No header files?

[AJ]

"Leandro Lucarella" <llucax at gmail.com> wrote in message 
news:20091024182205.GA28347 at llucax.com.ar...
> AJ, el 24 de octubre a las 01:44 me escribiste:
>> "Leandro Lucarella" <llucax at gmail.com> wrote in message
>> news:20091023125658.GA23090 at llucax.com.ar...
>> > AJ, el 22 de octubre a las 22:08 me escribiste:
>> >> > I'm still not clear on exactly what you plan on doing.
>> >> Ooops! No I don't! Yes, I will be "modeling" via header files.
>> >
>> > May I suggest something? If you want to "model via header files", write 
>> > .d
>> > files with just declarations (skip implementations). You'll end up with
>> > hand written .di (headers) files, but in a .d files, so you can still
>> > automatically generate .di files from it. Use the .d files as an import 
>> > in
>> > other modules (i.e., don't provide hand written .di files only). Then,
>> > when you want to start implementing your "model", just add the bodies 
>> > to
>> > the functions in the "modeled" .d file.
>> >
>> > The effect in the development process is the same as hand writing .h 
>> > files
>> > and progressively fill the .cpp files with the implementation, but you 
>> > are
>> > doing all the work in just *one* place (the .cpp/.d files) and 
>> > generating
>> > the headers files (.h/.di) from that.
>> >
>> > What problems do you find in that scheme?
>>
>> I'm convinced: D is not me.
>
> My question was honest, I'm really curious about what problems do see.

While I was being just a tad facetious, I was not nearly entirely so being. 
I cannot make an assessment of the D development process in depth at this 
time (but I'm hardly new at SW dev and can eval things at a high level and 
determine whether or not they have value for me (for example, if it's not a 
curly brace language, it's not for me)). I have no issue with my C++ process 
(and I even like it over the D one). To each there own. I need to get some 
software out the door, and one doesn't do that by destroying the old factory 
and putting up a new one.