By that argument, then Maya should cost $100.  You could get Visual Studio for $100 and develop Maya.  The cost of authoring tools is defintely related to how much value they can justify.  

To use Aquarium Desktop as an example-- with all the Aquarium screen savers out there, why has no one else made a decent aquarium desktop? The answer is that it's very very difficult to have fluid animation, alpha blending, and per-object scripting and DesktopX can do that. 

Now that we have a site (DesktopGadgets.com) for people to sell cool stuff, we don't want someone to simply pay $70 and be competing with us on products that generate alot of revenue.

I would bet that Stardock made more on DesktopX-generated content than Konfabulator did in sales to end users.  What we've learned, over the years, is that DesktopX is currently better suited as a development environment than as an end user application.

I don't know what the future has in store for DesktopX.  Windows Vista is going to change the dynamic a great deal.  Moreover, with Konfabulator free, it is a lot harder to justify putting time into the widget-creation elements of DesktopX (other than for DesktopX Pro). 

Simply put, DesktopX has tended to try to be 3 different things -- a widget creator, a desktop extender, and a quasi-alternative shell.  And those 3 things are so different that I think it makes it a) too hard to use for many people and stretches the QA resources on it too far so that it ends up being not quite satisfying in any of the 3 areas.

No because it requires immense effort still.  Visual Studio can't turn out a Word processor in a couple of weeks.  DesktopX Pro can build some very specific types of unique content without alot of effort.

Let's use Natural Desktop as an example -- with DesktopX Enterprise, we built that in a few weeks with just a couple people.  With Visual Studio, it would have taken many many many months to do it.

For building desktops, DesktopX is very specialized at the types of content it can create just as programs like Maya are specialized to create 3D models.   There were a lot of people at Stardock who didn't want DesktopX Pro or Enteprise to exist at all (particularly Enterprise).

If Konfabulator has shown anything to be true it is that the CONTENT matters more than the technology of the thing that uses the content.   DesktopX can create, relatively easily, a very specific type of unique content that people want.  So the question is which direction is best? Turn DesktopX into a quasi-internal-only tool to create other stuff to sell that is stand-alone? Or to make it available to other developers who might compete with content Stardock might want to sell? There's no easy answer.

Let me give you a totally unrelated example: The Political Machine.  Stardock developed a Direct3D backend for DesktopX to take its .dxpacks and run them in a 3D environment.  As a result, Stardock was able to crank out the user interfaces and screens for The Political Machine in a very short amount of time.  This allowed Stardock to cut development time on the game in half.  This led to the question -- should Stardock try to encourage/sell DesktopX Enterprise to other game developers to help them (And generate additional revenue) or should DesktopX Enterprise be used as a competitive advantage for Stardock's own internal game design teams?  Like most things, the answer ended up somewhere in between -- we decided that we would tell people how we wer able to create all these game screens so fast but we wouldn't encourage people too much to use DesktopX for game making unless they came to us and negotiated licensing.

The Desktop Pet is a good example and you'll see why when it comes out.  That was created with DesktopX Pro (not Enterprise).   The Desktop Pet will probably sell 5,000 to 10,000 copies on-line and if we get it into retail another 20,000 to 30,000 units.  You couldn't create a desktop pet with say Konfabulator or any of the other "widget" programs because they're currently too limited.  But DesktopX Pro, out of the box, can do it.  One developer and one 3D modeler in a couple months.  And there's nothing out ther that is anywhere near as sophisticated.  But there's nothing stopping someone else from doing the same thing as we're doing.  All some larger company would have to do is pick up a copy of DesktopX Pro for $69.95 and create "Virtual Rover" or something and have a hit leaving our pet in the dust.  So again, there is always some conflict over how DesktopX should be marketed.  Some might say "Why does Stardock even let DesktopX be available at all to the public?" 

So far, because people make trinkets like clocks or try to copy what's been done on Konfabulator, the issue with what DesktopX should and shouldn't do has never come to a head.  A few months back someone made an Aero web browser using DesktopX that was fully compatible with Internet Explorer (it uses its ActiveX control).  I know that raised some eye brows around here. 

Some people forget that DesktopX was developed during a different time in the skinning community.  The target market for it was skinners (same as WindowBlinds).  It was expected that the community would jump in and figure out all the neat things that can be done with it.   But that didn't happen.  And so 4 years later, Konfabulator shows up on the Mac and then eventually on Windows.  And unlike DesktopX, its developer included abunch of really cool content it made itself.  It wasn't made with skinners in mind, making a Konfabulator widget (especially when it was first released) was huge pain.  They focused purely on the consumers -- the users.  Imagine how different DesktopX's history might have been in 75% of the budget wasn't spend on the content creation elements -- the GUI for creating, manipulating, and exporting objects and desktops?  6 months alone were spent figuring out how to make ActiveX controls move seamlessly with other DekstopX objects.  Nothing else can do that.  But how many things use ActiveX controls? Hardly any.  It's incredibly powerful stuff but instead people debate which is "better" based on how pretty the little PNG files that display clocks and weather monitors and other trivialities. 

In other words, Stardock developed DesktopX under the misconception that the key to success was getting "Skinners" on board.  But that has changed in the past few years.  What we've seen both in our own experience is that consumers are driving success.  Consumers will choose what they find easiest/best to use and THEN the skinners will follow what the consumers are already doing.  It doesn't matter that Konfabulator and the rest have no real development environment for the skinner.  The skinner will sit down in some editor and peck out their XML and their JavaScript and deal with all the other hassles because that's where all the user are.

The problem is that DesktopX is still saddled with all this development environment stuff which adds a great deal of complexity and overhead to the actual environment (unless you use just the DesktopX run-time which we did in 3.0).  If there's a DesktopX 4, I think you'll find it a very very different animal.  It'll not care about the content creation, it'll use XAML what content creation help it provides will mostly be about making it easier to do simple things (widgets for instance) but fancier things would be done using Sparkle and what not.  But it's hard to say this early on.

Volume doesn't work in many many cases.  That is the reason why products like 3D Studio and Maya cost so much.