I have ideas about all of these issues, and probably know more about the history of our explorations than most people. What I lack is a lot of the important technical knowledge that SOMEbody needs to have and manage and distribute to those who need it.
Here are some of the issues:
License servers:
1. We need to upgrade ArcIMS to 4.0.1 and install the dongle. This means getting to ims2, in the Tucker machine room. At the same time, we need to ascertain the adequacy of the existing security of ims2, in the light of the preconference seminar I attended. Here are the points at issue, most of which Jit thinks we probably have covered:
- set up a firewall
- install ArcIMS on a standalone system
- install only the OS, Web server, and JRE components
- disable all non-required services [there's a list in the preconference materials]
- decrease the number of privileged programs
- ser IUSR_ims2 user privileges
- set file permissions on directory structures
- delete/move all sample Web sites
- deploy intrusion detection
2. ims1 needs to be physically moved and relocated from .192 and all that's really required is that the machine be always ON. Perhaps the room off of Parmly 307 would be an appropriate location?
3. Should the ArcINFO license server be set up differently, so that it's readily accessible via the network and off of .192? Could it run on another junker machine, out of the Parmly 307 closet? Who would do that setup?
Data: Because the desktop implementation of ArcGIS uses drive letter mappings to locate files, it has been necessary to put GIS data that's to be shared in a commonly-accessible place, most commonly \\acadproj\vol8\. This location has built-in problems, most specifically that access permission to \\acadproj\ for students has to be explicitly granted by Ruth Floyd, and is limited to specific \vol8\ folders --thus, my Global Stewardship students were granted access to \esrixx\ folders and to \global\, but they couldn't access other \\acadproj\ folders. I've adopted the convention of mapping \vol8\ as the R: drive --this matters if someone is trying to use a Project (an ArcGIS .mxd file) I've created, since ArcGIS stores paths to data layers under a drive letter. There is no systematic catalog or geodatabase available or even possible under this drive-letter mapped regime, and this is one of the basic ways in which our implementation of GIS is bush-league.
The alternative to the current unsatisfactory, limiting, limited, but more or less workable situation is to create and maintain a real spatial data warehouse, running in a real database environment (SQL Server, Oracle, etc...). The 'geography network' model that ESRI has developed for distributed data access (exemplified by geographynetwork.com and geodata.gov) is an ideal toward which to work, though how W&L or ACS or NITLE would actually get there is less than clear. It is quite clear that setting up and maintaining ArcSDE would be a task for a real database administrator --far too complex for an amateur to attempt.
ESRI Products: The necessary upgrade to 8.x isn't such a big challenge, though it does mean rewriting home-brewed tutorial pages. I have the installation CD for ArcView 8.3, and the procedure is pretty simple (just requires installation on a local machine, and pointing the Desktop Administrator License Manager toward ims1). There may be more challenges when 9.0 is released (maybe January?), but the additional features will make it worthwhile.
Facilities: Parmly 302 is the best we can do as a lab/teaching venue for the foreseeable future. My memo on hands-on computer classrooms did get some action, in the form of a tuning of the AC in the room so that it's less oppressively loud, but the AV issues haven't yet been addressed, and there are no fundds available in 2003-2004. This should be identified as a matter of high priority for 2004-2005.
Support: Our model for GIS and spatial data support is inadequate to daily needs, let alone to an expansion of the data visualization, analytical, and pedagogical applications of GIS technology. It consists of (1) things I've learned through several years of experimenting and workshops attended and given, (2) knowledge of several students who have gone further than I and other faculty (I'm thinking of Drew Heath and Jit, among current students), (3) the Geology-specific knowledge of Dave Harbor and Chris Connors, and (4) an absence of formal commitment to this aspect of the Information universe by Library and University Computing. It's unclear how this can change in the absence of funding for a 'GIS guy' to provide the expertise we don't have and can't easily develop, and similarly unclear how demand for such a position can be developed or articulated.
Consortial activities: NITLE is in the process of hiring a GIS Coordinator, though it's less than clear whether the position could have any influence on what happens at W&L, or in ACS. The experiments that Skip and I were doing, mostly with linkage of Access and ArcIMS, can only be continued with skills and knowledge that I don't have. I reluctantly conclude that they will remain as prototypes, though the Civil War project might attract attention outside of W&L. We do not have the resources to support further development or hosting of an expanded service.