Hi Samuel,

[I reorderd your post for my argument a bit]
>Another reason is for extensibility, we want our community to be able to improve and perfect the functionality and as you know it is easier to do that in VFP than in other languages, and you really know that, the code in the FFC and XSource would have taken a lot more of effort to write it in another language as C++ or C#.

That was my take as well: if there is something in .Net-runtimes or IP-free C++/C# helping me, I'll use it.

>So if not solution comes implemented in VFP and there is one in other language, we'll use it because we want to speed this as much as possible, but for our community right now the best language for this task is VFP.

I got that impression from your posts as well, but here on UT are so many different viewpoints that any statement is probably interpreted in more than a dozen of ways. Not critizising your posts - just something to keep in the back of your mind when writing the posts. The vfp roadmap was obviously interpreted differently - and I truly believe it was meant to be as open as possible without hurting the vfp community in other areas / even earlier than necessary.

>We don't want to reinvent the wheel, Kamel and Flash put a lot of effort, but Several of the functions are not compatible with the functionality of VFP, those functions have the intention but not the functionality... Some that come to my mind are the FOpen, FCreate, those even don't return a number!. For those we better come with our own implementation and what better than to do it in VFP language.

Perhaps the task for each function *should* be split into two tasks:
a) create a unit test ***always in vfp*** to check the actual results against vfp
b) the implementation of the new routine - with vfp being preferred, if no other implementation exists.

Then any implementation can be either checked through Netextender directly from vfp or through your online compiler and a local .Net installation, as I am sure anybody proposing a C++ implementation will have the .Net compiler for it to run basic testing.

I also think a more wiki like approach has more benefits than drawbacks. As long as tracking/version control is working <bg>.

regards

thomas