Not all feedback to my "Delphi is the VCL" post is here on the blog... although the feedback on the blog is very interesting, and includes even a post from Delphi R&D (btw, did you notice that Danny Thorpe posted a comment on this blog a couple of weeks ago?). There is a nice thread in non-tech and a post that made me blush, David I referring to the blog post as a "spectacular blog post". OK, I know self-congratulating on a blog post is childish, but non the less...

The real reason I wanted to write again is to provide some clarifications. The more I think about it and my views get more clear, the more I think this was a proper way to understand the roadmap, but at the same time a correct way to push Delphi further ahead in the future. Delphi is for the Windows platform, but this platform is a moving target. On the web side you have Windows Live services and Silverlight, but on the fat-client and client/server sides evolution is slower. Still, demand for small and lightweight software is there to stay. As I recently read in a Delphi 2007 review, a simple VB.NET application consumes many times the memory of a Delphi one. Not always relevant, but interesting. Some specific points:

  • I agree that CodeGear really needs to put extra effort in the development of the VCL, and I'm wondering if the community should help once more and how (I do have a couple of ideas).
  • I'd love to see things move in the Object/Relational mapping area (at a more or less complex level) in the WIn32 side. This can keep the VCL ahead of the object mapping Microsoft is working on.
  • I have used VCL.NET and made speed comparisons with WinForms and things are interesting. Add to this better support for Vista and it can be a starting point.
  • I disagree that .NET is a fad. But on WIn32, it is an option and not the only way to go.
  • Of course, Microsoft will not rewrite Office in .NET. But in Vista I haven't found even a small new little utility program written in .NET. I might be wrong, though. It is true that new software from Microsoft, like the Expressions programs, is written in .NET... but as that is specifically for graphical and XAML manipulation, the opposite would have been astonishing.
  • I feel constrained by GDI as well, and would like to have an OpenGL VCL. I guess this is very unlikely. GDI is weak, but quite optimized. And not all applications need the same level of graphic capabilities.
  • Having a XAML for the VCL as a DFM replacement would be interesting.
  • The real big open question remains, how do you move your VCL applications to the web. If you need to do so, of course, as many can happily live in the Win32 world.

Again, thanks for all the feedback, it helped me clarify my ideas to myself first, and possibly also to the readers of this blog.