ESRI Developer Conference
OK, the weekend is over. I say that as the conference was Friday and Saturday and felt like I was using my weekend up for it, but there you go.
Since it was the first conference of it's kind ever, it was pretty small, in the sense that we are all used to the masses as the San Diego International User conference with its teeming thousands. There were about 700 or 800 people attending total, which must have included the ESRI personnel. ESRI staff seemed to outnumber the attendees at times. I think the ratio must have been 1 ESRI person for every 3 attendees, and they were pretty much always accessible. Which is not true of the San Diego convention. I got all my questions answered before lunch.
Well, most of them. One thing about this conference is that you don't have a slew of people trying to break into the field or trying to make contacts, sell stuff, and the like. These are all programmers, not in the custom or mood to make lots of new friends or shmooze. I saw lots of people with blinders on. But there were friendly folks and certainly if I took the time to talk to someone there was conversation to be had.
Overall the sessions were pretty good, and depending on what you came for, the information for once was not at all too shallow. Some sessions were indeed over my head, as I am not well versed in C sharp or .NET. The conference seems to follow the ESRI vision of distributed GIS and web applications. Not entirely, but when in doubt that's what I began to expect. ESRI, as well as Microsoft, seems to be pulling us tooth and nail into the .NET environment whether we like it or not. If there was a coding example, it was in .NET. Or Java. Or C++. Although they say that ESRI will support VB6 into the foreseeable future, I think that I am going to have to learn other languages just so I can keep up with any examples they use in the future.
Unlike other conferences, there were no vendor booths. This was strange at first, until I realized that all the vendors are here, but are here as attendees. No one is here trying to find a service or application, they are here to learn how to build them.
Instead of vendor booths there was a new event during lunches and after the Friday sessions that I really did like. Birds of a Feather meetings is what they were called. They were more like round table discussions with about 20 or so people, and it was much more dynamic and interesting than most of what you get in the lecture rooms.
I went to two of these. One was Unit Testing, put on by Dave Bouwman of Sanborn. It was pretty good, although I didn't have much to contribute. Most of the guys there were seasoned programmers who write huge programs that make testing small parts of them incredibly difficult. But they are continuing the discussion over at ArcDeveloper.net going forward, and it will be interesting to see what they come up with.
The other one was Migrating from VB6 to VB.NET, which was the most helpful and informative session in the entire two days for me. ESRI staff were there to provide tips on what you'll need to watch out for, and we were given the opportunity to ask lots of questions.
One thing of note with .NET is that it has a wrapper that enables you to use COM objects, which is the only way you can work with ArcObjects at all. That causes programs to slow down if they have to make lots of repetitive calls to arcobjects, like in loops. The ESRI guys were actually recommending that we use VB6 custom code to wrap up loops and call those libraries from .NET applications. I asked when he thought ESRI would get around to changing over all the COM objects to the .NET framework and he replied that it wouldn't be for several years. At least 5 or 6. So we're going to be stuck in this VB6-.NET transition world for quite some time.
Dave Bouwman has a few posts about his experience there. I agree with what he was saying about the "coarse grained objects" mantra for the most part. I thought it was interesting given the crowd. Having said that, I think that I will find myself using those objects from time to time, as programming is only part of my job and sometimes I just need to move on knowing that the underlying code is fairly sound.
He is right about the ESRI persons spending too much time on 9.2, if nothing else because it caused them to spend too much time showing off Visual Studio 8, which we won't really be able to take advantage of until version 9.2 comes out. Which is when again? Anyone want to place bets that it won't happen until the end of the year?
In his last post Dave also brings up what I was feeling the entire time I was there. Sure I liked seeing the new stuff and the things that .NET can do, but there was a serious lack of information and support for those of us who are working with ArcObjects. VB6 folks. It's not going away, and VBA isn't going away either, even when people flip the switch over to .NET, VBA will still be the macro language.
James Fee at Spatially Adjusted posted about it here, here, here, here and here.
A couple of things about what James is saying that I feel the need to comment on. One is that his sense after the first day was that the people there were really in to .NET and how popular it was. True the sessions were packed, but those of us who are straight ArcObjects programmers for in-house applications didn't have much choice. There were very few sessions geared to helping us RIGHT NOW, and I think that ESRI needs to understand that. I didn't get to go to the closing session, but James commented on it.
Since it was the first conference of it's kind ever, it was pretty small, in the sense that we are all used to the masses as the San Diego International User conference with its teeming thousands. There were about 700 or 800 people attending total, which must have included the ESRI personnel. ESRI staff seemed to outnumber the attendees at times. I think the ratio must have been 1 ESRI person for every 3 attendees, and they were pretty much always accessible. Which is not true of the San Diego convention. I got all my questions answered before lunch.
Well, most of them. One thing about this conference is that you don't have a slew of people trying to break into the field or trying to make contacts, sell stuff, and the like. These are all programmers, not in the custom or mood to make lots of new friends or shmooze. I saw lots of people with blinders on. But there were friendly folks and certainly if I took the time to talk to someone there was conversation to be had.
Overall the sessions were pretty good, and depending on what you came for, the information for once was not at all too shallow. Some sessions were indeed over my head, as I am not well versed in C sharp or .NET. The conference seems to follow the ESRI vision of distributed GIS and web applications. Not entirely, but when in doubt that's what I began to expect. ESRI, as well as Microsoft, seems to be pulling us tooth and nail into the .NET environment whether we like it or not. If there was a coding example, it was in .NET. Or Java. Or C++. Although they say that ESRI will support VB6 into the foreseeable future, I think that I am going to have to learn other languages just so I can keep up with any examples they use in the future.
Unlike other conferences, there were no vendor booths. This was strange at first, until I realized that all the vendors are here, but are here as attendees. No one is here trying to find a service or application, they are here to learn how to build them.
Instead of vendor booths there was a new event during lunches and after the Friday sessions that I really did like. Birds of a Feather meetings is what they were called. They were more like round table discussions with about 20 or so people, and it was much more dynamic and interesting than most of what you get in the lecture rooms.
I went to two of these. One was Unit Testing, put on by Dave Bouwman of Sanborn. It was pretty good, although I didn't have much to contribute. Most of the guys there were seasoned programmers who write huge programs that make testing small parts of them incredibly difficult. But they are continuing the discussion over at ArcDeveloper.net going forward, and it will be interesting to see what they come up with.
The other one was Migrating from VB6 to VB.NET, which was the most helpful and informative session in the entire two days for me. ESRI staff were there to provide tips on what you'll need to watch out for, and we were given the opportunity to ask lots of questions.
One thing of note with .NET is that it has a wrapper that enables you to use COM objects, which is the only way you can work with ArcObjects at all. That causes programs to slow down if they have to make lots of repetitive calls to arcobjects, like in loops. The ESRI guys were actually recommending that we use VB6 custom code to wrap up loops and call those libraries from .NET applications. I asked when he thought ESRI would get around to changing over all the COM objects to the .NET framework and he replied that it wouldn't be for several years. At least 5 or 6. So we're going to be stuck in this VB6-.NET transition world for quite some time.
Dave Bouwman has a few posts about his experience there. I agree with what he was saying about the "coarse grained objects" mantra for the most part. I thought it was interesting given the crowd. Having said that, I think that I will find myself using those objects from time to time, as programming is only part of my job and sometimes I just need to move on knowing that the underlying code is fairly sound.
He is right about the ESRI persons spending too much time on 9.2, if nothing else because it caused them to spend too much time showing off Visual Studio 8, which we won't really be able to take advantage of until version 9.2 comes out. Which is when again? Anyone want to place bets that it won't happen until the end of the year?
In his last post Dave also brings up what I was feeling the entire time I was there. Sure I liked seeing the new stuff and the things that .NET can do, but there was a serious lack of information and support for those of us who are working with ArcObjects. VB6 folks. It's not going away, and VBA isn't going away either, even when people flip the switch over to .NET, VBA will still be the macro language.
James Fee at Spatially Adjusted posted about it here, here, here, here and here.
A couple of things about what James is saying that I feel the need to comment on. One is that his sense after the first day was that the people there were really in to .NET and how popular it was. True the sessions were packed, but those of us who are straight ArcObjects programmers for in-house applications didn't have much choice. There were very few sessions geared to helping us RIGHT NOW, and I think that ESRI needs to understand that. I didn't get to go to the closing session, but James commented on it.
At the closing session, Brian Goldin went over the some of the concerns that attendees had from the 2 day Developer Summit and it sounds like that ESRI is listening and wants feedback on how to change the event for the better next year. Generally people wanted an extra day, more technical talk and less marketing (Q&A sessions should be longer) and keeping the summit separate from both the Business Partner Conference and the User Conference. Quite a few people have said that they won't attend the UC this year because the Developer Summit was so much more valuable to them. In fact ESRI is looking at allowing ESRI customers with complimentary passes to the User Conference; apply those to the Developer Summit.Sweet. An extra day and we can use our complimentary pass for the Dev summit! That's going to be tres useful next year. I'm glad they made that an issue.
posted by Richard | 7:05 PM
|
3 comments
