With Windows Vista nearly here, the inevitable request to Microsoft for it to include built in OS skinning features has begun.

Putting aside the fact that vocal requests for skinning to be part of the OS didn't start up until third-party OS skinning programs started to come out, SHOULD they be part of the OS?

The first question is, how powerful should the skinning be?  Should they basically be resource replacement type skins (i.e. change the graphics to other graphics?). Should said skins allow controls and title bars/border graphics to be replaced (ala Msstyles).  Or should it allow for buttons to be moved, added, borders of independent thickness, animation, etc. (ala WindowBlinds).

Compatibility slides down as you allow for more sophisticated skinning.  One thing most people don't realize is that skinning programs themselves are pretty rock solid at this point.  But the SKINS themselves can cause problems for some people (i.e. slow downs, goofy graphical artificats, etc.). 

So the more power and flexibility you give to users, the more rope you are giving people to hang themselves.

Over here:
http://shellrevealed.com/forums/thread/332.aspx there is a discussion on whether skinning should be in the OS.

The people there are vigorously arguing that it should.  But what precisely is their reason?

From the discussion there it seems to me to be a matter of cost.  That is, a third-party program might be just as good (or better) than what Microsoft would include but they don't want to pay $20.  They unconciously want the millions of users who could care less about a feature to pay for the development and support.

There are 4 basic reasons I don't think such a feature should be part of the OS:

#1 Support.  Buggy skins cause problems. We've seen that. Msstyles, which are pretty mild in skinning power, do cause problems. Most people don't even realize that the app that hung randomly or just disappeared did so because one of the graphical resources in the .msstyles they were using messed up a given program.  WindowBlinds is no different in that regard. A badly made skin can cause compatibility problems.  But as long as it's not part of the OS, Microsoft doesn't have to deal with the support from that.

#2 Branding.  Microsoft is trying to have the Windows Vista experience be a specific brand. A specific look and feel. Putting skinning into the OS weakens that brand. Every Linux distribution could start including an Aero look and Microsoft's trade dress would be likely too weak to prevent that.

#3 ISVs.  In the short-term many into skinning people would think it great if skinning were part of the OS. But in the long term, it would be to their detriment.  It would eliminate the commercial market for skin engines.  Now maybe you're thinkig "Good! Those scummy corporate bastards should get real jobs."  Maybe that's even true. 

But once they're gone, they're gone.  At this stage in the game, it would be immensely difficult to pick up and create a new skinning engine from scratch. Note that zero third-party skinning programs were created in the past several years. None. WindowBlinds pre-dated Windows XP and so it was able to continue to evolve.  Other programs, Style XP and what not are just patching uxtheme.dll to use Microsoft's msstyles engine.

The net effect is that from that point forward, users would be dependent on whatever scraps Microsoft chose to throw to the skinning community. 

Stardock would be just fine in either case, it doesn't need the skinning market to thrive, it has lots of software.  So some user thinking that my argument here is based on commercial viability need a reality check.  The skinning market isn't THAT big. Enterprise software, PC games, are much more lucrative.

But you take away the ISV alternatives and you're left with whatever charity Microsoft is willing to hand out.  And once those ISV alternatives are dead, they ain't coming back. The skinning market is simply not big enough to justify the effort to create something from scratch at this point (it's worth mentioning that it took 8 years from the release of Windows to even start getting third-party programs and they were slow and buggy for their first couple years).

#4 ROI.  Return On Investment.  How many people would actually use such a feature? .05%? .1%?  Sure, in absolute numbers that's still hundreds of thousands of people.   But the cost in developing and supporting that feature would in effect mean that the 99.x% of users are forced to subsidize a feature that only a small group wants. 

And that small group can already get skinning -- from third parties. So it almost screams out "I want other people to pay for my hobby".

Not that I think there is something wrong with people requesting features to be part of the OS.  If someone wants a feature, they should ask for it.  I only start to object when people try to pass their desire for something for free as somehow being a principled stand when it's really just wanting to avoid having to pay for something or being inconvenienced by having to load up a third party program.

Before someone reads this and says "But third party programs add bloat" -- uxtheme.dll and WindowBlinds use trivial amounts of memory.  The VDM in Windows Vista uses typically 70 or more MEGABYTES of memory and it's primary purpose in Vista right now is for Aero glass.