[fpc-pascal] Standardization of Modern Pascal

[Marco van de Voort]

> I found your answers to my naive questions really interesting. And
> certainly I agree. Would it be possible to make them more visible on the
> web site of Free Pascal? In practice, include the texts you pointed to
> (iso-ansi discussion and delphilater) in the web site? Or did I missed
> them?

Problem is that they are my opinions, and not necessarily those of core.

> A few comments before I turn back to work.
> 
> I think that Modern Pascal is not so fuzzy.  Indeed my feeling is also
> that the Free Pascal community plus CodeGear's Delphi constitutes most
> of what is today Modern or Object Pascal. So the current implementations
> (FPC2.2 and Delphi2008?) are making the langage.
> 
> If this is so, there is no point to discuss a standard. As much as I can
> see, the situation, in practice, is analogous to Sun and Java. Only that
> CodeGear is not going to collaborate anyhow with Free Pascal. If there
> is something like a "reference" it is rather Delphi. In the sense that
> the FPC try to be compatible with Delphi.

All correct I think. But the reference has been delphi till 7, and after
that it is less clear.

> Considering that who wants a standard wants to write a code as
> independant of a given compiler, the question resumes to write code that
> can be easily ported between Delphi and Free Pascal. The answer shall be
> to write it Delphi6/7 compliant and forget about .NET. This is only how
> I see things.

> An important addition is that "Delphi extensions are examined (...) if
> it fits FPC tradition...". I believe you refer to this text?
> 
> http://www.stack.nl/~marcov/delphilater.txt

Yes. The problem with the later versions of Delphi was that they were
positioned as a means for easy transition from win32 to .NET. But that
didn't go so well with the users, and it is said that CG is splitting the
targets again. It was really a bad marketing decision in the first place.
Porting to .NET meant effectively a rewrite anyway.

But some of the later features may appear in FPC in time. Another problem is
that FPC has always had some functionality earlier than FPC. (function and
operator overloading e.g.) In the past, often if Delphi also started to
support it, some small changes were made so that the slightly different
Delphi syntax would work with FPC too.

However the set of features that is introduced with Tiburon (unicode and
generics) are signficantly larger, so a chance on a durable rift between FPC
and Delphi is larger.
 
> I understood also that programing concepts are more elaborate into FPC.
> You stated that generic classes (isn't it?) for instance, are already
> implemented while they are only scheduled in Delphi. While most
> exclusive features of Delphi are related to .NET functions. In that
> sense, it is relevant to port Delphi code to Free Pascal but it might be
> hasardous to port Free Pascal code on Delphi...
> 
> Concerning .NET, the concept is useless for a compiler that is far more
> crossplatform than the Microsoft's framework will ever be.

Well, regardless of your opinion of .NET (or Java, the difference is not
much), the problem is that a deep integration of one of them with FPC with it is not very
useful. The languages and compiler models are too far apart to have synergy,
see:

http://www.hu.freepascal.org/faq.var#dotnet