The very first PDC I attended was in 1998 in Denver (before I joined Microsoft). Don Box reckons this was an unremarkable conference by PDC standards, and he may be right. Nevertheless, I remember being blown away by the range of ideas Microsoft were working on. Some of those ideas have since become a central part of Microsoft's developer platform; others have since been quietly dropped.

One of the sessions that attracted a lot of attention five years ago was an architectural session by Mary Kirtland, covering the COM+ runtime environment. She demonstrated an experimental C++ compiler that supported declarative programming. The idea that one could express intent via attributes evolved into C# and of course today seems almost mundane in the .NET world. Creating a web service from a normal class with a simple [WebMethod] attribute still amazes people whenever they see it for the first time, though.

Another session that held my attention was an otherwise dry keynote from David Vaskevitch, which was something of a "State of the Union" address about the Microsoft technology roadmap. He had one slide that I'd love to dig up from the archives, which showed the building blocks of the platform vision. Five years on, we'll be handing out bits of two of the last pieces of the jigsaw puzzle as part of Longhorn: Forms+ (now Avalon) and Storage+ (now WinFS). My guess is that both have changed beyond all recognition since this original slide, but it's interesting to note that the visions have stood the test of time one way or another.

In fact, the longer I've worked for (or with) Microsoft, the more it seems to be the case that most ideas bear some fruit eventually. Even Hailstorm, which was widely viewed to have sunk without trace (certainly a search on microsoft.com reveals little of what was once a flagship PDC technology). But much of the vision has been incorporated into MSN for consumers; to a lesser extent, in the business arena by technologies such as SharePoint and Greenwich.

I'm looking forward to this PDC, not just for the detailed sessions on Longhorn, Whidbey and Yukon, but for the opportunity to hear some of the brightest minds putting forward their vision of software over an even longer term. I'd like to see a glimpse of the work we're starting now that won't bear fruit for another five years time. Hopefully Longhorn will be old hat by then!