One of the projects that I'm occasionally working on that continues to rear it's head no matter how many people try to cancel it (I am not making this up, but the client, my company, and the tech lead [me] have all tried to kill this project at one time or another ... to put this in perspective, this is almost every single person involved in the decision-making of this project in any way), is a push to retire Site Server and replace it with Crap-merce Server 2003. For the most part, the switch from Site to Commerce Server was painless, but the switch from Win2k to Win2k3 was not.
The object model is the same, which was a nice thing. However, if you store Commerce Server objects in the database, you can't open them with Site Server. The good news is that this process works in reverse, so there's no work to translate our data from Site Server to Commerce Server.
All-in-all, I wouldn't have picked Commerce Server as the platform for this project, as we utilize only a small subset of its functionality. It's just not worth it. However, the client was pretty insistant on keeping the Site Server/Commerce Server functionality (they get a technology audit from time to time, and the auditor likes the sites to leverage supported technology from reputable firms, and MS Commerce Server definitely fits that bill) so the process of retiring Site Server and replacing it with Commerce Server was not that difficult. It was actually pretty easy. If we were to redo the site from the ground up, I would think about Commerce Server as a platform, but the recent troubles I've had with upgrades to Windows 2003 has somewhat shaken my confidence in Microsoft's commitment to backwards compatibility (along with CS2006 only supporting .NET, which is an indicator that the CS team has the same level of commitment to backwards compatibility), so I'm not sure I'm all that interested in binding myself to another of their server packages. The OS (actually IIS) has been thorny enough, and it's nearly impossible to move from ASP to ASP.NET in a web environment of any significant size (if there are more than hundreds to the low-thousands of files, then I wouldn't even try without a sizable team), so locking ourselves into another product, even if it's the same vendor, is not something that seems likely to promote long-term sobriety in our development team.
/NOTE: if you've moved a site from ASP to ASP.NET, that's great. Unless it's a site with 19,764 files and you've got 1 100% allocated developer and 1 tech lead working about 25% of the time on the project, I don't think we're talking about the same task
In the latest of a series of moves that appear to be designed to get as many decision-makers as possible to want to find an alternative to MS products, we've been moving a lot of our websites from Win2k Server to Win2k3 Servers. The nicest thing I can say is that the move sucked less than "Memoirs of Geisha." Only most of our pages and some of our apps broke. It seems like backwards compatibility is no longer a concern, and those of us that have to use these server packages are the ones that suffer. Because of this, we've hesititated on .NET 2.0 and SQL 2k5, as well as Commerce Server 2003. There is a good chance that we will eventually use CS2003 (but only because we have to), but there is a zero percent chance of us using CS2006, should it arrive.
I wonder whether I'm just getting jaded by aggressive, destructive business practices now that I'm in a more senior role or whether Microsoft products have always been this developer-unfriendly and I was just too low-level to notice (I suspect it's the former). Our company has spent, I think, 600 hours in the past few months exclusively on moving 2 websites to Windows 2003 Server. Six-hundred hours. That time would have been better spent watching "Crouching Tiger, Hidden Dragon" 296 times (I'm including an extra 8 hours for beer runs and discussions of Michelle Yeoh's Awesomeness).