Board index » cppbuilder » Re: What is next for C++ Builder?

Re: What is next for C++ Builder?


2004-06-30 11:54:43 PM
cppbuilder60
Leroy Casterline wrote:
Quote
"Peter Agricola" < XXXX@XXXXX.COM >wrote:
>Real cross-platform with different integrated toolchains and RAD support for
>different frameworks and a superbe projectmanager resulting in ease of use
>and high productivity on different platforms.

Does this differ much from Eclipse?
As far as I know Eclipse's C++ support only works with gcc.
--
Gillmer J. Derge [TeamB]
 
 

Re:Re: What is next for C++ Builder?

Leroy Casterline wrote:
Quote

I wish I could get away from shutting down BCB so often, but one of the
linker bugs I'm facing seems to require that I delete the .TDS file, and
BCB holds that open even if I close my project. I wish they'd fix that,
too.
Yes, I have to delete the TDS file every so often too. The de{*word*81} stops hitting
breakpoints and deleting the .TDS and doing a Build makes the breakpoints work again.
Quote
Do you have reason to believe that their will be a BCB9?
Not with any certainty. I think it is a possibility. I have a hard time assigning a
probability to it.
Quote
D8 shipped at the tail end of last year. I think it was late January or
February before my SA version arrived (not that I've used it for
anything).
If they ship BCB v9 after D9 we won't see BCB v9 till some time next year.
Effectively we have to live with all the bugs for at least 6 months and maybe 12
months. Then again, maybe forever.
 

Re:Re: What is next for C++ Builder?

On Wed, 30 Jun 2004 13:01:23 UTC, "Dave Jewell"
< XXXX@XXXXX.COM >wrote:
Quote
"George Tihenea" < XXXX@XXXXX.COM >wrote

>Honestly, as much as I like CO++, I think Borland should focus all their
>efforts in Java. They must be the de facto Java dev tools provider if they
>want to survive because Windows platform is lost.

There's a lot to be said for your argument....maybe. :-)
However....

>I will be curious what is the position of JBuilder vs. other Java tools?
>Is it the market leader? Is it losing share to Sun or Oracle?

....my understanding is that JBuilder's market share is haemorrhaging badly,
partly as a result of freebies such as Eclipse.

One indicator about market share is the number of advertised jobs.
Worldwide search at monster.com shows
JBuilder: 164
Eclipse: 240
Delphi: 348
C++Builder:13 (actual number)
--
Per Johansson
Systems developer
per.johansson.name/
 

{smallsort}

Re:Re: What is next for C++ Builder?

"Gillmer J. Derge [TeamB]" wrote:
Quote
>>Real cross-platform with different integrated toolchains and RAD support
for
>>different frameworks and a superbe projectmanager resulting in ease of
use
>>and high productivity on different platforms.
>
>Does this differ much from Eclipse?

As far as I know Eclipse's C++ support only works with gcc.
And the RAD only with Java.
And it's sure not easy to use, that will say to get started with it. I once
looked at it but put it aside for a moment I have a _lot_ of time to install
it and get the pieces together.
Peter
 

Re:Re: What is next for C++ Builder?

"Peter Agricola" < XXXX@XXXXX.COM >escreveu na mensagem
Quote

This is nice, but not cross-platform. C++ is a cross-platform language and
I
want a cross-platform tool. Yes, cross-platform includes uncle Bill's
os's,
but others like Linux and Symbian are becoming increasingly important. I
will not ignore them.
Exactly!
Right on the target!
More. Borland announced that C++ Builder 6 was the beginning of such
cross-platform endeavour!
And CBX is a reassurance of it! The only fault is the lack of VCL support!
Why on earth it would hurt to be able to develop to Win32 at the same time
that others platforms would be at reach?
Saulo
 

Re:Re: What is next for C++ Builder?

Randall Parker < XXXX@XXXXX.COM >wrote:
Quote
If they ship BCB v9 after D9 we won't see BCB v9 till some time next year.
Effectively we have to live with all the bugs for at least 6 months and maybe 12
months. Then again, maybe forever.
I'm hoping that Delphi and BCB might be combined into one product. I had
reason to believe this might happen a few months ago, but don't remember
why I thought that. Perhaps something I read in Delphi/NonTech.
I think this would be a real win for the BCB community, and would add
the BCB customer base to the Delphi customer base. I think this would be
good for Borland, too.
I can wait for a year if I believe that it's coming. Wrapping BCB into
Delphi would make the wait more pleasant. My biggest concern is the
linker situation - I may be dead in the water sooner rather than later,
and I don't think my customer would be happy (or willing) to pay me to
break the project into DLL's, especially since I selected the
development platform.
 

Re:Re: What is next for C++ Builder?

"Peter Agricola" < XXXX@XXXXX.COM >wrote:
Quote
And the RAD only with Java.
And it's sure not easy to use, that will say to get started with it. I once
looked at it but put it aside for a moment I have a _lot_ of time to install
it and get the pieces together.
I've read that their is an Eclipse RAD C++ project well underway - on
this score, Eclipse and CBX seem to differ very little.
 

Re:Re: What is next for C++ Builder?

"Gillmer J. Derge [TeamB]" < XXXX@XXXXX.COM >wrote:
Quote
As far as I know Eclipse's C++ support only works with gcc.
I was under the impression that it could support other compilers, but
I'm likely wrong on that. I was under that impression that gcc was
fairly standards compliant, but I could well be wrong about that as
well. Do you know the score on gcc compliance?
 

Re:Re: What is next for C++ Builder?

Leroy Casterline wrote:
Quote
well. Do you know the score on gcc compliance?
I know it's very good. I couldn't assign a number to it. Certainly
better than BCB6, but that's a pretty low standard to hold yourself to. :-)
Either way, there are a variety of reasons my g++ may not be acceptable
for a particular project regardless of how well it conforms to the
standard. Some companies are afraid of free software and/or GPL
software. It certainly won't help you much with VCL projects. I'm not
sure how good the compiler or its runtime libraries are in terms of
performance (lousy, I think). ...
--
Gillmer J. Derge [TeamB]
 

Re:Re: What is next for C++ Builder?

Leroy Casterline < XXXX@XXXXX.COM >wrote in
Quote
<SNIP>
>IMHO, Borland had the right idea with C++BuilderX; they just need to
>fix it and give it a more conformant compiler.

Why do you think it was the right idea given (as you mention below)
Eclipse? What is it's compelling advantage over the freebies or IDE's
offered by the hardware vendors? (and yes, I've done ARM development
with CodeWimperer - not the best IDE but more than sufficient for most
purposes - and expensive enough that I doubt many managers would throw
it away so they could pay big bucks for CBX).

<SNIP>
The biggest problem with C++Builder was that it had a {*word*99}py IDE, IMO.
It was completely oriented towards RAD, but didn't take the needs of C++
coders into account -- code management, library management, change
control management, versioning...the whole thing is just borked in BCB.
What CBX *tried* to do was give C++ programmers a good IDE oriented
towards C++ guys rather than RAD-Pascal guys. So far so good; I
especially like the toolchain-agnostic approach, which is one of its main
advantages over a tool like Eclipse. Eclipse with CDT is tool-agnostic
*in theory*, but almost no one I know uses it for anything besides GCC
stuff because support is so hard to incorporate. And forget about trying
to get a non-GCC de{*word*81} working.
The main problem with CBX wasn't the *concept*, it was the *execution*.
The product was just too buggy and incomplete to use for all but trivial
tasks. There was no import tool for C++Builder projects (a mind-boggling
oversight, if you ask me). The BOE had lots of bugs. The build
configuration was confusing (and buggy as well).
It seems to me that Borland may be killing the whole C++ product line by
simply letting it wither on the vine. Doing nothing isn't an option:
many people (including myself) will simply use other products. I expect
that Eclipse's problems will be smoothed out eventually, and Microsoft's
C++ .NET product is about as good as it gets from Microsoft.
I'd be surprised if Borland's C++ market-share is more than a couple of
percentage points of the whole (if that). While this is mostly Borland's
own fault, it raises the question of whether or not Borland considers any
monies spent on C++ development as throwing good money after bad. I've
been wondering lately if the confused signals coming out of Borland on
the C++ front are indicative of some internal civil war: management suits
who want to kill the whole product line, moderates who want to keep it
alive under the Delphi penumbra, and purists who want a "programmer's
tool", i.e., C++BuilderX.
mr_organic
 

Re:Re: What is next for C++ Builder?

"Gillmer J. Derge [TeamB]" < XXXX@XXXXX.COM >writes:
[snip]
Quote
Either way, there are a variety of reasons my g++ may not be
acceptable for a particular project regardless of how well it conforms
to the standard. Some companies are afraid of free software and/or
GPL software.
Some people are afraid of being abducted by aliens, too.
Quote
It certainly won't help you much with VCL projects.
I'm not sure how good the compiler or its runtime libraries are in
terms of performance (lousy, I think). ...
For C++ code, it blows BCB out of the water. Furthermore, codegen bugs
are rare and {*word*118}. I trust the g++ optimizer, which is more I can
say of bcc32.
The real problem with GCC in Windows is lack of support for most MS
propietary C/C++ extensions. If you are on Standard C++ land, bcc32
only beats g++ on compiler speed and look *very* *bad* on every other
aspect.
--
Oscar
 

Re:Re: What is next for C++ Builder?

"Gillmer J. Derge [TeamB]" < XXXX@XXXXX.COM >wrote:
Quote
>well. Do you know the score on gcc compliance?

I know it's very good. I couldn't assign a number to it. Certainly
better than BCB6, but that's a pretty low standard to hold yourself to. :-)
<g>!
Quote
Either way, there are a variety of reasons my g++ may not be acceptable
for a particular project regardless of how well it conforms to the
standard. Some companies are afraid of free software and/or GPL
software. It certainly won't help you much with VCL projects. I'm not
sure how good the compiler or its runtime libraries are in terms of
performance (lousy, I think). ...
No, it won't help with VCL projects, and if performance is in the gutter
I guess it wouldn't matter much if it would. I don't recall reading
discussions regarding its performance, but gcc isn't a topic I follow so
I'm sure you have a better chance of being on-target on this than I do.
 

Re:Re: What is next for C++ Builder?

Oscar Fuentes < XXXX@XXXXX.COM >wrote:
Quote
Some people are afraid of being abducted by aliens, too.
Some fears may be irrational, but if they inhabit the little minds of
TPTB in the work place then irrationality is moot.
Quote
For C++ code, it blows BCB out of the water. Furthermore, codegen bugs
are rare and {*word*118}. I trust the g++ optimizer, which is more I can
say of bcc32.

The real problem with GCC in Windows is lack of support for most MS
propietary C/C++ extensions. If you are on Standard C++ land, bcc32
only beats g++ on compiler speed and look *very* *bad* on every other
aspect.
Do you know how gcc compares to other compilers (MS, Intel, etc) in
terms of code speed and size?
 

Re:Re: What is next for C++ Builder?

"mr_organic" < XXXX@XXXXX.COM >wrote:
Quote
The biggest problem with C++Builder was that it had a {*word*99}py IDE, IMO.
It was completely oriented towards RAD, but didn't take the needs of C++
coders into account -- code management, library management, change
control management, versioning...the whole thing is just borked in BCB.
I like the IDE, but I agree with you that the areas you mention have
been more or less^H^H^H^H^H^H^H^H^H^H^H^H nearly totally ignored, to the
detriment of both the product itself and the programmers who depend on
those features. But then that's true of the entire product - nearly
totally ignored for far too long.
Too bad Borland was unable to identify the problem; if they treat CBX
the same way they treated BCB it has no chance of success - and
indications to this point are that Borland has learned nothing from
their BCB experience.
Quote
Eclipse with CDT is tool-agnostic
*in theory*, but almost no one I know uses it for anything besides GCC
stuff because support is so hard to incorporate. And forget about trying
to get a non-GCC de{*word*81} working.
I didn't know that Eclipse has not realized the potential of agnosticity
(if that's a word).
Quote
The main problem with CBX wasn't the *concept*, it was the *execution*.
Yeah, I saw that myself. That, and the lack of promised bug fixes and
timely updates presents an ugly picture.
Quote
Doing nothing isn't an option:
many people (including myself) will simply use other products.
Given my large investment in 3rd party VCL components, all of which are
implemented in Delphi, my only real option is Delphi. Don't get me
wrong, I think Delphi is a great product, fast and powerful enough for
nearly any purpose. But I strongly prefer C/C++ syntax...
Quote
I expect
that Eclipse's problems will be smoothed out eventually, and Microsoft's
C++ .NET product is about as good as it gets from Microsoft.
From what I've read recently, I think your right about Eclipse. But in
that case, Borland has even less chance of making CBX fly. Regarding VS,
I'd say that it is much better than anything I'd expected from MS, but
only for .NET apps and I'm locked into Win32 for the foreseeable future.
Quote
I've
been wondering lately if the confused signals coming out of Borland on
the C++ front are indicative of some internal civil war
I've been thinking the same thing. I hope it matters (to me) who wins.
 

Re:Re: What is next for C++ Builder?

Leroy Casterline < XXXX@XXXXX.COM >writes:
[snip]
Quote
Do you know how gcc compares to other compilers (MS, Intel, etc) in
terms of code speed and size?
It strongly depends on the type of code. You can write test cases
that shows one compiler to be much faster than any other, but for my
real-life projects, with almost no floating point and middle to high
level of abstraction (lots of templates, types and tiny functions)
intel wins by a wide margin. GCC's performance is a bit behind MS. My
experience shows that tweaking GCC's optimization options often
brings faster executables than MS.
--
Oscar