand in the mouth of a river. Then down the coast at a beach by a submarine canyon head, you could continually sample the sediments looking for these marked grains to appear. You have a really nice experiment for the residence time and transport rate of sand in your local beach cell.

Back to Cs. There is often things we would like to do to label things on the Earth that are not environmentally sound, so we don't do them. I would not go dumping a giant amount of radioactive tracer into the ocean. Here is where the nuclear industry comes into action. It seems that we been having a nuclear disaster about once every decade or two. Chernobyl, England, Washington State / Hanford, etc. These are perfect marker events. The release very unusually compounds and elements into an environment. We then come along and look for the, say, early 90's Cs spike. How handy. Now if we were to really ramp up the nuclear industry, especially in the developing world, the number of accidents would go way up and give us all kinds of great radionuclide markers. Wonderful! Go nukes! But you might want to pass on the three eyed fish for your sushi.

Posted by Kurt | Permalink

02.07.2005 07:37

Random links - thesaurus

Nisus Thesaurus - This appears to be free. 0.4 will add a system-wide dictionary and thesaurus, but until then...

I think I live in the middle of a crow convention. Quite the ruckus.

MacOSX hint:
I dragged an AAC file to ~/Library/Sounds and launched iCal. To
my horror, the sound that I put in there didn't appear in the
pop-up list of alarm sounds. "Okay," I thought, "maybe it's
expecting an AIFF file."

Not having an AIFF lying around on my new machine, I decided to rename the AAC file to have a AIFF extension. Magically, the sound appeared in the alarm settings dialog, but I didn't expect it to play. Guess what? It did!

Posted by Kurt | Permalink

02.07.2005 07:06

Todo list

Here is my todo list for stuff outside my thesis that probably will not get done for a while. This is the "I have not forgotten" list. Unless a quick solution lands in my email inbox, I will not be doing these until after I get my thesis done!
  • Get fink cwp-su to release 38. At 37 now.
  • Get OpenDTect into fink... or even just try it
  • Upgrade acoc from 0.6.0 to 0.7.0 in fink. 0.7.0 bombed on me for some reason.
  • Get gmtplus into fink cvs assuming that it worked for Steve Cande on the Palmer
  • Add --enable-darwin-x11 to the Coin fink configure and test
  • Get fink mbsystem working with GMT 4. Waiting for mbsystem 5.0.6
  • Talk to Bill Danforth and get xsonar info fink
  • Update fink proj to 4.4.9
  • Add thuban to fink
  • Help get fink sqlite to version 3.
  • Add hsv2rgb and rgb2hsv to density
  • Release a new version of xcore that works on Mac OSX 10.3
  • Setup a real backup program like I described. Maybe hdup will do it.
  • Have nanoblogger add the DESC field to the index
  • Fix my phrase index for nb to use the article number
  • Fix my phrase index to list phrases by article
  • Robustify my phrase index. Maybe use python sets or regex?
  • add mactounix and unixtomac in density
  • Setup phpicalender
  • Why does my phrase index not include fink?
  • Get more of the tools at www.remotesensing.org/ into fink
  • Check out all the projects that use gdal: gdal projects
  • Get OpenEV on the Mac

Posted by Kurt | Permalink

02.06.2005 20:01

Fink shlib policy change

Uh oh. I have some shared lib packages.

From - TheSin Subject - Proposed Policy Change for all Maintainers IMPORTANT

With the upcoming shlibs change, the packaging engine needs to be changed a little and this will break many packages.

first let me explain the shlibs change. Basically fink will beable to automatically insert all share library deps to the Depends line by using Depends: {SHLIB_DEPS}, curl | curl-ssl. This is an example and I added curl | curl-ssl to show that we will still need to add runtime depends by hand, but now dylib pkgs.

Now currently fink treats Depends and builddepends in the pretty much the same manor. In the new layout Depends will be Install/Run tim depends, so not needed to build a package. And since in the new policy all -dev pkgs will be required to depend on there own -shlibs, we can rest assured that the -shlibs will be present at build time so no need to list them in the BuildDepends, except if a -dev isn't currently following that policy. But there are only a few that aren't at this point so that won't be a big change.

Anyhow, any maintainer wishing to help us out and help speed this new great tool up for us, please follow this next set of instructions to fix your pkgs:

all pkgs such be able fink build foo with out the depends line so comment it out and add what is needed to builddep line. This is to test build only, so use fink build to test this and not install. Also please don't commit it like this. Again this is only to test that the pkg will build sans the Depends line. Install and Run time deps are still required at this point and that includes the -shlibs pkgs so don't start removing those yet.

Also we NEED and I can't stress this enough, need proper shlibs fields. This will are receive a little more documentation I hope and much more clarification in the policy and packaging section of our website.

The format is as such

Shlibs: << %p/lib/libfoo.5.dylib 4.0.0 %n (>= 4.0-1) <<

1) installed lib name of the major versioned lib, that means the actual dylib not the one of the symlinks 2) the compat version of the lib, you can get this by running otool -L on 1) 3) the lowest version of the pkg that has this lib, should almost always be %n and a -1 revision of when the compat version of this lib first made it into a fink pkg, This value shouldn't change until the next time the compat version on that lib changes.

Also contrary to some pkgs and what we may have stated before please do NOT use foo-shlibs (>= 1.0-1) | foo-ssl-shlibs (>= 1.0-1). We may allow it for x11 and nox variants but for now please only one pkgname and version.

The more Maintainers do the faster we won't need to worry about Depends anymore :D

Posted by Kurt | Permalink

02.06.2005 18:36

blender brain dump

Last year, I did some brief work on terrain generation code to convert various input grids to different output formats. Basically, I wanted to be able to take gmt netcdf grids and output them into a formats that could be read by LightWave, Maya, Blender, and Povray. The first three all read WaveFront obj files which are pretty simple compared to things like the lightwave file format. LWO2, is um, ouch. Here is a sample OBJ file:
o sample1o
g sample1
mtllib sample1.mtl
usemtl sample1mtl
v 0 0 -1.88
v 1 0 -1
v 2 0 0
v 0 1 -1
v 1 1 0
v 2 1 -0.2
v 0 2 -0.35
v 1 2 -0.323
v 2 2 -0.301
vn 0.551207 0.551207 -0.626372
vn 0.568636 0.568636 -0.593691
vn 0.19245 0.19245 -0.7698
vn 0.59035 0.515392 -0.615405
vn 0.240578 0.128444 -0.774256
vn -0.119176 -0.198286 -0.963921
vn 0.334089 0.0551787 -0.796885
vn -0.0464649 -0.236557 -0.960263
vn 0.0218834 -0.100465 -0.9947
vt 0 1
vt 0.333333 1
vt 0.666667 1
vt 0 0.666667
vt 0.333333 0.666667
vt 0.666667 0.666667
vt 0 0.333333
vt 0.333333 0.333333
vt 0.666667 0.333333
f 1/1/1 4/4/4 2/2/2
f 4/4/4 5/5/5 2/2/2
f 2/2/2 5/5/5 3/3/3
f 5/5/5 6/6/6 3/3/3
f 4/4/4 7/7/7 5/5/5
f 7/7/7 8/8/8 5/5/5
f 5/5/5 8/8/8 6/6/6
f 8/8/8 9/9/9 6/6/6
If you want this thing to display textures, you need an mtl file to go with it. Here is the one that goes with this. Hack it to load the image you want. sample1.mtl:
newmtl sample1mtl
Ka 0 0 0
Kd 1 1 1
Ks 0.2 0.2 0.2
illum 2
Ns 25.6
map_Kd sample1.jpg
You will have to dig up the specs on the format to get into the nitty gritties. You can make things simpler if you just want geometry and can ignore the normals (vt) and the uv texture map (vt).

I tried blender 236 this evening for about two minutes. Major improvements over when I last used it. Do a File->Import->Wavefront and select this file. You will see 8 triangles folded over. Then do Render->Current Frame to see a rendered scene. Also, you may need to delete the default box from the scene with Object->Delete (at the bottom of the geometry window). Use the middle mouse and scroll wheel to zoom around the scene. That is all I can dredge out of my head in just a couple minutes.

Posted by Kurt | Permalink

02.06.2005 16:24

Thuban - pysqlite based geographic data viewer

In trying to find more info on pysqlite for my data import and processing scripts, I ran into Thuban. This is a Interactive Geographic Data Viewer that can read Shapefile, PostGIS, and GeoTiff. Looks nice and it uses wxWindows, proj, pysqlite, and optionally GDAL. I was unable to get the big gdal based GIS program to port to mac osx (what was it called?), but maybe this could be ported. It looks like all the prerequisites are already in fink. When I have to next deal with one of these data types I may give it a go.

CBEL list of python resources - Nice jumping off point.

Posted by Kurt | Permalink

02.06.2005 14:34

First python sqlite scripts

I just wrote my first python script using pysqlite. It loads in the magnetometer data from the current format used by both Bubba and Flo. The big bonus is that it is about a billion times faster than going from bash. My bash import code made about 100 exec calls for each k15 entry including the startup and shutdown of the sqlite command line interface. Talk about ways to slow yourself down. With perl or python all those awk like commands are just the use of the builtin split command. And the "cat file | while read line; do" loops are just "for line in magfile.xreadline():" which is a little more intuitive for me.

Here is the script! 20 lines of python. It could use a little error cheecking, especially around the "create table mag" command.

Posted by Kurt | Permalink

02.06.2005 12:45

Beagle 2 report pdf

Here is the pdf of the actual report that I quoted from before: Beagle 2 Report

Not as in depth and insightful as I had hoped.

Spent the morning measuring NRM's of cores 1,5,and 6 from the Santa Barbara Basin. Got a 160 measurements. Now I need to add Flow's datastream into the database. I think I will use python this time and see how it goes.

UNH Space Scientists Awarded $5 Million From NASA To Study Edge Of Solar System - Eberhard Mobius and Marty Lee will work on NASA's Interstellar Boundary Explorer, or IBEX, mission.

Posted by Kurt | Permalink

02.05.2005 12:08

Watch the robot run and Gee et al. 2005

Myche pointed me to Honda's Running Robot. Holly smokes. What will things be like in 20 years? Will the dreams of Hans Morevic come true? One of the points that stands out in my mind from my time at CMU was wandering in to talk to Hans. I did not really know who he was at the time, and I doubt he knew who I am. We was one of those unassuming people who is brilliant and it takes a while to realize this. He tolerated my questions for quite a while that late night. I think it was around 10 or 11 at night and I was waiting around to talk to meet up with some other robotics hackers at CMU. Maybe it was Liam who had an office right around the corner. CMU is a strange place. I spent hours in the highbay and labs by myself. But the number of wild projects and ideas that came by me was astronomical. It is hard to describe why CMU is what it is, but it is truly one of the hubs of the robotics world. I remember hanging out with a bunch of RoboGrads when we first say the Honda robot walk on a video. Everyone was floored.

Added a dynamically generated LaTeX table to my paper today. So this did not have to autogenerated, but I think it is really better this way. If I ever find a typo, I go back to the raw data tables and fix it. My next task is to build a better core photo picture. In working with Lisa on Friday trying to show where in the cores certain features are, it was instantly clear that having to add an offset of the top section is just unacceptable for a published paper. It is great with the database, I do all my work in section space and then the database warps everything into depth space. Section space is really the way to take your data. There is never any question where a particular bit of mud comes from in the actual liner. It is NOT good for analysis.

I am starting to dive into the background literature for the Santa Barbara Basin, slumps/slides, and recent AMS work. Today I read Gee et al. 2005. That is not Jeff Gee, but M. Gee of Manchester. This paper is a buried slump described from industrial 3D seismic survey off Angola. Some quotes and notes:

"The upper landslide surface tends to preserve depositional or stopping structures, whereas the basal landslide surface can yield valuable information about dynamic landslide emplacement."

What is a "counter-regional" growth fault?

This slided had a 40 meter head scarp and 9km long basal striations. The paper talks about basil morphology along with how and when slides are able to increase their volume, which basically means that they are picking up material. I originally though they were talking about increasing the water content/porosity of the sediments as they are transported, but that is not what they are talking about here.

The average slope of the basal surface is ~1.6°, with the steepest being 3°.

For the upper surface, "The rafted blocks have similar dimensions to these failure scars, although the block geometry suggests that they have rotated and plastically extended."

I have not read about the Grand Banks slide from the 1929 Earthquake before. It will be interesting to compare this to what comes out of the Sumatra tsunami of Dec 2004.

I have to thank Duncan Mallace for the link to the link to the in progress cruise of the HMS Scott that is in progress over the Sumatra earthquake area. http://www.intermargins.org/tsunami1.html#a

Posted by Kurt | Permalink

02.05.2005 07:26

Phoenix experimentation system

Check out the Phoenix system for electronics experimentation. This is very cool looking and would be useful for us in our labs. Could also be a lot cheaper! There is a Linux Gazette Article about this project.

Posted by Kurt | Permalink

02.04.2005 16:00

What to do with your photo ipod

Your ipod photo may not take pictures on its own but, steal your friends an make an ipod stereo scope.

So it's time to ignore the tech world and focus in on things. I have nuked 2/3rds of the RSS feeds from NewsFire. I am down to 18 and it may go lower.

Jason gave me a quick walk through on AF demagnetization this morning. It has been a long time since I last used this stuff so it was good to get a run through.

Sugarbush is back to life. I did an nmap and found way to much open so it was time to shut off services:
  chkconfig --del mysql
  chkconfig --del nfs  
  chkconfig --del smb
  chkconfig --del portmap
  chkconfig --del nfslock
  /etc/init.d/nfs stop
  /etc/init.d/portmap stop
  /etc/init.d/mysql stop
  /etc/init.d/smb stop
  /etc/init.d/nfslock stop
All acounts should have their old passwords and the ssh_host keys are all from the previous Mandrake 8.1 setup. Looks like sioseis, seismic unix, and xsonar got nuked. Whoops. Where were they installed? We need to get the latest on there anyway.

Posted by Kurt | Permalink

02.03.2005 21:34

Donald Knuth

I guess I must not be a real computer scientist. In reading Knuth's website, I realize that I do not own a single book by him. Oh wait, I own Literate Programming. Not really like his other books and it was hard to get into. Not like ready Deep C Secrets. The orange fish ran off with someone at Scripps and I can't remember who. Someday I should really study an algorithms book and maybe buy his classic algorithms volumes. No point in reinventing the wheel.

When is fink going to get sqlite3 and digest-sha1-pm584 is stuck at 2.10-3.

Today was a slow struggle getting Mandrake 10.1 on sugarbush working with Jenna. Here are some quick notes about odd things I ran into:
  • After about 20 Mac OSX 10.3 installs I have to say that Mandrake has some catching up to do.
  • Don't make your pokey DVD drive be the master. Slow CD reads. Like 1 hour on a P4 box. Ouch.
  • Paranoid mode is cool, but ouch is that not handy for a normal user. You can keep people from running "w" on your box. If I was the military, I would say that this is a great start.
  • Make sure you configure your mouse!!! We left the mouse as a "generic PS/2 or USB." After install, we rebooted and Mandrake ignored both the PS/2 mouse that was attached and a USB mouse I added.
  • The ucsd network is weird. If my netmask is ff.ff.0.0, my gateway is sitting way up on the top of the address range? I am not used to this.
  • Turning on and off services is not in an obvious place. I like to have an ftpd installed but turned off just in case of emergencies.
  • Very few people at ucsd realize that we have a campus ntp... bigben.ucsd.edu aka ntp.ucsd.edu.
  • I think I was spoiled as Stanford with network and sysadmin support back in the early 90's. They were kickin butt with AFS in 1993. Larry Schwimmer and co rocked from a user point of view. Hey wait, my xenon account is in afs space. Awesome!
  • Printer setup is insane. If I am setting up a network printer and have not setup the network, mind asking me to setup the network and maybe ping the printer? I gave up on the printer install.
  • The dependencies for the downloaded updates were seriously jacked up. I haven't even started messing with things and rpm is bitchin about unmet dependencies. Flashback. Should have installed debian.

I had fun trying to get points into fledermaus this evening. I needed to redo the planned cores file into one that had the actual GPS loc at the time the king-kong was poking a hole in the sea floor. It was easy enough to pull the values out of sqlite, but then I ran into all kinds of problems. Forgot the minus sign to say that I was in the WESTERN hemisphere. I did not see the equivalent to grdinfo for fledermaus sd files. That would be useful. The -info option did not do what I thought and stopped the command from working. Say what? But I did succeed and once this are working, they work great. Here is the data and the command:
-120.108 34.36116666 "1"
-120.1075 34.369999999999997 "2"
-120.05883333333334 34.365666666666669 "3"
-120.05716666666666 34.378999999999998 "4"
-120.02356666666667 34.337049999999998 "5"
-119.9772333333333 34.329983333333331 "6"
-120.03666666666666 34.287500000000001 "ODP"

/Library/Application\ Support/IVS/bin/cmdop \ mkpoint3d -in loc-fleder.dat -type ascii -format "x y" \ -out santabarbara-basin-cores.sd
If you have cores in the Santa Barbara Basin, EMAIL ME THE LAT/LONGS!!! I will add them to my data base and release a scene file with them all. Big Dave, this means you! Juan Carlos? I know there are hundreds of cores in the area.

Tomorrow Larry Smarr will be visiting JPL.

Posted by Kurt | Permalink

02.03.2005 09:10

Beagle 2 Leassons learnt

Check out the Lessons learnt from Beagle 2 and plans to implement recommendations from the Commission of Inquiry

I like these failure reports as it gives an insite to how these systems are designed and operated. this one is not the best ever, but I'll quote it here so that I can search it and not loose track of it.

The recommendations of the Commission of Inquiry:

Recommendation 1 Future lander missions should be under the responsibility of an Agency with appropriate capability and resources to manage it. The lander/orbiter mission should be managed as an integrated whole. Nationally-funded science instruments should be included in the lander on the same basis as on the orbiter.

Recommendation 2 For future science payloads which are critical to overall mission success or have a very high public profile, the ESA Executive should make a formal, comprehensive assessment of all aspects of the proposals including technical, management and finance, and advise Space Science Policy Committee (SPC) accordingly before acceptance. If the assessment is not positive, ESA should advise the SPC not to accept the proposal.

Recommendation 3 Sponsoring Agencies of nationally-funded contributions to ESA projects should ensure that the required financing is committed at the outset to meet the estimated Cost at Completion and require that a structured development programme is established.

Recommendation 4 In addition to the ESA-led reviews of interfaces, formal Project Reviews of nationally-funded contributions to ESA missions should be undertaken by the sponsoring Agency to a standard agreed with ESA and should cover the entire project.

Recommendation 5 When an independent review of a nationally-funded project, such as the Casani review of Beagle 2, is commissioned, it is essential that ESA and the Sponsoring Agency ensure that its recommendations are properly dispositioned and those which are agreed are actioned and followed up through a formal process.

Recommendation 6 For future projects, Heads of Agreement or similar formal arrangements between co-operating entities, ESA, and national sponsors, should be put in place at the outset of projects and should include formal consultations at key stages of the projects to jointly consider its status.

Recommendation 7 Fixed price contracting should be avoided solely as a mechanism for controlling costs, and used only where the sponsor and contractor are in alignment on the requirements and scope of the work and the sharing of risks between them. Both parties should be confident that the contractor has sufficient margins to manage his uncertainties and risks.

Recommendation 8 For future high-profile/high-risk projects, ESA and any Sponsoring Agency should manage the expectations of the outcome of the project in a balanced and objective way to prepare for both success and failure.

Recommendation 9 At the start of a programme, the funding authority (ies) should require that there is system-level documentation. This is necessary to provide all partners with the technical requirements for the project and sufficient design description and justification such that the margins and risks being taken in each partner's area of responsibility are visible.

Recommendation 10 Future planetary missions should be designed with robust margins to cope with the inherent uncertainties, and they should not be initiated without adequate and timely resources to achieve that.

Recommendation 11 Future planetary entry missions should include a minimum telemetry of critical performance measurements and spacecraft health status during mission critical phases such as entry and descent.

Recommendation 12 For future planetary entry missions, a more robust communications system should be used, allowing direct commanding of the lander for essential actuations and resets without software involvement - enabling recoveries in catastrophic situations.

Recommendation 13 Planetary probe missions involving high-level shocks from pyros and other events should undergo representative shock environmental testing at system level.

Recommendation 14 Adequate and realistic deployment tests should be performed, and sufficient time and resources must be available in the development of a new planetary mission.

Recommendation 15 Elimination of internal connectors for mass saving should be avoided if at all possible. But if unavoidable, a stringent system of check and independent crosscheck should be followed during the final wiring operation.

Recommendation 16 A back-up for the entry detection event (T0) must be included in the design of planetary entry probes.

Recommendation 17 Future planetary entry missions should include a release of the back cover and front shield, which is aerodynamically stable and analytically predictable to avoid uncontrolled contact of front shield with the lander.

Recommendation 18 Sufficient difference between ballistic coefficients of all separated items, e.g. back covers assembly and the main parachute, or other positive means, must be ensured to exclude collision after separation.

Recommendation 19 Adequate competencies in air bag and parachute technology must be available for future European planetary missions, making best use of existing expertise e.g. in USA and Russia.

Where is the full report?

Posted by Kurt | Permalink

02.03.2005 08:24


Using phpmyadmin, I attached to the WordPress database in MySql and there were all the comments. I have now forgotten the admin password that I changed it to. Doh! I was able to use phpmyadmin to change the comment_approved field to 1 for a comment and it appeared. Nice. It is possible to just let everyone comment? Or will Lisa's class pages get filled with spam? Can WordPress use TypeKey? Here is a thread discussing a TypeKey plugin for WordPress

Posted by Kurt | Permalink

02.03.2005 06:41

Seismic tools

I got an email from Freddy Aldersons about the release of some free seismic programs. Too bad he does not release the source. I can't put binary only stuff in fink. Sorry.

Geocenter just announced a release of SeisUP on freshmeat.net, but this is unfortunately not an F/OSS system.

Jenna is still asking for Landmark since it can pick a horizon from multiple lines and make a surface. Can't other programs do that too?

Gebco - General Bathymetric Chart of the Oceans.

Cardiff Univ, UK has a program that "features novel 3-D methods to map the seabed." They are using software by CodaOctopus. I have never heard of them. But that is not too surprising as I live in the land of MB-System. (ScienceDaily)

Maybe hdup2 would be a good simple backup system.

Check out the HVO Live Camera of Kilauea. Very cool! Kilauea was just on CNN.

I have always wanted to check out smalltalk since this is one of the birth places of object oriented programming. Maybe someday I will try GNU Smalltalk which is improving the portability with Mac OSX. I won't be trying this anytime soon!

Posted by Kurt | Permalink

02.02.2005 22:09


I have MySQL, phpMyAdmin, and WordPress all installed and working. It does not seem to want to take my comments. I need to figure that out. I click "Say It" and end up back on the same page. That's it for tonight.

Posted by Kurt | Permalink

02.02.2005 16:55

Google name polution

I have been running into a problem lately. For example, I am trying to find info postly by people who have hints on how to get MovableType working on Mac OSX. But many MovableType blogs have a link to MovableType.org on the page. This destroys my ability to search on say osx and movabletype. I get back lots of MovableType weblogs about Macs that have nothing to do with MovableType, but the blog is being done through the program. Arg!

Maybe I could install MovableType easier on a Linux box? I do not have a good one handy to try.

Posted by Kurt | Permalink

02.02.2005 16:40

Movable type frustration

I tried to help out Lisa with getting Movable Type installed on any Mac OSX box. After about three hours, I have not made it to where she got to. She was able to add entries but had things lock up when trying to add comments. She and I both got this message when trying to use the TypeKey signin:
  The site you're trying to comment on has not signed up for this
  feature. Please inform the site owner.
Looking at the internals of MovableType, I decided to attempt to start from the beginning on my laptop (10.3.7 + fink). I finally got sqlite working and got to try to add a comment and it blew up with this message. BTW, a lot of their cgi's spew warning at the end.
The 'year' parameter (undef) to DateTime::new was an 'undef', which is
not one of the allowed types: scalar Params::Validate::__ANON__('The
\'year\' parameter (undef) to DateTime::new was an \'unde...') called
at /Users/schwehr/Sites/blog/mt/extlib/Params/ValidatePP.pm line 467
\'year\' parameter (undef)') called at
/Users/schwehr/Sites/blog/mt/extlib/Params/ValidatePP.pm line 346
Params::Validate::validate('ARRAY(0xad0520)','HASH(0xaf2950)') called
at /Users/schwehr/Sites/blog/mt/extlib/DateTime.pm line 165
called at /Users/schwehr/Sites/blog/mt/lib/MT/Template/Context.pm line
called at /Users/schwehr/Sites/blog/mt/lib/MT/Template/Context.pm line


called at /Users/schwehr/Sites/blog/mt/lib/MT.pm line 496 MT::rebuild_entry('MT::App::CMS=HASH(0x8c3e3c)','Entry','MT::Entry=HASH(0x992ca0)','BuildDependencies',1,'OldPrevious',2,'OldNext',1,...) called at /Users/schwehr/Sites/blog/mt/lib/MT/App/CMS.pm line 3426 MT::App::CMS::rebuild_pages('MT::App::CMS=HASH(0x8c3e3c)') called at /Users/schwehr/Sites/blog/mt/lib/MT/App.pm line 352 eval {...} called at /Users/schwehr/Sites/blog/mt/lib/MT/App.pm line 301 MT::App::run('MT::App::CMS=HASH(0x8c3e3c)') called at /Users/schwehr/Sites/blog/mt/mt.cgi line 26 eval {...} called at /Users/schwehr/Sites/blog/mt/mt.cgi line 20
Perhaps it is the version of perl that ships with osx?
  perl --version
  This is perl, v5.8.1-RC3 built for darwin-thread-multi-2level
Briefly, here are some (but not all) of the steps I took:
sudo emacs /sw/etc/apache2/httpd.conf
    AddHandler cgi-script .cgi
And added this at the end:
        Options ExecCgi Includes MultiViews Indexes

emacs -nw ~/Site/blog/mt/mt.cfg and add: CGIPath http://localhost/~schwehr/blog/mt DataSource ./db ObjectDriver DBI::sqlite Database db/mtdb.sqlite

open http://localhost/~schwehr/blog/mt-check.cgi sudo cpan Answer the default to all questions and wait about 15 minute for cpan to sort itself out and get the mirror file. install DBI install SQLite All the other cpan modules I tried choked.

cd /Users/schwehr/Sites/blog/mt && chmod 755 mt-load.cgi

open http://localhost/~schwehr/blog/mt/mt-load.cgi

chmod 000 mt-load.cgi

open http://localhost/~schwehr/blog/mt.cgi

Go read the manual for the default login and how to setup...
What is the trick to Mac OSX 10.3.7 for Movable Type? Do I need to switch to fink perl?

I should really see if I can get comments working in nanoblogger, blosxom or pyblosxom.

Posted by Kurt | Permalink

02.02.2005 14:47

qsplat and movable type

qsplat can render hundred-million poly models. Wow.

Working on trying to install movable type on my laptop. Frustrating. I am trying the free for personal use instant license. cpan is cranky. I may have to hack all of the mt files to use /sw/bin/perl if I can't get this file to download! Finally... now I am in a windy maze of installation stuff.

Posted by Kurt | Permalink

02.02.2005 12:16

BPSIO update

MayaVI - VTK based data visualizer with a slick screenshot that makes me think of perfly for some reason. The VTK Pipeline seems much like the 1998 Viz network/shell interface for talking to objects. I want pickle/marshal for C++.

I fixed the core photo for BPSIO-04. Core 2 has only one section and it was mislabeled as 2-sec 2. Thanks Neal for catching this on Monday during my talk.

I changed around the bpsio database. The hext table is now a part of the ams table. Queries are much simpler now since the depth field is in with the raw ams data.

I got bibtex to export from EndNote 7:
  bibtex style in 
    Edit->output Styles ->Style Manager
  after you add bibtex to your favorites
  Your exported file will be bibtex with mac end of lines
It does not seem to handle labels correctly, but this is pretty good.

Technorati - search the blog sphere.

Posted by Kurt | Permalink

02.02.2005 06:17

Tsunami waves from space

Check out these two articles about the December tsunami waves being observed from JPL's MISR spacecraft:
JPL Mister 12-Jan-2005 release
From ScienceDaily: How Noise Protects Entire Marine Ecosystems. This is not talking about anthropogenic noise like seismics. And once you read the article, you might also think, "Duh!" Wave energy drives all kinds of important processes in the near shore environments. Agitation is critical.

18 Live Linux CDs reviewed if you want more choices over just knoppix. The article recommends PCLinuxOS (www.pclinuxonline.com). I have not tried this one. I still am amazed at knoppix, which saved our data from sugarbush when Mandrake 8 went belly up last month somewhere in the boot process (DevFS troubles?)

Wikipedia's entry on the competition for NASA's Crew Exploration Vehicle. I can't wait till we end the Space Shuttle and have de-orbit'ed the space station (that is going to be SCARY).

Software patents must die. I agree with the quote below. We need stricter software copyright protection with a REDUCED length of time. Maybe 25-30 years tops for a software copyright. And we need help with solidifying the open source licenses. That would really spur invention and economic growth. If you are living off of software copyrights from 25 years ago, you are holding up the whole community. Get with the times and stop using PCDOS 2.1.
"If you could not patent software algorithms or ideas, how much of the
money spent on writing software would go away? How much innovation
would disappear? How much investment in that innovation would
disappear? I don't think any would disappear," Behlendorf said
C++ RegEx could be really useful. PCRE is a similar library that is in fink. Uses the Perl 5 syntax, which I do not really know all that well.

PlansCalendar might be a good group calendar. Flat file or SQL support. Built on PERL and CGI.

I tried to update acoc in my fink local to 0.7.0, but it does not work. I tried to get back to 0.7.0. After fink remove, scanpackages, index, and removing the deb and info file, it was still in the database. The trick was "fink purge". There was a config file that fink leaves there in case you modified it. Why does this line fail in 0.7.0, but not in 0.6.0?
   require 'term/ansicolor'
digest-sha1-pm584 refuses to install after building. Conflicting with the perl core. Arg. New pstoedit version. That should be help with writing the Santa Barbara paper.

Posted by Kurt | Permalink

02.01.2005 19:58

iSight 640x480 or 320x240?

Chatting with Deano:

This message says that the iSight is 320x240 on the chip and then expanded to pretend it is 640x480. But I can't find the footnote on the apple isight page anymore. So which is it and are there two iSight generations?

His recommendation for firewire cameras is the point grey dragonflies (1024x768). That is better than the isight. They also have the "Bumblebee" which does stereo.

DirectedPerception makes pantilt units. dp systems are steppers running open loop. Is there something lighter? Matt suggested heavier

Posted by Kurt | Permalink

02.01.2005 15:38

Santa Barbara oil seep

To contrast to white_dune, there is a new Coin based visual designer that I would like to try: CoinDesigner

Marine Geoscience Data Management System - thanks to Julie for this link.

There is a BBC article that talks about new areas of seepage:Oiled Birds Prompt Study By UCSB Experts If you want to see some upclose pictures from Aug 2004, take a look at the photos from the BPSIO cruise. They think storm change, but could this be to some local structural shifting?

Posted by Kurt | Permalink

02.01.2005 15:24

gnuplot tutorial

gnuplot in a box... open a terminal, shell, xterm, or something and start typing:
cat << EOF > 1.dat
   # depth value table for core 1
   1 2 
   2 2.2
   3 3.3
   4 1.4

cat << EOF > 2.dat # depth value table for core 3 1 2.9 2 2.1 3 1.3 4 2.4 EOF

gnuplot plot '1.dat' with linespoints, '2.dat' with lines title 'core 2' quit

cat <<EOF > sample.gp set terminal pdf set output "sample.pdf" set title "This is a test" set xrange [1:4] set xlabel "depth" set ylabel "some value" plot cos(4*x)+2, '1.dat' with linespoints, '2.dat' with lines EOF

gnuplot sample.gp open sample.pdf

Posted by Kurt | Permalink

02.01.2005 08:53

Pmag seminar - Snowball,Sandgren - Sweden 9000-450 BP

Before jumping into stuff from the paper, it looks like bloglines has picked up my blog. About 40 hits from cr03.bloglines.com in the last 4 days to my index.xml.

My biggest frustration with journal articles today is that I can not just click on a reference to a paper that I am reading and get the journal article. I also can not just click on a data link and get a tar ball of the data in whatever format the authors use, let alone a standard format.

Snowball and Sandgren, 2004. "Geomagnetic field intensity changes in Sweden between 9000 and 450 cal BP: extending the record of "archaeomagnetic jerks" by means of lake sediments and the pseudo-Thellier technique," EPSL, v227, p361-376.

  AF  - alternating frequency 
  AMS - C14 dating in this paper.  Not Anisotropy of Magnetic Susc.
  ARM - Anhysteretic remanent magnetization
  DRM - detrital remanent magnetization
  IRM - Isothermal remanent magnetization
  MAF - Median acquisition field
  MDF - Median destructive field
  MS  - Magnetic susceptibility
  PSV - paleomagnetic secular variation
  PT-RPI - pseudo-Thellier relative palaeointensity
  RPI - relative paleointensity

gyttja - A nutrient-rich sedimentary peat consisting mainly of plankton, other plant and animal residues, and mud. It is deposited in water in a finely divided condition.

They used an unoriented 5m rod operated piston corer with PVC liner with an inner diameter of 67mm.

I am a little confused as to what material was used for the AMS 14C dating. I do not know what happens if you just send off a bulk sample to a lab. They ended up discounting the AMS 14C dating in the end. Ended up matching secular variation to a northern core that has a varve count.

It is weird to see the magnetic susceptibility grow steadily with depth. That is the opposite that I see with my cores where they have the susceptibility grow right at the top of the core. I need to go read what butler says about that in his book. The difference must be that these are fresh water lakes.

I have not done a deep reading, but I do not see any work in the paper to look for strange anisotropy problems with AMS or other type. Bias warning in that my thesis is AMS (not C14) based. The data does look pretty good at first glance so that maybe that kind of check is not critical.

What is the definition of a "jerk." Technically a discontinuity in the 2nd derivative. So this is interpreted to be rapid changes.

The PT-RPI check in not in Lisa's paper as far as we can tell. Some normalized thing. Can't be NRM over ARM

No hysteresis loops. Snowball thinks that it is biogenic magnetite. Has the right AF-MDF. Could be mag-hemite.

The wiggle matches in figure 6 are really good. "Amazing!" was the consensus.

The declination data is tough with twisting. Why did it twist with a rod piston corer. The PVC can twist with uneven forces. Same is true with air driven piston systems. Confirmed by Jason and Cathy.

From Roi: Lake Lusan in Israel has a great core from 70ka to almost present that is not having paleo-intensity yet. Possible complications from magnetic material.

Nice record in this paper, but needs independent dating. With independent data this is exactly the kind of data that Cathy et al would put into their model (CALS7K). For error modeling, age uncertainty is turned into increased intensity and direction error do to secular variation.

GUFM goes back 400 years to 1600AD (Jackson et al. 2000).

Posted by Kurt | Permalink

02.01.2005 08:27

Wanted - backup utility

I was talking to Myche over the weekend about need for a simple backup program for my laptop and desktop. I was sitting around watching a movie that was not too exciting, so I thought, "Let me give this a try." Hah! It was not quite so simple. I got totally nailed by spaces and funky extended characters in filenames. Doh!

Here is basically what I want. Once a [configurable time period, say a week], I would like a disk image to appear on my desktop. In that file, I want as many of the files that changed in that week to be in that image so the disk is as full as possible. I would also like a list of the files and md5sums to be cached somewhere so that it is easy to find later what cds contain that file. I could then just say burn that image when it is convenient for me.

The details of that are not as simple, but not to crazy as they could be. It would be nice to be able to select compressed or non-compressed, in a tar or not, and a list of regular expressions describing what files are not allowed in the backup. It might also be good to allow multiple CD images. In the case of a single image, it should keep going back in time looking for files that still fit and adding them to the image. For example, if I have a 1GB file next that doesn't fit, it skips that moves to the next file, if it fits, add it, if not, keep looking for the next older file to add.

Another thing that bit me what trying to create a disk image. I managed to improperly eject an image. I had to manually delete a /Volume/test with rm -rf. That was scary. I then had to logout and back in for the interface to be happier.

This morning, I thought that maybe tar could do this since it has the exclude files using a file containing regular expressions functionality, but it does no know how to add starting with the newest file in its list. Tar has these options which might be helpful:
I think maybe I just need to check out bacula which is in fink. Maybe it already supports making Mac OSX disk images or that could be already added. It sounds like it supports many of my desires. But no time now. Must get ready for the GP seminar soon.

Posted by Kurt | Permalink