I'm on my way to PDC.  This is the first time I've gone there. I'm only 34 and yet feel like a bit of an old-fogey in that I'm not as up on the latest .NET techniques as I probably should be.  I'm less a developer these days and more of a technologist.  I want to see what technologies Microsoft is working on and see what is going to be beneficial to our company and our customers and which parts we can take a pass on.

I'm on the plane looking through blogs that I've archived with Blog Navigator and I've enjoyed reading Scoble's in particular.  He works at Microsoft so his rallying up for Microsoft is understandable (just as I tend to be pretty jazzed about Stardock's latest stuff).  Half the "good stuff" apparently is after-hours.  The parties and such.  I'm known enough in the industry to be aware of most of the good parties but not important enough to be necessarily invited to some (such as the PDC "Influencers" party).  Going to a show like this is humbling. I feel like the Comic book guy who is king of his little comic book store but is powerless beyond its confines "I shall return to my store where I dispense insults rather than absorb them..."

Last PDC I sent our lead R&D developer there and he came back with a lot of useful feedback.  At the time, the Longhorn side-bar was a big thing but our developer said it was quite hacky in its implementation.  The side-bar disappeared for awhile and recently resurfaced. Perhaps it's better implemented now.

What I'm going to be interested in learning (and talking to those in the know) is just how far Microsoft intends to take "managed code".  The overall preferred Windows Vista API is called WinFX which is essentially the name of a group of managed code APIs.  Managed code lets traditional developers get a lot done very fast but makes it harder for developers who are doing things that the OS vendor didn't think of to actually do.  Desktop enhancement software requires techniques such as API hooking and message hooking and all kinds of ways to get into a process to extend the base OS feature set.  When a program like WindowBlinds works, it actually gets inside a given process and directs its paint calls to WindowBlinds to take care of it.  In this way, we can extend the feature set of the Windows drawing routines.  But managed code is a layer on top of all those APIs.  You can't really "hook" them.

I'm not so much worried about Microsoft breaking our stuff (Microsoft, nVidia, ATI, Nintendo, Alienware, Good Year, GE, Dell, and countless other large companies use our technologies).  What I want to make sure is that the new goodies in Windows Vista are going to be extensible. And by extensible I mean..actually extensible and not marketing extensible (i.e. marketing extensible is where some marketing guy at Microsoft says how extensible their system is but really what it means is that they allow developers to change the configuration to a number of pre-set definitions). 

For example, could I get into the Explorer process and make use of the new graphical capabilities to have different folders in explorer windows be different sized based on how much "stuff" is in them?  I'm not saying we want to do that, but it's a good litmus test.  On Windows Vista, there are so much more potential to do neat things than there was on Windows XP.  I am not looking for Microsoft to put in APIs that make it trivial to jump in and do things, I just want to make sure it's reasonably possible.  After all, Stardock essentially acts as a free R&D lab for Microsoft anyway (as any long-time user of Object Desktop can tell you).  We're willing to invest in trying cool new things with all the engines but we have to be able to do things that Microsoft hasn't necessarily thought of.  Hence, the need to jump in there and extend (and sometimes replace) painting routines, APIs, etc.

I'm heading there with my Windows XP Thinkpad equipped with the WindowBlinds 5 beta. Amongst its other capabilities, WindowBlinds can do semi-transparent window frames ala Aero -- but on Windows XP.  I know for certain that Microsoft never thought that was possible, especially at the speed we have it running (i.e. on a decent machine it shouldn't affect performance).  So one can imagine the kinds of things that could be possible to do with Windows Vista as long as Microsoft doesn't close the door.  API hooking and getting in process allow developers to gain the same benefits as a developer of an open source OS.