SDMG Notes for 17 May 2007

In attendance:

Terralyn Vandetta, Kelly Christiansen, Theresa Valentine, Heather May, Dave Hockman-Wert

Notes: Dave

Treats: Matt

Agenda:

1. Meetings:
  • Geospatial 2007: May 6th-11th. Website o Entertaining keynote speaker. Good schwag. Lots of interesting talks, which will be available online. Python panel was noteworthy; lots of people using GDAL, like Matt. Theresa was on the planning committee and thus had to work.
  • ESRI International User Conference: June 18-22 San Diego. Theresa and Heather going
  • Geoprocessing Group Meeting: Wednesday, May 23 from 10:00 - 12:00 location?
  • Society for Conservation GIS conference June 25-28 in Monterey, CA. Website
  • Open-source geospatial conference September 24-27 in Victoria, BC (Matt probably going).website
    2. Data Issues
  • Kelly is looking for Level 3 Ecoregions for Canada and the U.S.
  • Terralyn said that “data/fordata” will be moved to a PC server from UNIX.
  • If anyone still has data on 8-mm tape drives, it’s time to move it to something else. Machines capable of reading old media are not going to be around forever.
    3. Training needs/opportunities:
  • ArcServer training in July in Eugene: Keith, Theresa, Terralyn, and another Forest Service person are thinking of going.
  • ESRI Conference: send Theresa or Heather your questions/problems
  • Google Earth: need a training on KML and working with GIS data. Anyone know enough to lead this?
    4. ArcGIS 9.2 update/service packs
    ArcGIS v9.2 is available for installation on an individual basis. Service packs 1 and 2 are available.
    5. Round Table:
  • Kelly: Moving from USFS to OSU Forest Science! Also, in working with old ArcInfo workspaces in ArcCatalog, he discovered that ArcCatalog wanted to interpret the workspace as a point coverage, which led to some problems.
  • Dave: Wondered about NAIP image availability. Couldn’t find the latest year of orthoimagery for Tillamook County.
  • Jerry: College of Forestry Computing is thinking of moving computers out of FSL 231 to save money. USFS is charging about $8500 for that space. In the past, TRF funds from OSU were used to make these computers available for graduate students. Forestry Computing Group would have to pay now, which means costs would be spread out across all the groups.
  • Terralyn: With some of the new powerful computing techniques (clusters), it’s possible to generate huge volumes of data. For example, a cluster can run a model in 1.5 hours that generates so much data that it takes 9 hours to stitch it back together. This may be a netCDF limitation, so Neilson’s group is looking into other options.
  • Heather: Redesigning the GNN codebase, from C++ to Python/C++ combo.
  • Theresa: She heard from Mark Nay who was having trouble mosaicking TIFF images.
    5a. Field Trip!

    Terralyn led us down the hall to visit the cluster computing room. Mostly used by the Neilson group, there is an older cluster, called Vitis, consisting of 32 PC nodes networked together to allow for simultaneous processing. This cluster is mostly used for development now, because in the past few months, they purchased a nice new cluster with 12 sleek, stacked servers each containing two sets of duo-core processors, for a total of 48 compute nodes. The new cluster, named Lynx, has larger and faster servers than Vitis, so model runs that were taking a few days to finish can now be completed in a matter of hours.

    If anyone is interested in utilizing this kind of distributed computing power, it is possible to purchase a node to add to the existing cluster and then share available space when it is not being used. Contact Terralyn for more info.

    5. Next Meeting: June 28
  • Notes: Matt
  • Treats: Theresa
  • Speakers: ESRI Conference Report