zigGIS and Spatial References, Part 3

This post updates this post and this one.

One of our beloved 5 users ;-), James, pointed out recently that zigGIS still has issues with user-defined coordinate systems. Alas, this is true. When we implemented support for on-the-fly “reprojection,” we limited it to the standard EPSG SRIDs supported by both PostGIS and ArcGIS. This was mainly due to the fact that we were trying to fix a number of things at once.

I have begun to look at building in support for user-defined spatial references. The main difference I see is that the prj syntax is slightly different between PostGIS and ArcGIS. For example the prj strings for SRID 4326 (WGS84) are not the same. They seem to behave correctly but this creates a situation where ArcGIS will recognize all standard SRs as custom is I load them using the PostGIS strings. It’s not a huge deal, just interesting. However, Paolo did notice that ESRI and PostGIS data sources with (theoretically) the same SR can have a slight offset when overlaid in ArcMap. So maybe it’s more than just interesting.

I plan to leave the current code alone and just put in routines to handle custom SRs but maybe I’ll also play with the standard ones to see if it makes a difference.

6 thoughts on “zigGIS and Spatial References, Part 3

  1. Dewald,

    That sounds great. You have undertaken a very challenging project and I bet you and your team will have fun pulling it off. I have sent your comments to Abe and Paolo (the other developers of zigGIS) and we plan to discuss how best to take advantage of your generous offer.

    Bill

  2. Bill,

    Thanks for the info. If you need some production testing, I am happy to assist. We are currently migrating our web-mapping to a Mapserver-based solution using PHP (from an in-house mapping application created and since degraded from 2004 – using ESRI technolgies as a backbone).

    This means our desktop guys need to start migrating data from ArcGIS 9.1’s Personal Geodatabase / ArcSDE environment to PostGIS (we need to be fully opensource). So I can offer an extensive testing environment for you – especially back-testing multi-directional data connections. I myself am a C# / Python / PHP developer with GIS experience.

    Dewald

  3. Dewald,

    Thanks for stopping by. You’re correct about the WKT strings. In my next post, I discuss how I got around the problem. It seems to work well but need some more thorough testing. The code shown in that post has not been incorporated into the build yet so keep an eye out for an update.

    Bill

  4. Hi Bill,

    Isn’t ESRI meant to be implementing WKT standards in their PRJ files to ensure compatibility with Open Standards, because if so re-projecting data using user-defined projections (built on WKT strings) would not make a difference. If zigGIS uses WKT OGC Standards, then the problem should theoretically be mute.

    Dewald

  5. Bruce,

    Thanks for the feedback. I can see where that would be a performance hit for you. I forwarded you comments on the Abe and Paolo so that we can try to solve that problem in the near (hopefully) future. I’m glad you’re finding zigGIS useful, though.

    Bill

  6. Bill
    I am looking forward to converting our office to zigGis so you have more than 5! THe huge benefit right now is the persistence of data in the project file. That is the biggest issue right now in our current use of pgArc. The issue right now that is holding us up is zigGis connects to the entire layer from PostGIS instead of the display extent. The main application we use is a massive state-wide database of wells. The performance of zigGis is fine but ArcMap is not with 400K points.
    Thanks!
    Bruce Rindahl

Comments are closed.