Board index » cppbuilder » Re: Open Letter to the Borland C++ Developer Community

Re: Open Letter to the Borland C++ Developer Community


2003-08-07 03:42:42 AM
cppbuilder69
"Peter Agricola" wrote
Quote

No. I didn't know either, but yesterday John Kaster corrected me and gave
this link in C#-nontech.

bdn.borland.com/article/0,1410,29460,00.html
<sigh>OK, I stand corrected!
So, please mentally replace all my references to 'CLX' above with
'VisualCLX'.
(grumble, I've grown so used to despising CLX (pardon, VisualCLX),
I find it difficult to accept that I'm now all of a sudden writing CLX apps
by a redefinition of terms... it's like being told my favorite food is now
called 'liver'... grumble).
Kristofer
 
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Alex Bakaev [TeamB]" wrote
Quote

In my mind, the most exciting part was a place where Borland invited
people to participate in a beta program.
<...>
Why don't you apply for a beta program and if you qualify you may get
all your answers sooner?
Hmmm.... actually I might consider that. <g>
Kristofer
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Kristofer Skaug" wrote:.
Quote

What's so interesting about being standards-compliant?
Is there a market demand for this kind of software code?

Forgive me for asking, as I come from the Delphi camp
where the "standard" language is completely irrelevant -
(Delphi *is* the standard).

When your code is standards-compliant it can be compiled by every
standards-compliant compiler. So the standards-compliant code you write will
compile without changes in Borland compiler for Win32, it will compile
without changes in GCC for Linux, and if I would know the name of a
standards-compliant C++ compiler on Mac I would compile it for Mac too.
Peter
 

{smallsort}

Re:Re: Open Letter to the Borland C++ Developer Community

"Peter Agricola" wrote
Quote
When your code is standards-compliant it can be compiled by every
standards-compliant compiler. So the standards-compliant code you write
will
compile without changes in Borland compiler for Win32, it will compile
without changes in GCC for Linux, and if I would know the name of a
standards-compliant C++ compiler on Mac I would compile it for Mac too.
Ok - but isn't this mainly a matter of syntax checking?
I mean, *you* write the code, the compiler can only say ack or ugh.
Plus, if you're that kind of cross-platform wizard, you'll probably
already know the subtle differences between compilers and
program for the common denominator? Just like you should be
aware of OS specific code and know how to work around it.
Kristofer
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Alex Bakaev [TeamB]" wrote
Quote

Yes, there is. Just search for boost, loki, ACE/TAO.
A Standard-compliant compiler allows one to use the
rich libraries that are out there.
I never knew it was a problem for BCB to ingest standard C/C++?
<ducking, not sure if I stomped on someone's sore toe!>
Kristofer
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Rodrigo Gómez" wrote
Quote

If I may ask, and this is "OT", why did you moved from Delphi to BCB?
I didn't! I use Delphi (99%), but I like to try to follow what's going on in
the BCB camp because many of team members use BCB, and I'm sort of
responsible for software development tooling issues in general. We've been
having lots of fun integrating Delphi and BCB code and binaries. For
example, lots of our generic Delphi library code and forms go directly into
BCB projects as needed, and BCB code is fed back into Delphi through DLL's
etc.
Kristofer
 

Re:Re: Open Letter to the Borland C++ Developer Community

Quote
Forgive me for asking, as I come from the Delphi camp
where the "standard" language is completely irrelevant -
(Delphi *is* the standard).
If I may ask, and this is "OT", why did you moved from Delphi to BCB?
Regards,
Rodrigo Gómez
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Saulo I. Regis" < XXXX@XXXXX.COM >wrote in message
Quote

"Chris Uzdavinis (TeamB)" < XXXX@XXXXX.COM >escreveu na mensagem
news: XXXX@XXXXX.COM ...
>
>Blame the lawyers. :)
>
>--
>Chris(TeamB);

LAWYERS... Let people do their work!!!
That's what you would think. But unfortunately there are always some people who
take advantage of such releases if those disclaimers aren't in there. Imagine
the clause that they could decide not to release an announced product wasn't
there. Theoretically, it is possible that a business decided (in a not so bright
moment) to adopt C++ as their language expecting this Borland product--maybe
because they've seen other Borland products perform well before or maybe because
there was something different in their coffee. Anyway, now they spend money on
the machines and hiring developers and all that. Then, Borland changes
plans--for whatever reason--and the company just lost a bunch of money. You can
bet your grandma that that's a lawsuit waiting to happen. So they just add as
many disclaimers as possible to avoid any such litigation. It's sad but
necessary.
sm
 

Re:Re: Open Letter to the Borland C++ Developer Community

"liz" wrote
Quote

It was never fully compliant.
Some standard constructs it couldn't handle?
Kristofer
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Sebastian Moleski" < XXXX@XXXXX.COM >writes:
Quote
>>This is great news. I hope...
>>"Build a development tool for the entire C++ market."
>>
>>includes .Net developers and at least legacy support for VCL.
>
>C++ does not run on .Net. Managed C++ is not C++.
>
>VCL is not C++ either.

Based on what do you make the assertion that it is not C++. I don't
see any standard violations.
Both the VCL and Managed C++ adds features incompatible with C++. For
instance: you can't use multiple inheritance with VCL classes, nor you
can't create a VCL object on the stack, so a VCL class is not a C++
class.
--
Oscar
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Remy Lebeau (TeamB)" < XXXX@XXXXX.COM >escreveu na mensagem
Quote
>Will there be .NET compatability?

Why should there be? Borland already has two existing .NET solutions -
C#Builder and Delphi for .NET. Introducing a third full-blown solution
for
C++ would be overkill, and would take focus away from pure C++ as well, I
would think.

Well I'll have to disagree on that one. We are talking about multiplatform
and C++.
.NET is a new platform just like Windows, Linux, Solaris, ... are!
If .NET will be a big player C++ Builder should deliver code there too. At
least I'll be expecting it.
If Borland is promising "write once and run anywhere" with its Front-End and
Back-End architecture I'll expect to reach .NET from my C++ code and not to
be forced to re-write everything in Delphi or C# to get there.
This would not make sense.
Saulo
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Kristofer Skaug" < XXXX@XXXXX.COM >writes:
Quote
"Sebastian Moleski" wrote
>
><..>a result is that gcc might even accept the same code but with
different
>semantic (not good), miscompile the same code (just interpret it
differently) or
>just outright reject code that ought to work.

I see, but I think compilation isn't even just the beginning. If
you want to write code that works for different compilers on
different platforms, you still actually have to *try* that compiler
and *test* the output.
Testing is a mandatory practice for any serious development work,
isn't it? We should re-test after any significant change, either on
the source code, the development environment or the execution
environment. So nothing new here.
Quote
Any kind of compliance Borland can put into BCB's compiler won't
guarantee that it will work on another platform, even if it may seem
to compile ok there.
Correct. If you use platform-neutral code, it should work everywhere,
once you use a good compiler.
Quote
Anyway, I understand the *intention* behind standardization, but the
practice seems so hazardous that compliance from Borland's side
would seem but a first-round insurance.
What we expect from Borland is to provide a good C++ compiler. IF they
add a cross-platform application framework, that's great.
BTW, there are millions of lines of C code out there that works on
dozens of platforms. This is mostly due to the fact that C has mature
compilers. C++ is not at the same maturity level of C, but things are
going better.
--
Oscar
 

Re:Re: Open Letter to the Borland C++ Developer Community

Sebastian Moleski wrote:
Quote

Based on what do you make the assertion that it is not C++. I don't see any
standard violations.

sm


It doesn't allow stack objects derived from TObject.
.a
 

Re:Re: Open Letter to the Borland C++ Developer Community

Kristofer Skaug wrote:
Quote
"liz" wrote

>It was never fully compliant.


Some standard constructs it couldn't handle?

Kristofer


Yes. Some wouldn't compile outright and some a buggy.
.a
 

Re:Re: Open Letter to the Borland C++ Developer Community

"Oscar Fuentes" wrote
Quote

What we expect from Borland is to provide a good C++ compiler. IF they
add a cross-platform application framework, that's great.
So what's wrong with the current one?
It sounds as if it's now some kind of "new feature" that it should work
correctly?
Quote
BTW, there are millions of lines of C code out there that works on
dozens of platforms. This is mostly due to the fact that C has mature
compilers. C++ is not at the same maturity level of C, but things are
going better.
OK
Kristofer