The ESRI Dilemma

I’m in bit of a quandary, and so is my organization. Maybe I’m missing something obvious, over-analyzing the situation but I am not sure what to do with ESRI. Maybe others are in the same boat as well.

For years, ESRI was the biggest name in GIS. Still is. Many members of my organization, including myself, have been using ESRI for 10+ years. However, the following factors now come into play.

• Budgets are now tight and ESRI (software, training, conference) is expensive.
• Geospatial applications and concepts are everywhere, thanks to Google, Microsoft, e911, Homeland Security, environmental concerns, GPS, and in-car navigation.
• I can’t stand ArcGIS Server. Most of what we do is simply serving of maps, raster and vector. ArcIMS was OK for this but ArcGIS Server is bloated, slow, and complicated (not to mention expensive.)
• To me, not much has been chaning with ESRI for many years.
• ESRI is still used most often with many of our colleagues, people who do not have the flexibility to change.
• What will happen to ESRI when Jack retires? What will happen if it goes public?
• In my mind, analysis has always been the re-headed step child in ArcGIS. Yes, it’s there, but it’s limited and slow. ArcObjects is a bear to learn, which you’ll need to use if you want to do complex analysis.
• Maturity and stability of open source map servers.
• Same goes for javascript APIs, like OpenLayers, GeoMosse, ka-Map, MapFish, MapBuilder, MapGuide, and many others.
• Analysis using a variety of open source programs like GDAL, GRASS, Python, R, SAGA, QGIS, etc…, is getting easier to use. It’s also more reliable than Arc*, much faster, and more flexible. You need to write code vs point-and-click interfaces, but the programs are getting smaller and easier to understand.
• Postgres and PostGIS works wonderfully and can be used with many different packages. And spatial SQL queries can be used for analysis on its own.
• Open source desktop GUIs are looking good. QGIS, MapWindow, uDIG, gvSIG, OSSIM, SAGA are pretty good now and getting better. Not quite as nice as ArcMap though.
• Use of some related software, like Postgres and PHP, and operating systems (Mac and Linux are growing) is increasing in other, non-GIS areas.
• Open source licensing is pretty good!

So, what do we do?

• Do we drop ESRI now and put all money and effort into learning a FOSS4G stack?
• Do we switch to open source for analysis and web serving and keep ESRI on the desktop in ArcMap?
• Do we keep using and paying for ESRI while keeping an eye on open source?
• Do we ignore open source and keep moving forward with ESRI since we already have a significant investment? (I know this won’t work. Many of us have been looking ot open source because some Arc products just couldn’t cut it. This is how we got here.)
• Do we look for a middle of the road product, like Manifold or Idrisi, that has a commercial company behind with paid support but not as expensive or bloated like ESRI?

I don’t have the answers but I know I need to make recommendations on what to do, spending real money and affecting real people’s time. When budgets tighten, you want to create a situation where you are not vulnerable when it happens again. We need to budget for annual maintenance costs, training for employees, future projects, staff expertise… We need a long term strategy here and I have a feeling the answers are not going to be easy.