Kurt's Weblog

http://signonsandiego.lamphost.net/pdfs/10242007_0500hrs_Evac_FirePerem.pdf

which is from http://sdcountyemergency.com/

Yesterday, from the roof of JPL's 168, it looked like we were on Titan rather than in Pasadena.



(I changed the levels a bit in photoshop, but that is all that I did.)

Kiley just forward this map to me. This is really bad for San Diego.

fire map

I am bummed that I will not be able to be there for this year's UNH GIS day. Last years event was a huge success and a ton of fun.

GIS Day at UNH

Life in SoCal is being hard on a lot of people. In the Pasadena area, we are fine, but things are not good and even here the air quality. wwdn describes it like being back in the 70's with the smog. I have to agree.

Santa Ana winds are making things a bit crazy around here. The air is a little bit smoky around JPL, but it's really not bad. Other areas are suffering the brunt of the trouble. Hopefully this passes quickly.

I've been trying to get a shot like this for a while. Finally pulled it off. This is a awesome skyline!

Using an lxml etree to walk a tree of 100k objects is decidedly a bad idea. It chews up a huge amount of memory and takes forever to get the tree loaded (after 15 minutes, I stopped checking). I am moving some of my code to the iterparse style of handling XML which is several order of magnitude faster. By controlling what tags get trees built, it runs in less than a second now. Here is a tiny.rksml example of the xml source:

<?xml version="1.0" encoding="UTF-8"?>
<RPK_Set>
<State_History>
<Node Source="Kurt" Time= "1192994832.9159949">
<Knot Name="SHOULDER_AZ" Units="RADIANS">-0.619</Knot>
</Node>
<Node Source="Kurt" Time= "1192994832.9159949">
<Knot Name="SHOULDER_EL" Units="RADIANS">-0.1231</Knot>
</Node>
</State_History>
</RPK_Set>

And here is the code that can quickly walk the parts of the tree that I need:

#!/usr/bin/env python
from lxml import etree
for event,element in etree.iterparse('tiny.rksml',tag='Knot'):
    print event, etree.tostring(element)

And the resulting output:

./iterparse.py
end <Knot Name="SHOULDER_AZ" Units="RADIANS">-0.619</Knot>
end <Knot Name="SHOULDER_EL" Units="RADIANS">-0.1231</Knot>

For my actual program, this does require me to do a lot more book keeping. There is no peaking backwards or upwards of the current element unless I cache what I need.

Turning Ubuntu into a GIS workstation [perrygeo.net] is an interesting list of applications to get people going on GIS with linux. His list is:

• Postgresql/PostGIS : a relational database with vector spatial data handling
• GRASS : A full blown GIS analysis toolset
• Quantum GIS: A user-friendly graphical GIS application
• GDAL, Proj, Geos : Libraries and utilities for processing spatial data
• Mapserver : web mapping program and utilities
• Python bindings for QGIS, mapserver and GDAL
• GPSBabel : for converting between various GPS formats
• R : a high-end statistics package with spatial capabilities
• GMT : the Generic Mapping Tools for automated high-quality map output

Where are we with fink and what is my take? Postgresl/PostGIS is in great shape. GRASS is there, but I haven't used it in 10 years. QGIS is quite a bit out of date. gdal is okay, but needs some tweaking. Proj and Geos are up-to-date with the stable versions. Python bindings for gdal are there (the old one, not the NG version). The python bindings for mapserver and QGIS are not there. gpsbabel is good. R (called r-base in fink) is up to the latest.

Mapserver still needs to get installed. Rob and I are also working on featureserver. mbsystem is on my list as I work in the world of multibeam sonar data and it also has some handy GMT helpers. pyproj is still one version out of date. I also use Coin3D/density to render 3D objects that I write out in VRML. Sprinkle in a bunch more python packages and I think that about covers where we are at right now. The work is never finished.

Today I got my little model of the Phoenix Lander to start moving in simpleview. I think my time with C++ is over for the moment and I can move back to the happier land of python. Programming with libxml2 and std::vector is just not the same as breezing through a little lxml python. I still have to face the fact that some of my coordinate frames are upsidedown and/or backwards, but I can fix with that.

This evening, I went along to dinner with a fun crew. Judy, Adam, Ivan, and Jenny. We hit up Barn Burner for a huge plate-o-bbq. The drinks came in mason jars and we all got properly messy.

On the way out the door I thought, "Why would I possibly need my camera this evening?" Well, I didn't realize that I would be eating next to the fire pit and giant chicken. So instead, I have a cell phone picture that is even stranger with low light than my point and shoot. Hopefully Jenny posts some of her pics from her real camera.

Another fun picture from yesterday evening... Sunset captured in a window reflection.

Development of a framework for Mapping European Seabed Habitats (MESH)

The MESH Project started in 2004 and is made up of a consortium of
twelve partners from five European countries led by the UKs Joint
Nature Conservation Committee (JNCC), with financial support from the
ECs INTERREG IIIB NWE Programme.  The MESH partnership draws together
scientific and technical habitat-mapping skills, expertise in data
collation and its management, and proven practical experience in the
use of seabed-habitat maps for environmental management within
national regulatory frameworks.
...
The website is designed to give easy access to the hundreds of project
outputs that include an interactive mapping page, a catalogue of
mapping studies, practical guides, tool kits, data templates and
technical reports. These items cover most aspects of seabed habitat
mapping from planning surveys, fieldwork standards, data handling,
predictive modelling, producing maps to how to communicate your
results.

Totday is Walter Munk's 90th birthday. If you are in any sort marine sciences, you have heard of Walter. He is one of the pillars of the community.

After 10 hours of computer graphics and C++ my brain went. Jason commented, "my that is pretty colored text that you are staring at." I had forgotten how much more work it is to do anything in C++ compared to the land of scripting languages. It was time to get off of the lab and get some exercise. I've always wanted to get a photo of the lab from below in the arroyo. And now I have it. The area below JPL is a strange frisbee golf course and a couple times a frisbee went by my head, but there was nobody in sight. The course has some corners that people have to shoot around.



And I love driving under this bridge along the arroyo back to where I am staying. This road hardly feels like it is in LA.



It was fun being out for a walk in the pre-halloween days!





The Rose Parade building is a huge building that I've passed a ton of times. Had to snap that.



And as sunset came, I headed over the 110 and had a tamale and a strawberry icy at the farmers market. Yum.



All-in-all, an excellent walk.

The days are getting shorter and starting later...

I tried to take a dramatic photo of JPL and I definitely over did it. Not very subtle. Kind looks like what happens to older pictures from the 70's.



This is one of the houses right near where I am staying that the full on crew working from dark-to-dark filming. Guess the threat of the writer's guild strick is pushing them hard.



Here is another shot of the same house above. This street corner was setup with huge spot lights last night.



The urban jungle...

2007-10-19 2PM Chase 130 Video Classroom

Right Whales in the Wrong Place: Numerical Model of a North Atlantic
Right Whale Ship Strike
    Presenter: Igor Tsukrov
    Abstract:
The North Atlantic right whale is one of the most critically
endangered whales in the world (<350 individuals), and is subject to
high anthropogenic mortality. Of the 39 animals examined post-mortem
(between 1970 and 2006) 20 deaths (51.3%) resulted from ship strike
trauma. To reduce the likelihood of fatal collisions, speed
restrictions are being considered for vessels traversing critical
habitat, however the effects of speed on collision outcomes have not
been specifically evaluated from a biomechanics perspective. The goal
of this on-going study is to provide quantitative data on the
mechanics of a ship-whale collision. To achieve this goal a numerical
model of the collision was developed to determine forces acting on the
whale during impact, and to predict the extent of damage sustained by
the animal as a result of collision. Numerical simulations were
conducted using the finite element method on two levels: numerical
analysis of the entire collision event and detailed modeling of the
whale jaw bone (mandible) to predict its fracture. Finite element
analysis was supplemented by mechanical testing performed on a right
whale mandible. The developed numerical model is capable of
considering variations in collision parameters including whale mass,
ship speed, hull size and shape, and whale orientation. It is expected
to be an efficient tool to predict mortality due to whale ship
strikes, and to inform management decisions regarding effective
regulation of vessel speed in the right whale critical habitat.

It's little strange that I heard about this through the gpmag-l mailing list, but I'm on the wrong coast at the moment and have been wrapped up in my research lately.

Solid Earth Geophysicist
.
The Department of Earth Sciences at the University of New Hampshire invites 
applications for a tenure-track position in solid earth geophysics at the 
assistant professor level starting August 2008 or thereafter.  The 
Department of Earth Sciences (http://www.unh.edu/esci) conducts research in 
geology, oceanography, atmospheric sciences, and hydrology, and offers 
Bachelors, Masters and PhDs in the Earth Sciences.  The successful 
candidate will be expected to teach geophysics, an introductory earth 
sciences course as part of the core curriculum in geology, and graduate 
course(s) in his/her specialty, and to develop a strong externally funded 
research program involving graduate and undergraduate students.  Research 
specialization is open, but applicants with interests in tectonics that 
complement current departmental efforts are particularly encouraged to 
apply.  The Department has strong ties to the Institute for the Study of 
Earth, Oceans, and Space (http://www.eos.sr.unh.edu/) and the Center for 
Coastal and Ocean Mapping (http://www.ccom-jhc.unh.edu/).  Starting salary 
will be commensurate with experience and qualifications.  A Ph.D. at the 
time of appointment is expected.
.
Review of applications begins October 22, 2007 and will continue until the 
position is filled. Please send complete CV, statement of research and 
teaching interests, and names and addresses of three references to 
Geophysics Search Committee, Department of Earth Sciences, University of 
New Hampshire, 56 College Road, Durham, NH 03824.  UNH is committed to 
excellence through diversity among its faculty and strongly encourages 
women and minorities to apply.

This just went out today. I will be at both meetings and presenting to the working group.

Radio Technical Commission for Maritime Services (RTCM) to hold AIS
Working Group meeting in conjunction with eNavigation Conference
.
As part of a U. S. Coast Guard research and development project, RTCM
has created a working group to study the use of AIS capabilities in
support of vessel traffic management. The Working Group on Expanded
Use of AIS Within VTS has been established to solicit input from
stakeholder groups, review existing standards, consider technical
implications of expanded AIS use and potentially develop new
standards. To make it convenient for maritime industry members to
attend both eNavigation 2007 and the RTCM working group meeting, RTCM
has scheduled the workshop for November 15th at Coast Guard Sector
Seattle headquarters, immediately following eNavigation.
.
For more information about the working group, please contact CDR Brian
Tetreault.

ExpressView (by LizardTech) works like a champ for viewing HiRise Mars images on the Mac. Here you can see that I've loaded up one of the "quick look" images that is about 700MB compressed. As you zoom and pan, the app works to fetch the higher resolution tiles. On the right is the overview strip that shows the whole image. This image is large. Here is an except from the PDS (Planetary Data System) header... PSP_001752_1750_RED.LBL

OBJECT = UNCOMPRESSED_FILE
    FILE_NAME    = "PSP_001752_1750_RED.IMG"
    RECORD_TYPE  = FIXED_LENGTH
    RECORD_BYTES = 56814 <BYTES>
    FILE_RECORDS = 58617
    ^IMAGE       = "PSP_001752_1750_RED.IMG"
    OBJECT = IMAGE
        DESCRIPTION                = "HiRISE projected and mosaicked product"
        LINES                      = 58617
        LINE_SAMPLES               = 28407

PSP_001752_1750 - Evidence for Water and Wind Processes in Gale Crater



The HiRise color images are fantastic!



Then I tried to get the Photoshop CS3 plugin for JPEG 2000 to work. Norton cause photoshop to implode to the point where I had to uninstall and reinstall the whole deal.

I've added JPEG 2000 support to gdal in fink using the JasPer library. Unfortunetly, that means that HiRise Mars images will not load. Sorry. The other libraries are all non-source based distributions and MrSID seems to be MS Windows only.

You may see something like this (or a segfault) when trying to handle large images. Here I tried to take a small subset of an image.

gdal_translate -of PNG -srcwin 10 10 1000 1000 PSP_005372_1515_RED.QLOOK.JP2 foo.png
Input file size is 32369, 95295
error: cannot decode code stream
gdal_translate(22843) malloc: *** vm_allocate(size=3084607488) failed (error code=3)
gdal_translate(22843) malloc: *** error: can't allocate region
gdal_translate(22843) malloc: *** set a breakpoint in szone_error to debug
ERROR 2: Internalize failed
ERROR 1: GetBlockRef failed at X block offset 0, Y block offset 0
libpng error: No IDATs written into file
Abort trap

Today is jpip day for me.

The main component of Part 9 is a client-server protocol called
JPIP. JPIP may be implemented on top of HTTP, but is designed with a
view to other possible transports. To facilitate its deployment in
systems with varying degrees of complexity, JPIP handles several
different formats for the image data returned by the server: these
include ordinary image formats, such as complete JPEG or JPEG 2000
files, and two new types of incremental "stream" that use JPEG 2000's
"tiles" and "precincts" to take full advantage of its scalabilty
properties. JPIP also supports both stateless and stateful modes of
operation, enabling sophisticated cache-modelling to eliminate the
redundant transmission of data.

I've whipped up some initial fink info files for FeatureServer, PythonPaste, and MoinMoin. I haven't really run these through any sort of tests, so I can use some feedback. Still to go is adding apache2 mod-available files so these things can be used with mod-python. I am pretty sure that these programs will not yet correctly find their cfg and the static data files. The cgi scripts are in /sw/lib/cgi-bin.

The info files can be had from:

http://vislab-ccom.unh.edu/~schwehr/software/fink/

I've been slow to pick up the camera this week. This morning, I ventured up the hill to get some coffee at Busters. It is currently at the top of the list for a place to sit for an hour or two.



Across the street in the little park where they do movies on the side of the building during summer, I heard a racket that sounded like parrots. Yup, parrots. Probably the same deal as in SF where escaped pets meet up.



Then down in the arroyo, I was wondering what terrible event had occurred until I realized that what looks like the entire fire department is out do ladder practice. I always enjoyed the fire drills on research ships when we get to try out the fire hoses.

Remko has created info files for GMT 4.2.1 in fink. 4.2.0 now lives in stable. Also look for gshhs-{lores,hires} to get the coastline data.

BTW, if density's simpleview in fink craps out (gets stuck on load), you need to do

fink rebuild soqt25
fink rebuild density

Seems that changes in the QT library are not binary compatible.

I'm using simpleview to aid development of some geometric code and finally hit the breaking point with simpleview having a really lame starting configuration. A little searching got me the tricks on how to set the window size for SoQtExaminerViewer and SoCamera knows how to "viewAll."

Setting the window size at startup can be had by tweaking the QWidget that SoQt::init() returns.

#include 
...
  myWindow->resize(a.width_arg,a.height_arg);

Where are "a" contains my command line args from gengetopts.

Then for the camera view, it is easy enough to ask the camera to view the whole volume. "si" is a small class that has all of my scene info.

  si->camera->viewAll(si->root,myViewer->getViewportRegion(),10);

It is nice to finally have that done. Two items that have been on my "to do" list for 3 years.

Last night, I dropped by the SoCal Piggies (Pythhon Interest Group). As usual, python folks (me included) always get excited talking about python. Andrew and Thomas were there talking about MoinMoin. The discussions went on for 3 and 1/2 hours. Quite the marathon.

They gave a great overview of MoinMoin, both where it is currently and where it is going. I tried to create a fink package of moinmoin, but didn't quite get through it. I've only experienced MoinMoin from the user side, so there is a little bit to get figured out for setting up a server.

They pointed out a useful website that compares wikis:

WikiMatrix

Agency AwardNOAA | Digital CoastMapping the laws [gcn.com]

...
Digital Coast was built primarily using ESRIs suite of geographic
information system tools. But the intended primary audience for the
atlas is resource managers, most of whom have little or no experience
with desktop GIS tools. So developing programming tools that could
move the application to the Web was critical. Development of Googles
Keyhole Markup Language for displaying geospatial data on the Web,
along with tools for converting data from ESRI format to KML, were
especially important.
...
"It was an eye-opening experience for us, said Cindy Fowler, GIS
integration and development manager. As we started uncovering the
laws, we realized pretty quickly that all of these things wouldnt hold
up [against] the modern mapping standards."
...

Sometimes I get some highly amusing emails that go beyond the usual work chatter. And sometimes the are even real and not just telling me that I won the lottery in Turkmenistan. I wish we had a web cam on the roof!

Folks,
.
I've just been informed that the blasting on the new building site is
about to start --- in a little under an hour, in fact.
.
We don't know how much noise/vibration this is going to make, but you
should expect to see and feel it somewhat.  You'll probably hear
sirens before the actual blast itself.
.
I would suggest that you make sure that anything particularly
sensitive is saved to your hard-disc in case we have a power failure
as a result; it would probably be best to stay away from the windows
on the side of the building nearest the site.
.
I'll let you know more as I hear it.
.
Cheers,
Brian.

I'm seriously missing out. Maybe I'll get an earthquake to make up for not being there. Haven't felt a good one of those in a lot of years.

I just released a new version of density. It is a Coin3D / SimVoleon based set of tools for working with 3D polygon and volumetric data. This release fixes the offscreen render wrapper script (which apparently is not being used by anyone other than me). I also fixed some minor bugs in the fink packaging.

I'm running into troubles with offscreen rendering on the mac. It seems to crash the Apple X11 server. I was going to make and post a little movie using "render," but that won't happen until I figure out the bug.

ecmascript has finally bubbled high enough up my priority stack that I am working on learning javascript and ActionScript 3. Having Flash bundled into CS3 Web Complete can only be a good thing. I am also looking at Adobe AIR (formerly Apollo) and Flex.

There is a ton of potential with these platforms and Javascript is hiding in all sorts of important corners (openlayers, web dnd, etc).

A couple people have told me about these news items. I have been disconnected from my normal news sources lately, so I appreciate the pointers.

Whale death attributed to ship strike [latimes - 23-Sep-2007]

The blue whale found dead last week in the Santa Barbara Channel was
probably the third victim of a ship collision in two weeks, scientists
said Saturday as they conducted a post-mortem on the 60-ton creature.
.
As surf roiled around the massive carcass on a beach at Point Mugu,
biologists cut doorway-size openings in its belly and probed its
organs for tissue specimens. About an hour after they had peeled back
4-inch-thick sheets of snowy blubber, they started removing foot-long
bone fragments that had chipped away from the animal's 9-foot
ribs. Later, they found other fractured bones, including a smashed
cranium.
.
"It's definitely a ship strike," said Easter Moorman, a spokeswoman
for the Santa Barbara Museum of Natural History, which was directing
the necropsy of the not-quite fully grown male. "The animal died
instantly."
.
Three blue whales -- members of the largest species on Earth -- have
been discovered dead off Southern California in the last two weeks,
the most recent on Wednesday. Two found in the Santa Barbara Channel
were thoroughly examined by scientists, who concluded that they were
hit by ships. One found in Long Beach Harbor was towed out to sea, but
a biologist who viewed it said it probably had been hauled into port
on a ship's bow.
...

And

General Information About the Third Whale [sbnature.org]

General Information About the Third Whale:
.
    * ID Tag: SBMNH 2007-20
    * Stranding Location: Near Platform Gail and floated to south of Platform Grace
    * Size: About 60-65 feet long
    * Gender: Male (subadult)
    * Cause of Death: Ship Strike
    * Necropsy: Completed September 22, 2007
.
In September, There Have Been Three Blue Whale Strandings off the
Southern California Coast:
    * September 8, 2007: Long Beach Harbor
    * September 14, 2007: Ventura County, Hobson County Beach
    * September 21, 2007: Ventura County, Beach TBD

And finally... Emergency Ship Speed Limits Sought to Protect Blue Whales [environmental news service]

SAN FRANCISCO, California, September 25, 2007 (ENS) - The Center for
Biological Diversity today formally petitioned the federal government
to set speed limits for ships in the Santa Barbara Channel off
southern California to protect endangered blue whales.
.
At least three dead blue whales have been documented in southern
California over the past two weeks. The Santa Barbara Museum of
Natural History, which conducted analyses of the dead whales,
determined that ship strikes caused all three of the deaths.
.
It was reported the week of September 9 that there were about 100 blue
whales in the Santa Barbara Channel and that their primary food
source, tiny crustaceans called krill, was prevalent in the shipping
lanes.
.
Most large vessels plying the Santa Barbara Channel are heading to or
from the ports of Los Angeles and Long Beach. Several thousand
container ships transit the channel each year.
...
Museum scientists have a "working theory" that domoic acid, a toxin
associated with certain algal blooms, may be a contributing factor to
the whales' deaths.
...
"Whether the blue whales are being disoriented by military sonar,
toxic algae or something else entirely, what is actually killing them
is speeding ships," said Brendan Cummings, oceans program director for
the Center for Biological Diversity. "The single most effective thing
we can do to protect blue whales is to slow down large ships."
...

In addition to spending a large portion of the day reading about ldap, Jeff gave me a tour of the Sonora Desert Museum. Is is a quick trip from Tucson west along Speedway and up through Gates Pass.



There is a great section of butterflies. However, no hawk moths (Jeff current research subject).



There were some seriously crazy cacti in the park.



This is not a pig, but a javelina. These critters with huge teeth are native to the United States unlike pigs.



As cute as this cat is, it is certainly no house cat. The paws look bigger than my hands.

I finally was able to get some LDIF files to import into openldap. First off, I am using openldap23 from fink:

fink list -i openldap
Information about 6680 packages read in 0 seconds.
 i   openldap-ssl-shlibs       2.1.29-1025         Shared libraries for LDAP
 i   openldap23                2.3.27-10           LDAP directory services implementation
 i   openldap23-dev            2.3.27-10           Libraries and headers for LDAP development
 i   openldap23-shlibs         2.3.27-10           Shared libraries for LDAP

I ran into a number of stumbling blocks that were not covered in the docs that I have been reading.

First, I had to edit /sw/etc/openldap/slapd.conf. I edited the includes at the top have add cosine and inetorgperson schemas:

include		/sw/etc/openldap/schema/core.schema
include			/sw/etc/openldap/schema/cosine.schema
include				/sw/etc/openldap/schema/inetorgperson.schema

Then in the database at the bottom of the slapd.conf, I setup a dummy schwehr.org domain.

database	bdb
suffix			"dc=schwehr,dc=org"
rootdn				"cn=Manager,dc=schwehr,dc=org"
# Cleartext passwords, especially for the rootdn, should
# be avoid.  See slappasswd(8) and slapd.conf(5) for details.
# Use of strong authentication encouraged.
#rootpw	 	secret
rootpw			{SSHA}somelargehash



# The database directory MUST exist prior to running slapd AND 
# should only be accessible by the slapd and slap tools.
# Mode 700 recommended.
directory  /sw/var/openldap-data
# Indices to maintain
index	  objectClass	eq

Replace somelargehash above with the results of "sudo slappasswd". I then had to cleanup from previous slapd instances. First do a ps -aux | grep slapd and kill slapd if it is running. If things are well, you can do:

kill `cat /sw/var/run/slapd.pid`

Remove all previous database info. WARNING: If you have a real database, this will destroy all your data!!!!

cd /sw/var/openldap-data
sudo rm -f log.0000000001 alock __db.00* *.bdb

Make sure that you have a DB_CONFIG file.

cp DB_CONFIG.example DB_CONFIG

Now check your config file.

slaptest -f /sw/etc/openldap/slapd.conf

I first tried adding both the root and the people at the same time. This did not work. The combined file:

cat >> EOF > ~/people.ldif
## Build the root node.
dn: dc=schwehr,dc=org
dc: schwehr
objectClass: dcObject
objectClass: organizationalUnit
ou: Schwehr Dot Org
      
## Build the people ou.
dn: ou=people,dc=schwehr,dc=org
ou: people
objectClass: organizationalUnit
EOF

But that gave me an error. Why? I don't know.

slapadd -v -l ~/people.ldif 
bdb_db_open: Warning - No DB_CONFIG file found in directory /sw/var/openldap-data: (2)
Expect poor performance for suffix dc=schwehr,dc=org.
str2entry: entry -1 has multiple DNs "dc=schwehr,dc=org" and "ou=people,dc=schwehr,dc=org"
slapadd: could not parse entry (line=12)

Then I tried each piece separately and it worked!

% cat >> EOF > ~/people1.ldif
## Build the root node.
dn: dc=schwehr,dc=org
dc: schwehr
objectClass: dcObject
objectClass: organizationalUnit
ou: Schwehr Dot Org

When added gives:

% slapadd -v -l ~/people1.ldif 
added: "dc=schwehr,dc=org" (00000001)

And the second part with the people entry:

% cat >> EOF > ~/people2.ldif
## Build the people ou.
dn: ou=people,dc=schwehr,dc=org
ou: people
objectClass: organizationalUnit

That results in:

% slapadd -v -l ~/people2.ldif 
added: "ou=people,dc=schwehr,dc=org" (00000002)

Start the ldap server:

% slapd -f /sw/etc/openldap/slapd.conf

I used gq to browse the database:

gq &

And finally a slapcat to look at the contents of the database:

slapcat
dn: dc=schwehr,dc=org
dc: schwehr
objectClass: dcObject
objectClass: organizationalUnit
ou:: U2Nod2VociBEb3QgT3JnICAgICA=
structuralObjectClass: organizationalUnit
entryUUID: ed352594-0984-102c-9c76-7977ca58b6da
creatorsName: cn=Manager,dc=schwehr,dc=org
modifiersName: cn=Manager,dc=schwehr,dc=org
createTimestamp: 20071008005537Z
modifyTimestamp: 20071008005537Z
entryCSN: 20071008005537Z#000000#00#000000



dn: ou=people,dc=schwehr,dc=org
ou: people
objectClass: organizationalUnit
structuralObjectClass: organizationalUnit
entryUUID: 0a9635a6-0985-102c-89af-21c7102a7e23
creatorsName: cn=Manager,dc=schwehr,dc=org
modifiersName: cn=Manager,dc=schwehr,dc=org
createTimestamp: 20071008005626Z
modifyTimestamp: 20071008005626Z
entryCSN: 20071008005626Z#000000#00#000000

I'm looking into openldap and reading LDAP System Administration from Safari. I had started with How_to_setup_and_maintain_OpenLDAP_server_for_your_network on fedoranews.org and got a bit stuck.

From there, I found MigratonTools. It is a perl package with scripts to read user information from NetInfo and create export files (ldif) for ldapadd. However, the script does not seem to work. It will take a little tweaking once I know more to get it to be happy.

Once I have a demo ldap server (slapd) running, then I can do things such creating a fink package for python-ldap

While reading the ldap book I ran into this from /sw/etc/openldap/schema/cosine.schema...

Pardon me?

# 9.3.5.  Favourite Drink
#
#  The Favourite Drink attribute type specifies the favourite drink of
#  an object (or person).
#
#    favouriteDrink ATTRIBUTE
#        WITH ATTRIBUTE-SYNTAX
#            caseIgnoreStringSyntax
#            (SIZE (1 .. ub-favourite-drink))
#    ::= {pilotAttributeType 5}
#
attributetype ( 0.9.2342.19200300.100.1.5
        NAME ( 'drink' 'favouriteDrink' )
        DESC 'RFC1274: favorite drink'
        EQUALITY caseIgnoreMatch
        SUBSTR caseIgnoreSubstringsMatch
        SYNTAX 1.3.6.1.4.1.1466.115.121.1.15{256} )

In contrast to working on ldap and openlayers stuff much of today, Jeff and I took a hike up in Sabino Canyon in the late afternoon. It is an amazing spot totally unlike another other place I've been.



Jeff pointed out some crazy big grasshoppers. The same ones Kiley had sent me pictures of.



Even wilder was this guy who was just sitting in the road.



And there was a great sunset.

"I just need to dehydrate a brain."

I took a couple mile walk today and got some good pictures of the Tucson foothills.

Steve showed me a trick on how to make ellipses and circles in Photoshop to highlight features in images. I had been using Illustrator, but it is better to do that in Photoshop and preserve the image size.

First make a new layer and select it in the layer window.

First switch to the Elliptical Marque Tool in the Tools window.



Select a region. Here I am highlighting the AC unit on top of some random house in Tucson. The highlight edges are where the ellipse will go.



Now select Edit -> Stroke.



Pick your color and how many pixels thick the line will be.



Hitting okay will reveal the ellipse.



Now to may a semi-transparent filling by doing Edit -> Fill. Make the Opacity lower. Here I used 20%. Select Use: Color (or reselect it) to get the color picker to pop up.



I now have a shaded and highlight region.



Here is using the Magic Wand tool to outline/highlight an irregular region.



Here is after changing the Stroke and Fill.



And the final result.

I took a walk this morning around the Tucson Spacecraft Operation Center (SOC). The weather was perfect and not many people were out and about at sunrise. The mountains surrounding the Tucson area are really spectacular. I don't think this picture does them justice.



And I had to take yet another image of the mural on the side of the SOC. The image is a good illustration of sperical lense distorion.



The road to nowhere...

During the recent ORT, I had to pull off a quick movie under tight time constraints. All of my usual routes of making movies were out: no VNC moves, scan converters, xvidcap, Jing, CamStudio, etc. This was on a linux box that was locked down (with respect to the time frame). What to do? I came up with an X11 solution. Time to go old school. I had the ability to slow down the 3D rendering and had imagemagick on the machine, so here is my solution complete with a little python...

The key was the X Windows Dump (xwd) program that has been around for ever. It is not elegant, but it really can get the job done. The first task was to get the 3D graphics X11 window id. This helps to constrain the data volume such that I am not capturing all 2kx1k pixels on the display. It works something like this (which is a quick dump on my Mac).

xwininfo
xwininfo: Please select the window about which you
          would like information by clicking the
          mouse in that window.

I then clicked in the window (here just an xterm):

xwininfo: Window id: 0xa0000e "xterm"



  Absolute upper-left X:  635
  Absolute upper-left Y:  291
  Relative upper-left X:  0
  Relative upper-left Y:  22
...

The key is that "Window id". Now I could ask xwd to grab that window id. But I want to create a sequence of numbered frames. I could do this in bash, perl, python or what ever, but Steve taunted me about python as he is a very good perl programmer. Therefore, python it was...

#!/usr/bin/env python
import os,sys
for count in range(500):
    filename='/tmp/foo%03d.xwd' % count
    print filename
    os.system("xwd -id 0x2800004 -out "+filename )
    os.system('sleep .2')

That is definitely just a hack (like not calling time.sleep()) and I have since created a complete program with an extensive command line interface, but it serves the purpose of illustration. Running that dumps a whole mess of xwd images into /tmp. From that I needed to make an animated gif that is a subset of the original frames. There were parts of the display that I can't release. The imagemagick convert program does it all. It will crop and make an animated gif with a specified delay between each frame.

convert -crop 640x480+303+80 -delay 100 /tmp/foo???.xwd /tmp/movie.gif

And that's it. I loaded the gif into firefox and had a little movie. The results looked better when Steve put the frames through QuickTime pro, but the gif did get the job done.

And that is a nice little old school hack. Like I thought that I would ever pull out xwd or xwininfo again.

For this ORT, I tried to use the Aqua version of Open Office on a MacBook Pro. OO seems capable, but I hit a number of snags and I will absolutely have to get MS Office on this computer. First off, when OO saved a power point file out, it stuck symbol garbage in front of each numbered bullet. The numbers are then in PowerPoint, but I don't know how to get at the symbols and remove them. Then there were the crashes. I got a batch of them while editing a slide in the middle of a meeting. Then the final trouble was that I had trouble with drag-n-drop (aka DND). The big OO root window gets in the way and the document often (did it ever?) open with a dnd event.

However, in favor of OO, the progress is quite substantial when compared to when I last tried it a couple of years ago. A lot of the engineering team uses OO on the linux boxes in the SOC with good results.

Operational Readiness Test number 6 (ORT-6) is done for my team. The sequencing teams still have a couple days left and I have a couple tasks tomorrow here in the Tucson SOC (Spacecraft Operation Center). For my roles and the SSV team tasks, things went extremely well. It is great to see a lot of familiar faces across the teams from MPF (that would be Mars Pathfinder), MPL (Mars Polar Lander), MER (Mars Exploration Rovers), and many field tests.

I just realized that I had not started blogging during MER. My first blog entry was in 04 December 2004, almost 3 years ago. I stopped working full-time for MER in mid-May of 2004.

Here are some pictures from around the ORT.

Steve and Paul demonstrated how much fun IPRW is. Chris did a bang up job rewriting and improving the MER tool chain. Progress moves forward.



Here is a view of the press area. Don't expect this space to even be recognizable from the picture come mission time.



Here is proof that I was actually here. You might see me in the background of other peoples pictures, but I don't often get a chance to stop to be in pictures.



And it's not all work. Kiley, Jeff, and I went one evening to go bowling. The 50 cent beers were, well... um. They were in little dixie cups. We had to wear wrist bands that said "Ok For Alcohol." That just sounds strange. Pretty funny!

I was hoping to find some background info on wikipedia on iron oxides, but ran out of luck because of not knowing how to spell the mineral.

Hematite is there:

http://en.wikipedia.org/wiki/Hematite

There is an overall list of minerals:

http://en.wikipedia.org/wiki/List_of_minerals

Then I found:

Goethite

The project is talking lot about blogs that cover the Phoenix Mars Lander. I found one myself that caught my eye:

Junkbot Phoenix Mars Lander Launches with Ancient RISC Processor Aboard

...
As reported by Colleen Taylor at Electronic News, the brains of the
Phoenix Mars Lander consists of a BAE, formerly Lockheed Martin,
formerly IBM Federal Systems RAD6000 computer, which has a fascinating
heritage that leads directly to the original RISC processor, John
Cockes IBM 801 project. The purpose of that project, started in 1974,
was to create a processor for a programmable telephone switch. It
represented a backlash against extreme CISC (complex instruction-set
computer) as epitomized by IBMs 360 computer architecture. The problem
with the extreme-CISC design approach was that it required that the
machines clock be slowed to accommodate a lot of logic in each stage
of the CPU. The telephone-switch processor needed about 10 MIPS when
the best IBM 360 implementation could only achieve about 2 MIPS.
...

Chalk another major item on the to do list. SIM just released Coin 2.5.0. I would really like to spend some time checking out the new SoGeo nodes. It will probably be a while before I have time to get a fink package setup and tested.

The announcement:

Coin 2.5.0 Released
===================



Systems in Motion is proud to release v2.5.0 of the Coin 3D graphics
library.  Coin 2.5.0 is a minor release.  It extends the functionality
of previous versions of Coin 2.* in upward compatible ways.  It also
fixes bugs and problems found in releases 2.0.0 through 2.4.6, and
includes updates to the documentation.  The API and ABI of the library
have been considerably extended for this version, but not in any way
that could conflict with older versions of Coin 2.*.



Coin 3.0.0 is expected to be released on april 1st 2008, and we will
aim for minor or major releases every 6 months, with random micro-
releases for the usual bugfixes inbetween when/if deemed necessary.






NEWS File Excerpt:
==================
* new:
   - geo nodes: SoGeoOrigin, SoGeoCoordinate, SoGeoLocation,
 SoGeoSeparator
   - shader nodes: SoVertexShader, SoFragmentShader, SoGeometryShader,
     SoShaderProgram, SoShaderParameter
   - shadow generation: SoShadowGroup, SoShadowCulling, SoShadowStyle
   - new camera node SoFrustumCamera
   - new node SoTextureCubeMap
   - a lot of new SbVec and SbBox classes
   - a lot of new field classes (mostly Vec and Box fields)
   - implemented stubbed SoSimplifyAction,
     added SoGlobalSimplifyAction and SoShapeSimplifyAction as stub classes
   - SoReorganizeAction improvements
   - SoFile:: setSearchOK(), getSearchOK()
   - SoGLRenderAction::setSortedObjectOrderStrategy()
   - SbTypeInfo.h template type information library

Right now we are in the middle of Operational Readiness Test #6 (ORT6) for the Phoenix Mars Lander. Things are going well and there are lots of good work going on. There are quite a few wikis going within the project helping to document things and the project has been talking about lots of digital media technologies.

I got excited when people started talking about digg, slashdot, live blogging (err... what I am doing at the moment), gizmodo, del.icio.us, seeding youtube, etc. Woohoo! The world I normally live in and problably a lot of the other engineers on the team. The SpaceGrant program is thinking about facebook. This stuff really helps build a commnity both inside and outside of the spacecraft team.

I love working on forward looking projects. The immediacy is such a kick.

Check out http://phoenix.lpl.arizona.edu/, where people on the project are blogging.