Notes from the 2006 ESRI User’s Conference

This year your intrepid scribe managed to get out to San Diego for the ESRI UC. I thought I’d leaf back through my notes (read: odd pictographic scribbles on cocktail napkins) and post some of the highlights.

Labeling & Annotation

  • With Maplex built in and some better label placement and annotation editing tools in 9.2, ArcGIS will have better label and annotation capabilities. If you care. I don’t. I haven’t made a decent looking map in years. And that one was an accident.
  • In 9.2, you will be able to overlay multiple map services. This is a huge improvement. ArcIMS boxes tend to get map service creep over time (we’ve got a server with almost 30 of the darn things on it), and hopefully this will help users break layers into logical groupings/services and pick and choose what they need rather than slapping a new map service on the server for every little thing.
  • The big “wow” for most people was the new designer site. While the old designer spat out a fairly useless HTML viewer application, the new designer spits out a nifty .NET or JAVA web application which you can then pull up in Visual Studio or Eclipse and edit in a decent IDE rather than run about trying to configure obscure text files. It is a much improved interface as well, being a fairly accurate Google Earth ripoff (a good thing). You can zoom in and out with the mouse wheel, and ArcIMS tiles the images to support that like Google Maps does. My only question is whether the server will tank as some yo yo moves his or her mouse wheel back and forth for half an hour. Each zoom makes numerous image calls, not just one. It will be interesting to see how badly it will punish a server.
  • For .NET developers, you get real objects to work with, not just a “link” you can pass XML through. There are objects for the map, tool bars, table of contents, etc. It should make development a bit easier.
  • A biggie for admins, you can now control whether ArcIMS will let users connect and download data. This is something that should have been in since 4.0.

  • I sat through a horrific Python session. To spell out a few things - Python is an open source scripting language. It’s used by developers all over the world. Its main features are it’s elegant coding design and the rather odd use of signification white space. Namely, code blocks such as loops are demarkated by indention rather than by begin or end syntax. ESRI just added some libraries for Python so you can do geoprocessing with it. ESRI didn’t invent Python, and Python isn’t going away (i.e. this isn’t Avenue or some other odd little language, so don’t sweat learning it). Model Builder will spit out Python scripts, which you can then schedule to automate tasks. It’s a fairly neat language, and I’ve been looking for an excuse to play with it.
Geodatabase Stuff
  • Replication (push-pull, not transactional) is now a part of the geodatabase. This is a pretty neat development if you find yourself needing to move data around an organization. DBMS replication is OK, but it requires a pretty deep knowledge of SDE to get it working correctly, and it often doesn’t work across DB platforms. With this method, you can plug your laptop in to the network, run a Python script to sync your local database with the enterprise database, and hit the road with the latest stuff. It is also handy because only changes are replicated, making for quicker update times.
  • There are general disconnected editing enhancements tied to the above.
  • There is now a personal SDE, which is free and supports 1 editor, a workgroup SDE, which is less expensive than full SDE and supports up to 10 editors, and then the enterprise SDE. The non-enterprise SDE’s have no RDBMS backend, but the enterprise one still does. Versioning is supported on all of them. Personal geodatabases can now also be stored as a file rather than a MDB. Confused yet? Long live shape files.
  • You no longer have to specify XY extents of layers. That would be nice.
  • You can edit non-versioned simple feature sets. That would also be nice.
  • Although there are some new features, the big thing here is bug fixes. Over 4,000 bugs were fixed from 9.1 to 9.2 release candidate. Of course, those are fixed bugs. They weren’t fessing up as to the number of actual bugs.
  • There are a number of improvements to cartographic features, with rules to automate symbology.
  • CAD support is improved.
  • Lots of usability tweaks. For example, you can hit a combination of eight or so keys to get some dohickey to pop up that is a lot better than the old dohickey. Or something like that. I’m afraid I don’t have the free mental storage space to house a bunch of new keyboard shortcuts.
  • They seemed hyped up about their improved documentation. Well….whatever. It will still suck. It’s the nature of documentation to suck. I don’t think I would trust documentation that didn’t suck to some degree.

GIS Interoperability
  • I’m still unconvinced about ESRI’s commitment here. I can’t say I blame them - if I owned 95% of the industry, I wouldn’t see letting some open source yo-yo connect to and use my data/software as a particularly positive development. That being said, ArcGIS will (finally) support WFS out of the box at 9.2. It’ll read GML files too.
  • On an irritating note, to read data directly from my OGC-compliant Postgres database, I have to buy a $2,000 Data Interoperability extension. Sigh.
Other Notes
  • Geeze a lot of people go to that conference. I kept hearing 14-15k people being bandied about this year. It’s almost outgrown the San Diego Convention Center, and that place is humungous.
  • I think Google scared the pants off of ESRI. ArcExplorer and the new ArcIMS designer application have a distinct Google Maps/Earth flavor to them. I think that’s a good thing.