Board index » cppbuilder » Re: Um, isn't it too late?

2003-08-12 02:35:23 AM
cppbuilder39
< XXXX@XXXXX.COM >wrote in message
have to use the VCL in projects, you can create raw-C++ projects as well.
Gambit
 

I have used Borland products since TurboC 1.0 when I was teaching myself C
programming. I have tried Microsoft products periodically through the
years, but the reason that I keep with Borland is that I can just "use it".
Early on (TurboC 1.0) I needed a de{*word*81}. Borland didn't have one, so I
bought the Microsoft C compiler, but I didn't understand the de{*word*81} and I
had no one to ask. Then, TurboC 1.5 arrived with it's de{*word*81} and it was
obvious to me how to use it. I know that Microsoft makes a fine product,
but the issue with "how do I" within the IDE keeps coming up for me. I use
VC6 on one project because that's what it was developed in and I only make
minor changes.
So, isn't it too late? Not for me. The thing that keep me with Borland is
that same thing many keep bi*ching about. The IDE and RAD.
-Michael
< XXXX@XXXXX.COM >wrote in message

Doubtful.
Name them.
. Ed

{smallsort}

"Ed Mulroy [TeamB]" < XXXX@XXXXX.COM >writes:
Diagnostics. IMO, EDG and GCC issues better diagnostics. I was told
that MSVC++ 7.1 is quite good too. Fixing template-related errors with
Borland is specially difficult.
Performance of executables. Compared against Intel and GCC, BCB shows
a disadvantage of about 50% for some of my apps that uses lots of
inlines, templates, data abstraction and other "advanced"
features. This difference grows if one fine-tunes the optimization
options of Intel/GCC.
stability. BCB has too much critical bugs for my taste. (Exception
handling, inlining, wrong code poduced for some templates,...)
Support. Intel releases an update every 3-4 weeks. GCC shines on this
aspect, too.
Third-party product availability. Often, you will find that, library
and auxiliary product vendors ignores Borland. IMO, even GCC/MinGW is
on a better position due to the Linux tools and libraries that were
ported to Win32.
--
Oscar
 

XXXX@XXXXX.COM wrote:
believe that C++ is too complex, use some simpler language.
was behind C++ Builder in C++ compliance, but Borland has created a RAD C++
environment for 8-9 years ago while VC++ did so only recently. The real
question is: why would anyone bother to use VC++ ?
plenty of them.
.NET 2002 or VS .NET 2003 yet in over two years. But I am sure you will be
scurrying to use them shortly.
their "supporters" no doubt. A very clever troll, but no intelligent person
is fooled by it.
 

MS who was rather poor in compliance now has released a new compiler with
better compliance. From your comment I take it that when on manufacturer
releases an updated version of their compiler, then you feel the others with
competing products should withdraw from the market or at least should not be
used. I dissagree.
bothers me. I do not know it as a great compiler. As far as I know it does
not run under Windows which, given that you also compare to MS, is the
platform under discussion.
. Ed

In article < XXXX@XXXXX.COM >,
Ed Mulroy [TeamB] < XXXX@XXXXX.COM >wrote:
for a long time.
That said, that alone does not necessarily mean instability,
certainly doesn't have to mean it in the sense that something
is wrong. I agree though that folks should tend to be
working with something stable. OTOH, not producing upgrades
also in and of itself doesn't mean the product is stable.
I guess what I'm saying is that w/o a context, not much
can be said either way, since each way has pros and cons.
--
Greg Comeau/4.3.3:Full C++03 core language support + more Windows backends
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

You and I spoke of GCC, not of MinGW. I know little about MinGW. If I
remember correctly people run a Win95 emulator to run GCC under Windows.
That about guarantees that the compiler, even if perfect, would still have
performance problems.
. Ed

"Ed Mulroy [TeamB]" < XXXX@XXXXX.COM >writes:
philosophically, I'd take a correct but slow compiler over a fast
but buggy compiler because I value compliance more than compile
speed. (Though compile speed is important too.)
--
Chris(TeamB);
 

"Ed Mulroy [TeamB]" < XXXX@XXXXX.COM >writes:
releases are far better than C++ Builder's.
 

"David B. Held" < XXXX@XXXXX.COM >writes:
Win32 application. Ditto for the executables produced with them.
As far as the runtime environment is concerned, MinGW's applications
are very similar to MSVC's, as they use MSVCRT.DLL as its C library.
fastest C++ compiler, while GCC is the slowest. A 3x difference is
typical.
Plus GCC/MinGW has weak support for some Windows-specific features,
like COM, MS deviations (MFC), structured exceptions, etc.
That said, GCC is better than Borland on every other feature you
expect from a modern C++ compiler. IMO, GCC definitively is the best
for cross-platform development and the second best for learning C/C++,
after EDG/Comeau.
--
Oscar
 

Oscar Fuentes wrote:
[...]
templates: GCC catches many errors at the point of definition that bcc
ignores until the point of instantiation.
Cheers,
--
Nicola Musatti
Team Thai Kingdom
 

Nicola Musatti < XXXX@XXXXX.COM >writes:
--
Chris(TeamB);
 

Ed Mulroy [TeamB] wrote:
provide support to our customers and we *know* that it is as buggy
internal beta)
- standard compliance (code written in VC++ can be compiled in other C++
compilers but not in BCB)
- optimization quality
- ease of project management (this is a question of choice but I find
VC's approach to project configurations to be much easier to handle)
--
Eugene Mayevski
Flexible cross-platform information exchange framework
www.msgconnect.com/
 

Delphi whenever reasonably possible. But, I'm not totally incompetent
at C++, and do have to use it some.
I used to work at a small shop that primarily used Borland's tools. I
moved to a new company that uses primarily Visual Studio. We are
moving to .NET, but still have a lot of customers using our legacy
products. As such, we have both VS 6 and VS.NET 2003 in house usage.
I clearly see that Delphi is much better than VB, at least in the VS 6
incarnation, I haven't used the .NET version yet.
C++. I don't think that C++ Builder is quite as old as you say, I was
thinking more like 6 or 7 years, though you could be right.
Also, my understanding is that you are correct, Borland's compiler was
more standards compliant that MS for many years. On the other hand, I
believe MS's compiler has been less buggy than Borland's for many
years, along with better runtime performance for many years. It
appears that while the market might have preferred all four (RAD,
compliant, low bugs, high performance) the market seems to have
preferred the product with the last two instead of the first two. But
if the market already preferred the product with the last two, how
much more so now that VS scores best in all four?
Now of course it is at least possible for a new Borland product to
score better or best in all four areas, but honestly, I just don't see
it happening. I could be wrong, but I don't see it. The amount of
investment that Borland would have to do would be enormous.
pure C++ programmers, the VCL isn't ideal being based on Delphi and
its object model. And I can see how C++ Builder users would be
inclined to want a real C++ based framework. But at least one thing
it has going for it is a very, very rich third party market support.
Any new framework, from Borland, MS, or anybody else, will take time
to get the kind of market support that the VCL has. But a framework
has got to really offer something that the other guy's doesn't or it
just won't ever get that kind of market support. OWL suffered from
being just not enough better than MS's framework to be worth
developing for. Some people did, but not enough.
When the VCL came out, that really had something going for it that
nothing MS had could touch. Now, I find it hard to imagine that
Borland will be able to do that again. I could be wrong, I just don't
see it.