"DINKUMWARE LTD is the last remaining commercial supplier of Standard C++ libraries" - Dinkumware's website

[Piotr Dobrogost]

Hi

I just read this

"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"

on http://www.dinkumware.com/competitors.aspx

{ mod note: the quoted paragraph continues "and has long been the only
commercial supplier of Standard C libraries." -mod }

If that was due to the fact there are free comparable implementations
of the Standard C++ library I could understand this. But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).

Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?

What does it tell about the Standard?
What does it tell about the C++ in general?


Piotr Dobrogost

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[j***@gmail.com]

Post by Piotr Dobrogost
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"
onhttp://www.dinkumware.com/competitors.aspx
{ mod note: the quoted paragraph continues "and has long been the only
commercial supplier of Standard C libraries." -mod }
If that was due to the fact there are free comparable implementations
of the Standard C++ library I could understand this. But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?

No. To get a little off topic, what's scary is when a company like
Oracle comes in and buys your programming language (Java), putting the
entire future of that programming language into limbo. What's not
scary are the several free open source implementations of the C++
standard.

Post by Piotr Dobrogost
What does it tell about the Standard?
What does it tell about the C++ in general?

I don't know. Tell me.


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[p***@1.google.dobrogost.pl]

Post by j***@gmail.com
No. To get a little off topic, what's scary is when a company like
Oracle comes in and buys your programming language (Java), putting the

Nothing changes, Java was proprietary and still is :)
If someone could ever trust one company then now he is forced to trust
another one.
I think users are going to have as much influence on current owner as
they had on the previous one...

Post by j***@gmail.com
What's not scary are the several free open source implementations of the C++
standard.

Yes, that's a very good thing and I'm aware of it.
I'm only surprised as to what degree they are non conforming and
incomplete in regard to TR1 as shown by Dinkumaware's tests.

Post by j***@gmail.com

Post by Piotr Dobrogost
What does it tell about the Standard?
What does it tell about the C++ in general?

I don't know. Tell me.

Some parts of the stdlib are very seldom used?
If so, why have they been put there?
It's really very hard to implement some parts of the stdlib?
If such project as gcc hasn't implemented them who ever is going to
create them (besides Dinkumware :) ?


Piotr Dobrogost

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Mathias Gaunard]

Post by Piotr Dobrogost
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

Are you aware for example, that Microsoft implementations of C# and
all their related .NET technologies, are not conforming to the
specifications they emit themselves?

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[p***@1.google.dobrogost.pl]

Post by Mathias Gaunard
Are you aware for example, that Microsoft implementations of C# and
all their related .NET technologies, are not conforming to the
specifications they emit themselves?

No, I'm not. I knew they didn't care about standards (C++ being the
best example I know) but this is new to me.
To what extend? Do you have some more information on it?


Piotr Dobrogost

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Mathias Gaunard]

Post by p***@1.google.dobrogost.pl

Post by Mathias Gaunard
Are you aware for example, that Microsoft implementations of C# and
all their related .NET technologies, are not conforming to the
specifications they emit themselves?

No, I'm not. I knew they didn't care about standards (C++ being the
best example I know) but this is new to me.
To what extend? Do you have some more information on it?

Actually, I read that on wikipedia a long time ago, but there is no
citation.
I guess that's not necessarily true then.

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Adam Badura]

Post by Mathias Gaunard

Post by p***@1.google.dobrogost.pl

Post by Mathias Gaunard
Are you aware for example, that Microsoft implementations of C# and
all their related .NET technologies, are not conforming to the
specifications they emit themselves?

No, I'm not. I knew they didn't care about standards (C++ being the
best example I know) but this is new to me.
To what extend? Do you have some more information on it?

Actually, I read that on wikipedia a long time ago, but there is no
citation.
I guess that's not necessarily true then.

I confirm that there is such problem.

Some time ago I was investigating collections sublibrary. I needed
that because I was going to implement a new collection and I was not
entirely sure what the specifications meant so I checked the
implementation. But then I skipped that project and never returned to
that task.

What I found out is that implementation of collections did not throw
on errors as the interfaces declared. However I do not recall more
accurately what was the exact problem.

You can however find it out yourself. All you have to do is to
download .NET Reflector, open MS assemblies and look for
implementation inaccuracies.

Adam Badura

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Pavel Minaev]

Post by Mathias Gaunard

Post by p***@1.google.dobrogost.pl

Post by Mathias Gaunard
Are you aware for example, that Microsoft implementations of C# and
all their related .NET technologies, are not conforming to the
specifications they emit themselves?

No, I'm not. I knew they didn't care about standards (C++ being the
best example I know) but this is new to me.
To what extend? Do you have some more information on it?

Actually, I read that on wikipedia a long time ago, but there is no
citation.
I guess that's not necessarily true then.

It's true that MS implementations do allow some things that the C# and
CLI specs do not (such as e.g. unboxing an int to an uint, or
returning a safe pointer so long as it's to a field and not a local),
but I'm not aware of anything that the spec requires, but the compiler/
runtime doesn't implement.


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Marsh Ray]

Post by Piotr Dobrogost
Hi
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"
[...]
But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?

Very little. TR1 is not part of any current approved Standard.

Post by Piotr Dobrogost
What does it tell about the C++ in general?

To me it says:

1. Dinkumware has an implementation of TR1. Dinkumware has a test
suite for TR1 implementations. Dinkumware is good at passing their own
tests.

2. The general term "C++" implies a large set of both formally and
informally specified libraries.

3. A conforming C++ implementation, even just a subset of the
libraries, is a lot of hard work.

4. A viable, yet mature, ecosystem exists for C++ standard libraries
with multiple implementations. Boost and Gnu have free but imperfect
implementations of TR1.

5. The ISO committe alternately lags and leads actual implementations
depending on which part of the release cycle we're in.

6. We developers still need to use some discretion and good judgement
before writing code dependent on specific libraries, even those
written up by the ISO committee.

7. Porting C++ code between platforms and compilers may continue to
occasionally require a bit more work than just changing the "CXX="
line in the makefile.

- Marsh



--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Jeff Schwab]

Post by Piotr Dobrogost
Hi
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"
on http://www.dinkumware.com/competitors.aspx
{ mod note: the quoted paragraph continues "and has long been the only
commercial supplier of Standard C libraries." -mod }
If that was due to the fact there are free comparable implementations
of the Standard C++ library I could understand this. But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

It tells us 100% compliance has not proven necessary. That's partially
because C++ is a bitch to implement, and partially because the vast
majority of C++ developers use only a subset of the language, anyway.

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Jerry Coffin]

In article <7e903ab6-67d0-4723-a7ce-78b47d951a12
@r37g2000yqn.googlegroups.com>, ***@1.google.dobrogost.pl says...

[ ... ]

Post by Piotr Dobrogost
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

I would tend to look at things from the other direction for a moment:
how many language standards have even one conforming implementation? The
only ones of which I'm aware are those that define the compiler itself
as the standard, so perfect conformance is inevitable (albeit, it's
usually perfect conformance with a standard that's known to be itself be
rather buggy).

In all honesty, I should also (no insult to P.J. intended) admit that I
have some doubt about the Dinkumware library really having perfect
compliance either -- I can't say I've personally seen/found a lot of
bugs in their recent versions, but it's enough code that I have a hard
time believing it's all really perfect (even in just one respect).

--
Later,
Jerry.

The universe is a figment of its own imagination.

[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[Joe Smith]

Post by Jerry Coffin
In article <7e903ab6-67d0-4723-a7ce-78b47d951a12
@r37g2000yqn.googlegroups.com>, ***@1.google.dobrogost.pl says...
[ ... ]

Post by Piotr Dobrogost
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

how many language standards have even one conforming implementation? The
only ones of which I'm aware are those that define the compiler itself
as the standard, so perfect conformance is inevitable (albeit, it's
usually perfect conformance with a standard that's known to be itself be
rather buggy).
In all honesty, I should also (no insult to P.J. intended) admit that I
have some doubt about the Dinkumware library really having perfect
compliance either -- I can't say I've personally seen/found a lot of
bugs in their recent versions, but it's enough code that I have a hard
time believing it's all really perfect (even in just one respect).

Perfect, no. Dinkumware knows that they could not possibly be perfect,
especially where the standard is self-contradictory. Instead they try for
the best possible scores on various testsuites, resolving any known issues
in the Stasndard by guessing, at least untill the standards comitee has
ruled on a defect report. Then they update the library to use the proposed
solution if possible, or if not possible (uses new language features), then
I'm guessing they try to emulate it as closely as possible. Since the
standard does not change very frequently, I tend to suspect that in general
the number of bugs in Dinkumware's library largely decreases over time. I'd
also be willing to bet that they have a frightening number of unit tests to
help avoid regessions.

AIUI EDG does pretty much the same thing with their compiler frontend. The
result is that using Comeau C/C++ (based on EDG), along with Dinkumware
libraries probabbly comes as close to the standard as one can currently get.

I must admit I'm not really clear on the releationship between Comeau C/C++
and EDG. I know Comeau uses the EDG front-end. I know they output low level
C. I know EDG comes with a sample backend that does that. Now what does that
leave Comeau C/C++ doing, besides providing support? I'm guessing Comeau
wrote a middle-end (optimizer), and probably have a heavilly modified
backend. Since the only C++ feature that does not map well at all to C code
is exceptions, they presumably cause a bit on an issue with using C as a
high level assembly as well. I would therefore not be at all surprised to
find that Comeau C/C++'s backend tries to utilize any extentions the
targetted C compiler provides to improve performance of exceptions. As to
the extent that Comeau C/C++'s fornt end differs from EDG's, I have
absolutely no idea.

Hmm. It seems my message has relatively little fact, and much conjecture. Oh
well, hope somebody finds this worth reading.


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

[John Nagle]

Post by Jerry Coffin
In article <7e903ab6-67d0-4723-a7ce-78b47d951a12
@r37g2000yqn.googlegroups.com>, ***@1.google.dobrogost.pl says...
[ ... ]

Post by Piotr Dobrogost
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

It demonstrates that the C++ standards committee isn't doing their job.
Too much effort is going into template la-la land, and not enough into
cleaning up the base language and making it safer.

Hence the industry migration to Java, C#, and even Objective-C, which
has made a comeback as the native language for the iPhone.

John Nagle
Animats

[Hakusa]

Post by John Nagle

Post by Jerry Coffin
In article <7e903ab6-67d0-4723-a7ce-78b47d951a12
@r37g2000yqn.googlegroups.com>, ***@1.google.dobrogost.pl says...
[ ... ]

Post by Piotr Dobrogost
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

It demonstrates that the C++ standards committee isn't doing their job.
Too much effort is going into template la-la land, and not enough into
cleaning up the base language and making it safer.

Is that their job? I don't know too much about it, but if I had to
guess, I'd say their job is to standardize across platforms and
compilers an interface that abstracts OS- and hardware-specific
conventions, while making the language adapt to new technologies and
idioms. Maybe even standardize incredibly common tasks, thus making
programming in the language easier and more hazard-free.

By my definition, I'd say they're doing a great job. By my definition,
the reason the standard library is so hard to implement is because
it's designed to cover up many of the common pitfalls of the language
that are the source of oh so many bugs. This is why std::vector is
such an important class: It helps remove common errors in coding
dealing with memory acquisition such as acquiring too little and
overflowing. And now that templates are becoming so widely used and
lauded, they're entering template la-la land so you don't have to. Tag
dispatching, if I'm not mistaken, is being replaced by concepts, a
much easier way to implement the same thing. Concepts also make the
language safer and when you can't call a function with type T, you'll
know why.

Were the standard library easy to implement, wouldn't it be kind of
pointless? If just any coder could do it, it wouldn't be helpful.

[Jerry Coffin]

In article <49f72d9b$0$1643$***@news.sonic.net>, ***@animats.com
says...

[ ... ]

Post by John Nagle
It demonstrates that the C++ standards committee isn't doing their job.
Too much effort is going into template la-la land, and not enough into
cleaning up the base language and making it safer.

The committee tends, primarily, to respond to input from the community.
They tend to respond best (most quickly and positively) to real
proposals that provide real input. It's easy to sit back and say things
like "they're not doing their job", but much more difficult to come up
with a real proposal that says something like: "Section x.y/z of the
standard is ambiguous, and should be reworded to read: '...'. This would
make the language cleaner and safer in the following ways:".

Everybody I've ever talked to on the C++ committee has struck me as a
committed, hard-working person. Some are certainly interested in
templates -- but quite a few others are interested in other aspects of
the language and its use. I've yet to see or hear any of them reject a
serious suggestion simply because it wasn't related to templates (or
whatever area happened to interest them).

Post by John Nagle
Hence the industry migration to Java, C#, and even Objective-C, which
has made a comeback as the native language for the iPhone.

Yes, I'm sure sponsorship by Oracle/Sun, Microsoft and Apple can't
possibly have any effect on anything. Of course, all of these companies
are also entirely altruistic, so none of them would ever have any profit
motive behind sponsoring a specific language either!

Don't get me wrong: I think all of these languages have some points in
their favor -- but in each case, one of the major points in its favor is
a huge amount of development being put into it by a few large companies
rather than anything really inherent to the language itself.

[David Abrahams]

Post by Piotr Dobrogost
Hi
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"
on http://www.dinkumware.com/competitors.aspx
{ mod note: the quoted paragraph continues "and has long been the only
commercial supplier of Standard C libraries." -mod }
If that was due to the fact there are free comparable implementations
of the Standard C++ library I could understand this. But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).

Just to be very clear, Boost isn't an implementation of standard C++,
and doesn't claim to be. I think the documentation of Boost.TR1 could
be a bit clearer about its aims w.r.t. conformance, but it is very
upfront about existence of known non-conformances.

As for GCC, it would be interesting to see how Dinkumware fares when run
against the test suite used to validate the GCC standard library ;-)

[Francis Glassborow]

Post by David Abrahams

Post by Piotr Dobrogost
Hi
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"
on http://www.dinkumware.com/competitors.aspx
{ mod note: the quoted paragraph continues "and has long been the only
commercial supplier of Standard C libraries." -mod }
If that was due to the fact there are free comparable implementations
of the Standard C++ library I could understand this. But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).

IMNSHO this kind of percentage claim is an abuse of percentages (an all
too common amongst marketing people)

Note that the above figures were arrived at by estimating that the
special math functions and the C99 library were equivalent to 75% of the
whole. Not surprisingly, Boost implements neither of those.

I know that implementing the special math functions is very difficult
(well if you want usable results) and it is greatly to the credit of
Dinkumware that they made the effort to do so. However, realistically
the user community that needs them (and those that do really need them)
is fairly small which is why these have now been parcelled off into
their own free-standing standard.

Post by David Abrahams
Just to be very clear, Boost isn't an implementation of standard C++,
and doesn't claim to be. I think the documentation of Boost.TR1 could
be a bit clearer about its aims w.r.t. conformance, but it is very
upfront about existence of known non-conformances.

And I suspect that taking the Boost code and modifying it to meet the
specs of TR1 would not be that difficult.
No

Post by David Abrahams
As for GCC, it would be interesting to see how Dinkumware fares when run
against the test suite used to validate the GCC standard library ;-)

[Mathias Gaunard]

On 1 mai, 20:47, Francis Glassborow

Post by Francis Glassborow
Note that the above figures were arrived at by estimating that the
special math functions and the C99 library were equivalent to 75% of the
whole. Not surprisingly, Boost implements neither of those.

Actually, it does, albeit incompletely.
It didn't though when the testsuite was run by dinkumware. Their
results are not up to date.

[j***@googlemail.com]

As the author of the Boost TR1 wrappers let me just jump in a bit
here:

* As noted previously, Boost is not a standadrd library
implementation, that would be kind of pointless anyway, rather it aims
to promote standard library evolution and the *next* version of the
standard library amoungst other things.
* The Dinkumware page is hopelessly out of date - in fact it appears
to predate the introduction of the Boost.TR1 wrappers altogether.
* We now have the TR1 unordered containers and the special math stuff
in Boost.TR1 as well (even if the latter is a little incomplete still
- some parts can only really be implemented by compiler vendors using
assembly - and that's not what a portable library like Boost is all
about).
* Since I also happen to be the author of Boost.Regex, and for that
matter the Regex chapter of the TR1, I'm a little bemused by their
claim that Boost.Regex is so non conforming. Heck, their regex
library doesn't pass *my* tests for that matter! Since my tests are
open source they can easily fix that or at least check that claim for
themselves, but since the tests that Boost apparently fails are closed
source there is no way I can check their claim. Maybe it says
something about standard conformance, maybe it doesn't. What I will
say is this: when I wrote the Boost TR1 tests there was only one
implementation to test against - ours. Once other implementations
became available they showed up some bugs in the our tests (mistaken
assumptions on my part), some lack of clarity in the text of the
standard, and yes, quite a few non-conformities in other
implementations, including the one that Dinkumware supplied to
Microsoft. But you know what: this is normal folks!!! No matter how
carefully you write your code and your tests, some differences between
implementations always show up, and yes, even between two very high
quality (strictly IMO of course) implementations. Heavens, if it was
easy we'd all be doing it.
* Finally, it's not my job to defend GCC, but I do happen to know that
their TR1 implementation is more or less mothballed now, so that they
can concentrate on the next C++ standard: it looks to me as an outside
observer as if they're aiming for a pretty complete job at that.

Anyhow, I hope this sheds some light on the Boost side of things.

John Maddock.

[Tony]

Post by Piotr Dobrogost
Hi
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"
on http://www.dinkumware.com/competitors.aspx
{ mod note: the quoted paragraph continues "and has long been the only
commercial supplier of Standard C libraries." -mod }
If that was due to the fact there are free comparable implementations
of the Standard C++ library I could understand this. But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

That's it's proprietary via it's complexity? That it's easier and better to
develop a new language rather than deal with C++'s many shortcomings? That
it's for rocket scientists only and there are relatively few rocket
scientists or wannabes?

[c***@mailvault.com]

Post by Tony

Post by Piotr Dobrogost
Hi
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of Standard
C++ libraries, (...)"
onhttp://www.dinkumware.com/competitors.aspx
{ mod note: the quoted paragraph continues "and has long been the only
commercial supplier of Standard C libraries." -mod }
If that was due to the fact there are free comparable implementations
of the Standard C++ library I could understand this. But taking TR1 as
a test case Dinkumware estimates conformance of free implementations
as 15% for Boost, and 12% for Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

That's it's proprietary via it's complexity? That it's easier and better to
develop a new language rather than deal with C++'s many shortcomings?

Many have tried to knock off C++ over the years -- Sun/Java,
Microsoft/C#, Digital Mars/D to name a few. It's not like
thousands of people haven't tried to replace C++ for decades.
If what you say were true, wouldn't there be an heir apparent
by now?

Post by Tony
That
it's for rocket scientists only

It's useful to a variety of scientists and to other
professions as well.


Brian Wood
Ebenezer Enterprises
www.webEbenezer.net

[Tony]

Post by c***@mailvault.com

Post by Tony

Post by Piotr Dobrogost
Hi
I just read this
"DINKUMWARE LTD is the last remaining commercial supplier of
Standard C++ libraries, (...)"
onhttp://www.dinkumware.com/competitors.aspx
{ mod note: the quoted paragraph continues "and has long been the
only commercial supplier of Standard C libraries." -mod }
If that was due to the fact there are free comparable
implementations of the Standard C++ library I could understand
this. But taking TR1 as a test case Dinkumware estimates
conformance of free implementations as 15% for Boost, and 12% for
Gcc in a comparison to their 100%
(http://www.dinkumware.com/tr1_compare.aspx).
Isn't it scary there is only _one_ company in the world producing
conforming implementation of the Standard C++ library?
What does it tell about the Standard?
What does it tell about the C++ in general?

That's it's proprietary via it's complexity? That it's easier and
better to develop a new language rather than deal with C++'s many
shortcomings?

Many have tried to knock off C++ over the years -- Sun/Java,
Microsoft/C#, Digital Mars/D to name a few. It's not like
thousands of people haven't tried to replace C++ for decades.
If what you say were true, wouldn't there be an heir apparent
by now?

The goal is not "to knock off C++" but rather to realize something closer to
an ideal language. I use C++ also, but a highly stylized form of it that
certainly is not of the prescribed variety and I'm sure many many others use
their own particular dialect of the language also. All those dialects are or
point toward something other than C++. I could counter your question easily
with something like: "If C++ is THE GP language, why are there still so many
other languages and why are there so many more under development? Why has it
not conquered all genres of programming?".

Post by c***@mailvault.com

Post by Tony
That
it's for rocket scientists only

It's useful to a variety of scientists and to other
professions as well.

Look up the metaphor (or is it an analogy?) "rocket scientist".