The future of C++ and VC++ ... 
Author Message
 The future of C++ and VC++ ...

Hi,
There has been some talk about the future directions of C++ that was
published on Dr. Dobbs Tech Net Cast. Please make an attempt to see (and
also hear) this before you comment here :) -
http://www.*-*-*.com/

First of all let me say that in general I am a supporter of VC++. But having
said that I also want to voice some concerns here. My biggest worry is that
will VC++ be upto it? Will there be enough support in VC++ *then* for us
mere mortal users to be able to use the new libraries that the Std committee
will come up with? I somehow think that there was a feeling there that some
compilers wouldn't become good enough and hence spoil the common good. There
were lots of remarks about Visual C++ also. And some good (favorable) ones
also I'd dare say :)

Before someone says  - "Oh No! Here we go again on that tangent ..." I'd
like to say that I am *genuinely* concerned. What kind of a job will VC++ do
tracking the future developments? I'd like some people from the VC++ team to
comment here.

I remember that at one time one member of a VC++ team even boasted that
while the Std was still not finalized they had even experimentally
implemented some features being considered. And that they were later bitten
by those features not quite making it or something. That was quite some
exceptional tracking. While I am not looking at something like that,
however, do we have any assurances from the VC++ team that they will do an
honest job of tracking these things?

VC++ being the {*filter*} compiler on the {*filter*} computer platform on earth
gives it quite some power. Think about that. There is no point in talking
new ideas if you may not be able to use it usefully in the future. So I
guess it would be nice if we all had a better idea about that.

Thanks,
Shiv



Mon, 13 Oct 2003 21:49:33 GMT  
 The future of C++ and VC++ ...


Quote:
> Hi,
> There has been some talk about the future directions of C++ that was
> published on Dr. Dobbs Tech Net Cast. Please make an attempt to see (and
> also hear) this before you comment here :) -
> http://www.*-*-*.com/

WOW. That WAS interesting.

Quote:
> First of all let me say that in general I am a supporter of VC++. But having

Me too. And VC7 is really great at things like ... optimization, to name one.

Quote:
> said that I also want to voice some concerns here. My biggest worry is that
> will VC++ be upto it? Will there be enough support in VC++ *then* for us
> mere mortal users to be able to use the new libraries that the Std committee
> will come up with?

I was wondering the same.

Quote:
> I somehow think that there was a feeling there that some
> compilers wouldn't become good enough and hence spoil the common good. There
> were lots of remarks about Visual C++ also. And some good (favorable) ones
> also I'd dare say :)

Yes, but they mostly were to VC6.

Quote:
> Before someone says  - "Oh No! Here we go again on that tangent ..." I'd
> like to say that I am *genuinely* concerned. What kind of a job will VC++ do
> tracking the future developments? I'd like some people from the VC++ team to
> comment here.

Me too. That IS important.

Ronald already told us that next version will be mostly toward compliance. And
we know some points in the standard are not clear per-se so 100% compliancy is
almost nonsense. But between VC7 and 100% ISO there's a pretty long way. I have
no doubt that the compiler will get "more compliant". Expecially because there
are a lot of libraries now that do make use of stuff that is not yet there in
VC7 (boost, for all).

But that is exactly my concern: is VC gonna support boost & C or is it gonna
support ISO ? Ok, VC6 was out before the standard was. Ok .NET took a lot of
effort. So what? The standard is not something that pops up from a wormhole.
Read ahead.

Quote:
> I remember that at one time one member of a VC++ team even boasted that
> while the Std was still not finalized they had even experimentally
> implemented some features being considered. And that they were later bitten
> by those features not quite making it or something. That was quite some
> exceptional tracking. While I am not looking at something like that,
> however, do we have any assurances from the VC++ team that they will do an
> honest job of tracking these things?

If honest==commercially wise you can rest very assured! But your example is like
stretching things the other way. As you say nobody requires it from MS that they
support features before they are in the standard. But at the same time a company
like MS should not take years to get compliant. The standard does differ from
the standard draft. But MS has people in the comitee. And many changes to the
draft are not really a surprise (fiew are, but most are not).

Quote:
> VC++ being the {*filter*} compiler on the {*filter*} computer platform on earth
> gives it quite some power. Think about that.

I do. And that scares me. If MS is gonna support the standard (and lately it
seems like they are trying to support standards) then C++ will live. If MS is
not ... well, the MP3 recalls the story of Pascal and Borland.

Currently C++ is the language I use most. But if VC does steer away from ISO
and/or it does not catch up with the time lost for .NET ... I'm pretty sure I'll
be forced to go back and split projects between Ada for embedded and Smalltalk
for large.

Quote:
> There is no point in talking
> new ideas if you may not be able to use it usefully in the future. So I
> guess it would be nice if we all had a better idea about that.

Most importantly: large part of the development of libraries like boost was done
on Linux. And this not only because Linux is free. Also because it's native of
more compliant compilers. If VC was having a point release in late '99 or early
2000 and that was as compliant as VC8 promises to be, eventually things would be
different.

Also note there's a difference between library and core. Of course I'd like to
have VC come with a good and compliant library. But FIRST I want the core to be
fixed. If there's a bug in the library I can fix it. But if there's a bug in the
core ... VC's not open source, is it?

Many of the ideas discussed in the meeting have to do with core, baut MOST of
them have to do with library. Therefore, again, bringing the core as close to
ISO as possible in as short time as possible will have the side effect of
letting us experiment with new library ideas.

Also, even if it is possible that some of the current core specs will change,
still the current standard is a landmark. I absolutelly would rather have a
compiler that does support the current standard even if we know things will
change in the future (unless we are talking of an ambiguous or unimplementable
feature).

--

Andrea Ferro

---------
Brainbench C++ Master. Scored higher than 97% of previous takers
Scores: Overall 4.46, Conceptual 5.0, Problem-Solving 5.0
More info http://www.*-*-*.com/



Mon, 13 Oct 2003 23:55:24 GMT  
 The future of C++ and VC++ ...
Hi Andrea,

Quote:
> Me too. And VC7 is really great at things like ... optimization, to name

one.

Indeed VC7 is *much* better (so much better that I find a pain to go back to
6!). I also like the new IDE much better than the old one :)

Quote:
> If honest==commercially wise you can rest very assured! But your example
is like
> stretching things the other way. As you say nobody requires it from MS
that they
> support features before they are in the standard. But at the same time a
company
> like MS should not take years to get compliant.

Agreed. Also, I do honestly believe there's another peril there. MS does a
lot of things based on what it perceives to be the customer's wants (even
though they come up with some of their own), so, regardless of how much some
of the VC++ team wants a better compiler, if MS as a corporation doesn't
feel pressed to provide it, tehy won't get the necessary resources to do so.
This _is_ the time to let MS know about it so that a compliant VC++ 8.0 does
indeed happen (anybody want to start a petition? <g>)

Quote:
> But MS has people in the comitee.

They should have. Perhaps Ronald can comment, but it was my understanding
that last time, before the standard was finalized, MS lost voting rights
because of missing two sessions in a row. Let me just say this: That is
*aboslutely* unacceptable for one of the major C++ compiler players out
there. Period. There _are_ no buts. When work on the standard is formally
restarted, I expect MS to get much more involved. If something like this
were to happen again, I promised I'd go to redmond just to kick some butts
around! :)

Quote:
> Also note there's a difference between library and core. Of course I'd
like to
> have VC come with a good and compliant library. But FIRST I want the core
to be
> fixed.

Aboslutely agree. Once the core is there, I can just go and get an updated
library from Dinkumware and be done with it. But if it isn't...well, you
know how it works.

--
Tomas Restrepo [VC++ MVP]

http://www.mvps.org/windev/



Tue, 14 Oct 2003 01:33:53 GMT  
 The future of C++ and VC++ ...
From the past discussion on this newsgroup, it has become clear that
Microsoft has demanded that VC++ meet Microsoft's vision of .NET before it
meets any C++ standard. That is, VC++ compliance with the C++ standard is
secondary (worse, it could even be said to be non-existent) to Microsoft's
stated goals.

There is nothing I see that would compel Microsoft to continue C++
compliance. Not only would compliance be an unnecessary expense, but it
would allow developers to remain outside of the .NET strategy. The latter
has got to be seen as counter productive by Microsoft.



Quote:
> Hi,
> There has been some talk about the future directions of C++ that was
> published on Dr. Dobbs Tech Net Cast. Please make an attempt to see (and
> also hear) this before you comment here :) -
> http://technetcast.ddj.com/tnc_play_stream.html?stream_id=560

> First of all let me say that in general I am a supporter of VC++. But
having
> said that I also want to voice some concerns here. My biggest worry is
that
> will VC++ be upto it? Will there be enough support in VC++ *then* for us
> mere mortal users to be able to use the new libraries that the Std
committee
> will come up with?



Tue, 14 Oct 2003 01:48:15 GMT  
 The future of C++ and VC++ ...

Quote:
> Hi Andrea,
..
> > If honest==commercially wise you can rest very assured! But your example
> is like
> > stretching things the other way. As you say nobody requires it from MS
> that they
> > support features before they are in the standard. But at the same time a
> company
> > like MS should not take years to get compliant.

> Agreed. Also, I do honestly believe there's another peril there. MS does a
> lot of things based on what it perceives to be the customer's wants (even
> though they come up with some of their own),

I read that as "MS does a lot of things based on what it perceives to be the
customer's wants and on what it perceives should be the customer wants".

Quote:
> so, regardless of how much some
> of the VC++ team wants a better compiler, if MS as a corporation doesn't
> feel pressed to provide it, tehy won't get the necessary resources to do so.
> This _is_ the time to let MS know about it so that a compliant VC++ 8.0 does
> indeed happen (anybody want to start a petition? <g>)

The big problem is that MS as a corporation "feels" the user community on the
base of user community response, but they consider a response only an active
feedback. No feedback absence is considered a response. So the fact that those
wanting standard stuff just go to open-source compilers (and OS) is perceived as
either:
1) no or just fiew complaints received, therefore no actual interest
or
2) ISO is important to those radicals of GNU more than to serious business
developers

Things seem to be getting a little bit better lately, but still ...

Quote:
> > But MS has people in the comitee.

> They should have. Perhaps Ronald can comment, but it was my understanding
> that last time, before the standard was finalized, MS lost voting rights
> because of missing two sessions in a row.

That means not they have no people in the comitee. It means they were not
interested in comitee proceedings. And that is exactly how much MS as a
corporation rated ISO.

Quote:
> Let me just say this: That is
> *aboslutely* unacceptable for one of the major C++ compiler players out
> there. Period. There _are_ no buts. When work on the standard is formally
> restarted, I expect MS to get much more involved. If something like this
> were to happen again, I promised I'd go to redmond just to kick some butts
> around! :)

Actually having MS voting in the comitee is something I do not actually pay care
to. From the point of view of historical "embrace and extend" equation, MS would
better keep hands off the extend side. I DO want MS to show INTEREST in the
comitee proceedings more on the embrace side of the equation. And voting is more
extend than embrace.

On the other side this means not that MS should just sit and watch. Expecially
because COM, .NET, Memory Mapped Files and other MS stuff is going to be on
topic as usefull de-facto industry standard stuff. Still, exactly for that
reason, MS should be in the discussion as an advisor of what they perceive these
tecnologies future can be. Then let the comitee decide on the language.

I'm not saying MS should not vote. I'm saying it is not extremely important
wether it votes or not. It should speak the technical stuff and it should
embrace the result. Let the comitee decide on the extend part. MS will do extend
by taking part to the comitee, not by taking over the language (or the comitee).

Quote:
> > Also note there's a difference between library and core. Of course I'd
> like to
> > have VC come with a good and compliant library. But FIRST I want the core
> to be
> > fixed.

> Aboslutely agree. Once the core is there, I can just go and get an updated
> library from Dinkumware and be done with it. But if it isn't...well, you
> know how it works.

And if P.J. & C has not it right yet, I can still debug the library and correct
it (that is again embrace and extend, isn't it). Try that with the compiler!

--

Andrea Ferro

---------
Brainbench C++ Master. Scored higher than 97% of previous takers
Scores: Overall 4.46, Conceptual 5.0, Problem-Solving 5.0
More info http://www.brainbench.com/transcript.jsp?pid=2522556



Tue, 14 Oct 2003 01:53:06 GMT  
 The future of C++ and VC++ ...

Quote:
> From the past discussion on this newsgroup, it has become clear that
> Microsoft has demanded that VC++ meet Microsoft's vision of .NET before it
> meets any C++ standard. That is, VC++ compliance with the C++ standard is
> secondary (worse, it could even be said to be non-existent) to Microsoft's
> stated goals.

That is exatcly the concern. But now .NET is supported bu VC. So the question
becomes: is ISO now gonna be top priority or are we to expect VC.NUTs ?

Quote:
> There is nothing I see that would compel Microsoft to continue C++
> compliance.

Well. Maybe not compell, but concern for sure yes. The point is how much
concern.

Quote:
> Not only would compliance be an unnecessary expense, but it
> would allow developers to remain outside of the .NET strategy. The latter
> has got to be seen as counter productive by Microsoft.

Nonsense. If that was so then VC would be just banned! For .NET development I
would really really pick C# or VB (among the current languages) and Smalltalk or
python or even Java (if it will be supported on .NET, we know MS cannot do it,
but others can). The type system and the memory management native to .NET just
happen to nicely fit all languages but C++!

There are many reasons to keep C++ alive for Windows. One is that device drivers
and OSs are not actually developed in VB. One is that even if .NET is strategic
to win the development platform war in the internetworked server era (figure the
www wars we had. They were a networking-for-umans fight. Next is going to be
networking-for-computers. And .NET is IE/IIS of these) that means not that all
code is gonna be .NET.

CAD/CAM, Embedded and Real-time systems, Avionics, Tactical environments,
communication, media codecs, satellite etc. There's a LOT of development that is
going to still be done outside of any VM. And that rules out Smalltalk, VB, Java
and all .NET development. They need Assembler, C, C++, Ada and similar "iron lev
el" tools.

I know that CAD/CAM could also make good use of .NET. Also. I know that
communications could also make good use of .NET. Also. But that's mostly at the
presentation level. I do not foresee DSP chips programmable in VB! I do not
think a MPEGII codec would benefit from running on a VM (it would be a portable
nonsense!). I'll never ever develop tactical SW to run on .NET or JVM (actually
not even do it in C++!!).

Are you really sure NT wants to stay out of those worlds as both target OS and
development OS? I would not put NT on some embedded systems, but I would on
others. And I would try to use it as a development platform for all of them.

To be honest I'd like VC to not only be more ISO compliant, but to also target
non NT systems!! It would be a dream to be able to use VC all the way throu to
producing rommable binary for non intel architectures! It would be pretty nice
to be able to have the final output be not a windows executable but an
architecture independetly optimized single C source to feed to an embedded
development tool!

But I'm dreaming. However ISO, IMO is a "feature" we can ask from MS.

--

Andrea Ferro

---------
Brainbench C++ Master. Scored higher than 97% of previous takers
Scores: Overall 4.46, Conceptual 5.0, Problem-Solving 5.0
More info http://www.brainbench.com/transcript.jsp?pid=2522556



Tue, 14 Oct 2003 03:18:07 GMT  
 The future of C++ and VC++ ...
Hi Andrea,

Quote:
> I read that as "MS does a lot of things based on what it perceives to be
the
> customer's wants and on what it perceives should be the customer wants".

Yes, that's a good way to put it :)

Quote:
> The big problem is that MS as a corporation "feels" the user community on
the
> base of user community response, but they consider a response only an
active
> feedback. No feedback absence is considered a response. So the fact that
those
> wanting standard stuff just go to open-source compilers (and OS) is
perceived as
> either
> 1) no or just fiew complaints received, therefore no actual interest
> or
> 2) ISO is important to those radicals of GNU more than to serious business
> developers

Exactly, which is why I say it is time to "speak up".

Quote:
> That means not they have no people in the comitee. It means they were not
> interested in comitee proceedings.

I know that.

Quote:

> Actually having MS voting in the comitee is something I do not actually
pay care
> to

I do for one simple reason: If MS feels that they cannot influence how the
standard goes (hoewever lightly), then I have to be suspicious of how much
would MS be willing to cover the expenses of getting someone out there to
attend the meetings. That doesn't mean I want MS to take over the ISO
comittee, it just means I think we might have a better chance of getting MS
more concerned and involved in standard C++.

Then again, I might just be wrong...

--
Tomas Restrepo



Tue, 14 Oct 2003 04:11:12 GMT  
 The future of C++ and VC++ ...
Quote:
> > First of all let me say that in general I am a supporter of VC++. But
having

> Me too. And VC7 is really great at things like ... optimization, to name

one.

Compile speed also as one. I've seen horror stories of projects where a
large body of STL heavy C++ code takes a couple hours to compile on a well
known *NIX OS and its C++ compiler but that compiles in 20 minutes flat
using VC++! :) This with the *NIX one being a dual CPU machine. Go figure.

Quote:
> I was wondering the same.

I hope more people wonder and ask here :)

Quote:
> Yes, but they mostly were to VC6.

Nopes. Even VC7 was also mentioned. If I was correct that was Scott Meyers?
And he was once spotted on this news group asking some setup related
questions about VC7. Seemed mighty pleased that MS was supporting the new
for scope. Was that one of the things that MS voted against in the
committee? :)

Quote:
> are a lot of libraries now that do make use of stuff that is not yet there
in
> VC7 (boost, for all).

But boost does have so many workaounds for VC++. That brings one of my fears
... We are all gonna be deprived of good libraries and the ability to write
code using innovative techniques because of everyone waiting for VC++ to
catch up ... And if there is another huge gap be{*filter*} VC7 and VC8 then I am
positively gonna think of using less of C++ ...

Quote:
> I do. And that scares me. If MS is gonna support the standard (and lately
it
> seems like they are trying to support standards) then C++ will live. If MS
is
> not ... well, the MP3 recalls the story of Pascal and Borland.

Thats why I say I might think of giving up on C++ as my main tool and I'm
sure a lot of others also will if things don't improve. It would be so sad
for one of the best programming languages. Lets just hope that this thinking
is not cassandra-ish ...

Quote:
> Many of the ideas discussed in the meeting have to do with core, baut MOST
of
> them have to do with library. Therefore, again, bringing the core as close
to
> ISO as possible in as short time as possible will have the side effect of
> letting us experiment with new library ideas.

Quite true. Exactly my sentiments. Only I am worried it shouldn't be so late
to cause us to miss the boat ... And to drive more and more VC++ users into
the don't-care-give-up state ...

Quote:

> still the current standard is a landmark. I absolutelly would rather have
a
> compiler that does support the current standard even if we know things
will
> change in the future (unless we are talking of an ambiguous or
unimplementable
> feature).

Me too. And also hope that the small/minor changes that might come in the
later std are also done soon enough that we don't have to wait another 3-4
years for those minor things also. They wouldn't matter so much since they
would be minor. But they'd be pretty darn irritating like the for scope as
of now.

Thanks,
Shiv



Tue, 14 Oct 2003 04:13:42 GMT  
 The future of C++ and VC++ ...

Quote:
> Indeed VC7 is *much* better (so much better that I find a pain to go back
to
> 6!). I also like the new IDE much better than the old one :)

Indeed Intellisense in VC7 does seem much more like an advanced C++ user
than the amateur it was in VC6 :) That alone is worth it ..

Quote:
> This _is_ the time to let MS know about it so that a compliant VC++ 8.0
does
> indeed happen (anybody want to start a petition? <g>)

And soon enough also :) ... Start the petition and I'll whole heartedly join
in with quite a few friends ...

Quote:
> They should have. Perhaps Ronald can comment, but it was my understanding
> that last time, before the standard was finalized, MS lost voting rights
> because of missing two sessions in a row. Let me just say this: That is
> *aboslutely* unacceptable for one of the major C++ compiler players out
> there. Period. There _are_ no buts. When work on the standard is formally

Tsk tsk tsk ... I wonder what excuse there is for that!

Quote:
> Aboslutely agree. Once the core is there, I can just go and get an updated
> library from Dinkumware and be done with it. But if it isn't...well, you
> know how it works.

Yeah. That and there will be small tweaks to the core which if not followed
by VC++ will once again land us in the same kind of scenario if not that
much aggravated but irritating nonetheless.

Thanks,
Shiv



Tue, 14 Oct 2003 04:21:52 GMT  
 The future of C++ and VC++ ...

Quote:

> Microsoft has demanded that VC++ meet Microsoft's vision of .NET before it
> meets any C++ standard. That is, VC++ compliance with the C++ standard is
> secondary (worse, it could even be said to be non-existent) to

While I have nothing against the work done for .NET, it is indeed very sad
that they couldn't find more engineers to work side by side on the
compliance issues. Given the resources they can muster, they could just
possibly have pulled that feat also. Now that would really have impressed
the world! But sigh ... Thats all wishful thinking now ...
(I personally also blame Java for getting us VC++ users in this shit)

Quote:
> There is nothing I see that would compel Microsoft to continue C++
> compliance. Not only would compliance be an unnecessary expense, but it
> would allow developers to remain outside of the .NET strategy. The latter
> has got to be seen as counter productive by Microsoft.

Thats like putting it very radically. So we are all fools to continue using
C++ and VC++ for that matter and we should just give up and go home? :) That
would be a very big insult to the legions of C/C++ programmers who even by
MS figures are the single largest community of developers after Visual Basic
...

I think in some way we all VC++ users are also to blame for not shouting
enough that we care about these things. Also as VC++ users of not promoting
enough of Std C++ library features over equivalent platform specific
libraries. Sure when you do not have something in the Std or it doesn't fit
the bill then by all means use platform specific libs. But when something
exists in the Std and it fits the bill then we *should* be using more of it.
We also have our work cut out in some way of educating people around us. I
try and do my bit. And I'm sure quite a few of you too do this. But if a
significantly lot more people also do then it would be all the more better.

Thanks,
Shiv



Tue, 14 Oct 2003 04:24:26 GMT  
 The future of C++ and VC++ ...
We since  have regained voting right, and in fact have attended all meetings
during the last 2 years.

Jonathan Caves, one of the leads of the compiler is the primary Microsoft
delegate and I am the alternate.

-Ronald-


Quote:
> Hi Andrea,

> > Me too. And VC7 is really great at things like ... optimization, to name
> one.

> Indeed VC7 is *much* better (so much better that I find a pain to go back
to
> 6!). I also like the new IDE much better than the old one :)

> > If honest==commercially wise you can rest very assured! But your example
> is like
> > stretching things the other way. As you say nobody requires it from MS
> that they
> > support features before they are in the standard. But at the same time a
> company
> > like MS should not take years to get compliant.

> Agreed. Also, I do honestly believe there's another peril there. MS does a
> lot of things based on what it perceives to be the customer's wants (even
> though they come up with some of their own), so, regardless of how much
some
> of the VC++ team wants a better compiler, if MS as a corporation doesn't
> feel pressed to provide it, tehy won't get the necessary resources to do
so.
> This _is_ the time to let MS know about it so that a compliant VC++ 8.0
does
> indeed happen (anybody want to start a petition? <g>)

> > But MS has people in the comitee.

> They should have. Perhaps Ronald can comment, but it was my understanding
> that last time, before the standard was finalized, MS lost voting rights
> because of missing two sessions in a row. Let me just say this: That is
> *aboslutely* unacceptable for one of the major C++ compiler players out
> there. Period. There _are_ no buts. When work on the standard is formally
> restarted, I expect MS to get much more involved. If something like this
> were to happen again, I promised I'd go to redmond just to kick some butts
> around! :)

> > Also note there's a difference between library and core. Of course I'd
> like to
> > have VC come with a good and compliant library. But FIRST I want the
core
> to be
> > fixed.

> Aboslutely agree. Once the core is there, I can just go and get an updated
> library from Dinkumware and be done with it. But if it isn't...well, you
> know how it works.

> --
> Tomas Restrepo [VC++ MVP]

> http://www.mvps.org/windev/



Tue, 14 Oct 2003 05:15:36 GMT  
 The future of C++ and VC++ ...

..

Quote:
> > feedback. No feedback absence is considered a response. So the fact that
> those
> > wanting standard stuff just go to open-source compilers (and OS) is
> perceived as
> > either
> > 1) no or just fiew complaints received, therefore no actual interest
> > or
> > 2) ISO is important to those radicals of GNU more than to serious business
> > developers

> Exactly, which is why I say it is time to "speak up".

Those of us still playing around with VC are speaking. But we are those that
need ISO compliance for less than 100% of the code or that need other MS stuff
for more then 0% of code. The majority of people actually needing better
compliancy also need not other MS proprietary stuff. Hence they just forget VC
and go to other compilers.
Their silence is their speaking.
In general C++ NGs, each time the talk get ISO, VC (that's 6 actually) is rated
second worst (Borland succeeds in sometime being less compliant).

Quote:
> I do for one simple reason: If MS feels that they cannot influence how the
> standard goes (hoewever lightly), then I have to be suspicious of how much
> would MS be willing to cover the expenses of getting someone out there to
> attend the meetings. That doesn't mean I want MS to take over the ISO
> comittee, it just means I think we might have a better chance of getting MS
> more concerned and involved in standard C++.

ROTFL. Man, MS defitively DOES have the money to send ONE person there! I mean
one per group, of course. Believe me: there will be much more C++ people going
from VC6 to VC7 because of the compliancy than because of .NET. This may not be
true for VB. But for VC it is. And there'll be people going from VC to C#.NET.
Do you really believe C++ under Windows is used only to develop windows
applications? It is not. And VC is at risk of loosing that part.

Also I did not mean that MS should not vote. Of course it can be part of it. But
as a matter of fact the discussion is more important than the vote, in deciding.
If I was MS I'd rather be the attorney than the jury!

--

Andrea Ferro

---------
Brainbench C++ Master. Scored higher than 97% of previous takers
Scores: Overall 4.46, Conceptual 5.0, Problem-Solving 5.0
More info http://www.brainbench.com/transcript.jsp?pid=2522556



Tue, 14 Oct 2003 04:46:21 GMT  
 The future of C++ and VC++ ...


...

Quote:
> While I have nothing against the work done for .NET, it is indeed very sad
> that they couldn't find more engineers to work side by side on the
> compliance issues. Given the resources they can muster, they could just
> possibly have pulled that feat also. Now that would really have impressed
> the world! But sigh ... Thats all wishful thinking now ...
> (I personally also blame Java for getting us VC++ users in this shit)

hmmm. Now that MS is out of Java ... are they moving VJ stuff to C# or to VC? If
I remember well MS had a huge hiring campaign fiew years ago. And that was for
.NET and C#. But VC an VJ were already stuffed. So if the new hiring was for C#
maybe VJ team could go to VC.

Quote:
> I think in some way we all VC++ users are also to blame for not shouting
> enough that we care about these things. Also as VC++ users of not promoting
> enough of Std C++ library features over equivalent platform specific
> libraries. Sure when you do not have something in the Std or it doesn't fit
> the bill then by all means use platform specific libs. But when something
> exists in the Std and it fits the bill then we *should* be using more of it.
> We also have our work cut out in some way of educating people around us. I
> try and do my bit. And I'm sure quite a few of you too do this. But if a
> significantly lot more people also do then it would be all the more better.

Reread that keeping this in mind: the platform specific stuff works as expected,
the standard stuff works in a platform specific way.

--

Andrea Ferro

---------
Brainbench C++ Master. Scored higher than 97% of previous takers
Scores: Overall 4.46, Conceptual 5.0, Problem-Solving 5.0
More info http://www.brainbench.com/transcript.jsp?pid=2522556



Tue, 14 Oct 2003 05:19:00 GMT  
 The future of C++ and VC++ ...


...

Quote:
> Nopes. Even VC7 was also mentioned. If I was correct that was Scott Meyers?
> And he was once spotted on this news group asking some setup related
> questions about VC7. Seemed mighty pleased that MS was supporting the new
> for scope. Was that one of the things that MS voted against in the
> committee? :)

Yes. Scott is one of the fiew guys that is playing with many compilers. But the
larger thread we had on comp.std.c++ was mainly related to the mechanics of
making generic and specific versions of functions. And in it's specific code MS
failed twice: once becouse of partial ordering of templated functions (without
that there's no real way of making generic and specific versions work!). The
second ... well, he did not pick ints as the basic type to overload on, he
picked chars. And as soon as a string literal came into the game VC7 failed
again because strings are not const!

My feeling is that Scott has not bein playing with C++ daily on Windows platform
for the last years. Therefore he's sort of "taking a look" at things. And you
know: we can appreciate VC7 because it's much better than VC6, but from an
outsider point of view, when code is core not coding, the IDE, compiler speed,
optimizations, proprietary extensions, platform APIs and stuff are just
no-check. And VC compared to other compilers ...

Here're tables from those posts:

Quote:
>                   //  VC   BCC  MWCW g++ Como
>                   //  --- ----- ---- --- ----
>   f(x, y);        //   G    G     G   G    -
>   f(x, pc);       //   S1   S1    A   A    -
>   f(x, "abc");    //   A    S2    A   A    A
>   f(x, p);        //   S2   S2    A   A    -
>                   //  VC   BCC  MWCW g++ Comeau WS5  WS6u1 g++-3
>                   //  --- ----- ---- --- ------ ---- ----- -----
>   f(x, pc);       //   S1   S1    A   A    S1    S1    S1    S1
>   f(x, "abc");    //   A    S2    A   A    A     A     S1    S1

VC is 6 and correct answers were MWCW's (and g++ old version, go figure!).

And VC7 is not any better than VC6 on those examples. It was better on another
one, but I have lost the posts so I cannot quote.

Quote:
> > are a lot of libraries now that do make use of stuff that is not yet there
> in
> > VC7 (boost, for all).

> But boost does have so many workaounds for VC++. That brings one of my fears
> ... We are all gonna be deprived of good libraries and the ability to write
> code using innovative techniques because of everyone waiting for VC++ to
> catch up ... And if there is another huge gap be{*filter*} VC7 and VC8 then I am
> positively gonna think of using less of C++ ...

Actually part of boost functionality is just not working on VC.

But the point is again: they must develop under Linux (or Windows ports of Linux
compilers) and then add all sort of workaround for VC.

That means: someone is using VC and VC is one platform to be supported (if
possible). But serious stuff is done somewere else.

ARGGHHH. Not surprised that people do not go to MS to ask for ISO. They just
stay away from MS as long as they can! That is something I want to be changed!

Quote:
> > I do. And that scares me. If MS is gonna support the standard (and lately
> it
> > seems like they are trying to support standards) then C++ will live. If MS
> is
> > not ... well, the MP3 recalls the story of Pascal and Borland.

> Thats why I say I might think of giving up on C++ as my main tool and I'm
> sure a lot of others also will if things don't improve. It would be so sad
> for one of the best programming languages. Lets just hope that this thinking
> is not cassandra-ish ...

Yep. I'm talking with a company that has openings for C++ development. And they
do both on-board and ground satelite SW. If I'll go there I'll already have to
cope with so many embedded stuff that unless VC becomes really flexible I'll
have to kiss it goodbye even as a "sketch" tool.

--

Andrea Ferro

---------
Brainbench C++ Master. Scored higher than 97% of previous takers
Scores: Overall 4.46, Conceptual 5.0, Problem-Solving 5.0
More info http://www.*-*-*.com/



Tue, 14 Oct 2003 05:11:35 GMT  
 The future of C++ and VC++ ...
Hi Ronald,

Quote:
> Jonathan Caves, one of the leads of the compiler is the primary Microsoft
> delegate and I am the alternate.

That does indeed ease my mind! :)

--
Tomas Restrepo



Tue, 14 Oct 2003 06:03:15 GMT  
 
 [ 25 post ]  Go to page: [1] [2]

 Relevant Pages 

1. Future of C++ and Windows Programming :: C++

2. FUTURE SOFTWARE FUTURES

3. Hard Questions about Using VC Technologies in the Future

4. Abou C++ future

5. future of c++ ?

6. Future of VC?

7. C/C++ Future - A standard proposal for numbers

8. continue to use ISO C instead of C++, java or future hypes

9. Generics (templates in C++)- in our future?

10. Generics (like C++ templates) in our future?

11. C++ and the future of C

12. Future of Watcom C/C++

 

 
Powered by phpBB® Forum Software