High Expectations
DDJ: Another hypothesis that I was going to forward, but I think you've preemptively shot it down, is that maybe Visual Studio 2003 set expectations artificially high. In a lot of ways, Visual Studio 2003 was just a massive service pack for Visual Studio .NET. It's logical that 2003 was very stable because it didn't introduce very many new features. It was mainly a big cleanup of the original Visual Studio .NET. However, even so, compared to the original Visual Studio .NET, Visual Studio 2005 is less stable.
Kathleen: It's worse by degree. It's not night and day. Even with the stability problems of 2005, I would choose to work with 2005 over 2003 any day of the week because of the features in 2005. I won't go back.
Bill V.: Agreed.
Kathleen: 2005 isn't so bad that you feel like you can't use the product. However, I do know people who are not moving to 2005 because of what they've heard about stability. They don't want to use something that still feels like a beta.
DDJ: I think some people would say some of these issues aren't technically stability problems. Instead, it's working as designed, the proverbial, "It's a feature!" For example, if the user control throws an exception at design time, or if there's an error in the designer generated code because you've changed the control and that code is no longer valid, then the WSOD, by design, is what you're supposed to see. Do you feel that this is really stability problems, or are things working as designed, but the design isn't really adequate in some areas?
Bill V.: I think a lot of these issues did show up in Ladybug (Editors note: also known as the MSDN Product Feedback Center.), and they were tagged as "Won't fix", or "By design". This happened for issue after issue after issue. Now we're looking at Orcas, and we're talking all these new features, we're talking about LINQ and things like that, and I think Orcas should the 2003 type of release. It should be bug fixes, stability issues, and working really hard on the documentation and discoverability of features that they already have. Push all those new features out to the Hawaii version.
DDJ: Is there any way, today, to obtain the hotfix without actually picking up the phone and calling PSS?
Bill V.: That's totally ridiculous. I called PSS; I talked to three different people. It was a painful waste of almost an hour of my time that I'll never get back.
DDJ: There was a petition. I looked up on the MSDN Product Feedback center this morning, and there was a recommendation that Microsoft not ship Visual Studio 2005. The recommendation was, "Have a Beta 3. It's not ready to release." I don't know when the voting closed on that feedback, but this morning, it had something like 230 votes. For something on the MSDN Product Feedback center, that's a lot. Writers for trade publications picked up on it and wrote articles saying that Microsoft's customers feel that it's not ready. On the other hand, all we had to look at were CTPs, and with CTPs, all bets are off. They are a snapshot, but the quality of a CTP is unlikely to accurately reflect the quality of the final product. Did people see this coming? Should Microsoft have known? Did Microsoft know that the quality wasn't going to be near that of 2003?
Kathleen: I think to get a stable product, we have to have something in the field that's about as good as what we have right now. Now I would prefer to call a Beta, and shake down something that's 98 percent of the way there. I think that would get the final product correct. But we desperately needed this releases to ship last fall. That's particularly true in Windows Forms because we can't release production stuff on Beta software. If I was inside of Microsoft, and someone had said to me, "You make the call. Do you ship now or wait?" I'm not sure I would have said to wait. Even though it's a pain to work with, there were a lot of people out there waiting for this to release. They had stuff that they wanted to get to the field. The .NET runtime is stable, so I'm not sure I would have delayed the release because of the development environment issues.
DDJ: Let me get other people's impressions on that. Given what we know now, if you had to make the call to ship back in November, would you have pulled the lever to ship it? (silence)
Bill V.: That's a tough call.
Billy H.: It is.
DDJ: Let me maybe make a distinction. Are we talking about something that's a tool problem, or is it the tool and the Framework? In other words, is it an issue where the development environment is unstable, but the underlying .NET Framework is solid, so that you can have very high confidence that the applications that you build against the .NET Framework will run well, and be stable for your customers, even if the tool that you're using to build them gives you grief?
Bill V.: Some of the problems related to Data Access are in the Framework, and the tools can only be as good as the Framework. I think it's both.
Billy H.:I don't have anything in production yet on 2005, but I haven't seen any issues with the Framework that concern me.
Kathleen: In my experience with Windows Forms, it's pretty good. There have been a couple issues, but I'm not worried about release.
Bill V.: One thing about the Framework, if you do run into an issue, you just code around it. You just know that there's a pothole in the road at mile 45, and you drive around it. It doesn't have to affect the stability of your application.
