W Randolph Franklin home page
... (old version) Login


This page is for WRF's notes of possible interest to the whole research team. (Note for individuals will be put on that person's wiki page in a box.)

Many of the wiki pages have not been updated in years. (As of 5/19/14, there are 1257 wiki pages. You can see the last modified date in the footer.) I mention them here because they are still worth reading.

Pages in this group

GeoStar2014 /
Everyone
GeoStar2014
MarcusAndrade
RecentChanges
Thrust3DGrid

Announcements

  1. Conference: SIAM GD17, Pittsburgh July 10-12. I was invited to give a minisymposium talk.
  2. Conference: S3PM-2017 :: International Convention on Shape, Solid, Structure, & Physical Modeling, Berkeley June 19-23. Marcus and I attended this last year. For the ACM Symposium on Solid and Physical Modeling (SPM) abstracts are due Feb 18, and papers a week later. For Shape Modeling International (SMI), abstracts are due March 28 and papers a week later. This is an excellent conference. We should definitely submit.
  3. Calendar. It now has my summer schedule. (6/6/16)
  4. GeoXeon describes the computer.
  5. My copies of papers from conferences like http://geometrysummit.org/ are on my dropbox account here .
  6. You both are excellent in parallel computing, both OpenMP and CUDA. There is a great interest in parallel computing in the GIS community, but no one seems to know much about it yet. So, think of ways to exploit your temporary advantage. E.g., write papers.
  7. I deleted Cookbook/MarkupExtensions when updating pmwiki. The old version caused thousands of warnings. The new version doesn't work. I don't think it's used. However please report any problems. (2015-09-24)
  8. This paper Geometric Modeling Problems in Industrial CAD/CAM/CAE from a 2007 SIAM Geometric Design conference might be interesting. It mentions the need for parallel algorithms and says that meshing is important. (2015-09-24)
  9. Google has a new compression algorithm that might be worth looking at. (2015-09-24)
  10. FWCG 2015: submission deadline is 2015-09-25. (2015-09-02)
  11. I'd like to propose to GRASS to include our best SW. This would provide great publicity.
    1. Is this ok?
    2. Do you have the time to do this? (I don't know what support GRASS can give.) (2015-09-02)
  12. I encourage people to submit to the ACM SIGSPATIAL GIS PhD program. (2015-09-11)
  13. I'm probably in town until FWCG. Let's meet thrice weekly: Mon and Fri at 3pm and Wed at 1pm. (2015-09-11)
  14. Think about the programming contest. It would be fun to win it. That would also demo the validity of our methodology. (6/1/15)
  15. I'm planning a group visit to Jianting Zhang at CCNY. Does Wed April 15 look good? Everyone is welcome, and I encourage everyone to come. This would be a day trip, either driving or taking the train.
    I'll forward our recent email, which describes his interests.
    We might also talk to Peter Braß (aka Brass). When Barb Cutler and I ran the Fall Workshop in Computational Geometry here at RPI a few years ago, he was an invited speaker.
    He wrote an excellent book, Advanced Data Structures, from which I'm learning things. I encourage you to look at it.
    The purpose of this visit is to consider possible future collaborative research proposals and projects.
    Please research the people that we'll be visiting.
    Jianting asked me to give a talk. I proposed that Wenli and Salles give a joint talk instead. You can repurpose the several talks that you gave at conferences last fall.
    (3/28/15)
  16. Wenli got graphics working on geoxeon as follows:
    "First I removed CUDA and nvidia-*, and tried to install CUDA with the accompanying driver. However, due to some error that I forgot, the driver won't install. Then I installed nvidia-current and CUDA without the driver, but CUDA 7 requires driver 346 to run. I downloaded the driver from http://www.nvidia.com/Download/driverResults.aspx/82252/en-us (which is now dead), and I had to select 'No' in one of the on-screen instructions to install it. Then everything just worked."
    Thanks. (3/28/15)
  17. The file /etc/motd is displaying whenever anyone logs in. 'motd' = 'message of the day'. (3/24/15)
  18. I added the CUDA bin (/local/cuda/bin) and lib (/local/cuda/lib64) dirs to the geoxeon system default (/etc/environment and /etc/ld.so.conf.d), so you don't need to specify them. (3/4/15)
    No. You still have to include /local/cuda/bin in your path. /etc/environment is not always processed.
  19. xelatex is a newer version of pdflatex that does fonts and accents better. I'm switching to it when possible. (2/10/15)
  20. NGA goes open source with a public geospatial tool kit https://github.com/ngageoint/mrgeo. It uses Hadoop. (1/28/15)
  21. I added a swap partition to geoxeon, to prevent the system crashing if if it runs out of memory. However if any processes use it, performance, especially interactive performance, will fall off a cliff. E.g., do not start 4 simultaneous processes each using 50GB. Start 2 and hold the others until those 2 finish. There are tools that do this automatically. (1/28/15)
  22. ZFS snapshots should now be working; see /tank/home/.zfs/snapshot/ . (1/23/15)
  23. AFS is now working again; see /afs/rpi.edu/home etc. (1/23/15)
  24. My reason for mentioning these conferences is for you to check out their programs to learn about current hot topics, to keep your own research current. (12/11/14)
  25. The 2014 SuperComputing conference looks interesting. (12/11/14)
  26. For Dec 8-9, I was at the UCSB Spatial Search Specialist Meeting. This was more information and library science than relevant CS. The attendees' position papers are on the site. (12/11/14)
  27. Two relevant professional societies for terrain are the American Geophysical Union (AGU), conference, and the European Geosciences Union (EGU), conference. Both meetings are very large. The societies have several open journals that may be relevant, both for reading and for submitting to. You might consider checking them out. (12/11/14)
  28. Increasing public visibility for our publications: (11/20/14)
    1. We should each try to use the same form of our name on all our published material.
    2. We should create entries in google scholar, dblp, etc.
  29. My Advice To Doctoral Candidacy Examinees (9/30/14)
  30. The NSF Interdisciplinary Research in Hazards and Disasters solicitation that I'm trying to put together a proposal for is here: http://www.nsf.gov/publications/pub_summ.jsp?WT.z_pims_id=504804&ods_key=nsf14581 (9/29/14) .
  31. New Hutchinson ANUDEM paper (9/17/14)
    http://www.geomorphometry.org/Hutchinson2013
    The full text is online free. Although I think that hydrography is mature (i.e., fewer interesting things to do), you never know.
  32. Interesting: (9/8/14)
    1. https://www.sites.google.com/site/digitalandscapes/
    2. http://vterrain.org/
    3. http://www.geomorphometry.org/
    4. Advances in Digital Terrain Analysis by Qiming Zhou, Brian Lees, Guo-an Tang, especially the chapter by Peter Shary.
  33. Updated meeting times: Mon 2-3:45, Thurs 3:10-3:45. (9/3/14)
  34. FWCG'14, the 24th Annual Fall Workshop on Computational Geometry will be at the University of Connecticut (Oct 31-Nov 1, 2014). Submissions due Oct 3 on https://easychair.org/account/signin.cgi?key=15757378.uhaYCpUAk2Gk4MGp (9/3/14)
  35. Making a poster My old file doesn't work properly now. However lots of ideas are here: http://tex.stackexchange.com/questions/341/how-to-create-posters-using-latex (8/23/14)

On Geoxeon, my files of possible interest to you

are in /opt/wrf/

  1. /opt/wrf/union3/
    1. Program to find the volume and area of the union of many random cubes.
    2. Demonstrates uniform grid and my local topological formulae.
    3. Runs 10x faster (clock time) using 32 OpenMP threads on geoxeon.

Current members

  1. Dan Benedetti
  2. Wenli Li
  3. Salles Maghalães

History

GeoStar was the name of the DARPA program, of which I was one of the six (later 5) projects. The name GeoStar was my idea.

GeoStar is the wiki group aimed at DARPA to describe the project. This is rather complete, and is an excellent place to start.

GeoStar2 is the internal wiki group, for RPI people working on the DARPA project.

Although the DARPA project ended, it is convenient to maintain the organization. GeoStar2009 has some updated research stuff.

GeoStar2014 ? is this page, with generally interesting internal research material.

Upcoming deadlines

(none)

Sites of possible interest

File formats

Overlay

It reads maps in the Harvard Odyssey format. It's easy to use, but it seems that no one else uses Odyssey. This means that we can't find data in that format, and can't find programs that convert. That's not an enormous problem, but is a little annoying.

The ARC section of ESRI's E00 format appeared to be the same as Odyssey, but E00 appears to be deprecated, and SHP files are not even topological.

By topological, I mean that the common boundary (arc or polyline) between two polygons of a map is stored only once. That prevents inconsistencies and slivers between adjacent polygons. However it opens the possibility of problems at the nodes or vertices where three polygons meet.

Misc

  1. https://en.wikipedia.org/wiki/GeoJSON
    This looks interesting because it is simple and readers exist.
    It is not topological. Its topological extension, TopoJSON, appears not to be widely used.
  2. https://en.wikipedia.org/wiki/Geography_Markup_Language
  3. http://avce00.maptools.org/docs/v7_e00_cover.html
    ESRI's proprietary topological format. Readers and writers exist. ESRI has deprecated it in favor of SHP files.
  4. ESRI SHP
    Not topological.

SW

  1. GRASS
    Lots of packages. Open.
    We should benchmark our results against GRASS whenever possible. I expect us to be considerably better.
  2. ESRI ArcGIS.
    The industry standard, and so can't be ignored.
    I don't have a current version, but can get it if there's interest.
    We should benchmark our results against ArcGIS whenever possible. I expect us to be considerably better.
  3. OGR simple feature library
    http://www.gdal.org/ogr/
    C++ code, reads many formats.
    Part of GDAL.
  4. Mathematica
    It can import SHP files, http://reference.wolfram.com/mathematica/ref/format/SHP.html
    It makes beautiful pictures.
    Beware the BS component. Sometimes there is less than meets the eye with Mathematica. E.g., the most powerful interpolation functions do not describe their algorithms, are described as being only for making pictures.

Interesting sites

  1. Smallworld Technical Paper No. 1 - Ten Difficult Problems in Building a GIS Richard G. Newell & David G. Theriault. Old but still interesting.