>|Hey Microsoft, if you're listening and you are planning a
>|response to Borland's scathing ad in the latest PC Week, how
>|about this -- "Sure we may be slow, but at least you can
>|build real applications with our product".

> Again, as one Microsoft C++ programmer, I can only say that my personal
> experience comparing the C7 and the Borland 3.0 products completely contradicts
> the Borland advertising claims.  Again, I encourage serious programmers
> to ignore the advertising claims, and compare *product* as actually
> applied to your real, actual, programming tasks.  Competitive update
> policies should make it affordable for people to make these comparisons,
> thus choosing the best compiler for their particular programming projects.
> Again, a particular programming team has more to gain or lose from their
> choice of compiler than either Borland or Microsoft.  Take some time
> to learn the different products, libraries, debuggers, etc, lest you make
> mistakes in your analysis.  A day or two is not sufficient to accurately perform
> these analyses.

   I am making great use of templates in Borland C++ 3.0. Can Microsoft
C++ 7.0 support templates at least as easily as Borland. Borlands add claims
you can't. Our programming group will not be able to use Microsofts
compiler if they can't support templates. Templates are very, very
important in our software design.

--
Joe Fisher
fisherj@ucs.orst.edu
Oregon State University

In article <1992Apr9.035806.3662@talon.ucs.orst.edu> fisherj@ucs.orst.edu (Joseph A. Fisher) writes:
>
>   I am making great use of templates in Borland C++ 3.0. Can Microsoft
>C++ 7.0 support templates at least as easily as Borland. Borlands add claims
>you can't. Our programming group will not be able to use Microsofts
>compiler if they can't support templates. Templates are very, very
>important in our software design.
>

 Microsoft provides a template preprocessor.  To use it, you write your
template definition, and run the preprocessor with some command line switches
to tell it which instantiation you need.  This produces compilable C++ code.
It's up to you to keep track of which instantiations you need.  As I recall,
it supports class templates but not function templates.
 -- Pete

See recent edition of MS Journal for full details.

>>   I am making great use of templates in Borland C++ 3.0. Can Microsoft
>> C++ 7.0 support templates at least as easily as Borland. Borlands add claims
>> you can't. Our programming group will not be able to use Microsofts

Templates are C++ 3.0 spec.  MS C++ is at 2.1 spec.  For templates they
supply a "preprocessor" which you run on your code with templates and then
run the compiler.  MS position is that spec 3.0 (i.e. templates) are
"controversial".

However, Do you use exceptions? Part of 3.0 spec.  Does Borland 3.0 have exceptions.
No....

In article <1992Apr10.162148.29104@borland.com>, pete@borland.com (Pete Becker) writes:
|> In article <1992Apr9.035806.3662@talon.ucs.orst.edu> fisherj@ucs.orst.edu (Joseph A. Fisher) writes:
|> >
|> >   I am making great use of templates in Borland C++ 3.0. Can Microsoft
|> >C++ 7.0 support templates at least as easily as Borland. Borlands add claims
|> >you can't. Our programming group will not be able to use Microsofts
|> >compiler if they can't support templates. Templates are very, very
|> >important in our software design.
|> >

According to the May issue of Computer language, the Microsoft C++ 7.0 supports version 2.1 of
the language.  Borland supports version 3.0 (??).  There aren't templates in 2.1.

AC

In article <HR.92Apr13105516@dirac.i-kinetics.com> hr@i-kinetics.com (Human Resources) writes:
>MS position is that spec 3.0 (i.e. templates) are "controversial".

Hm, I wouldn't have said "controversial" -- I would have said "ill-defined."
 [[but then again I'm not "Microsoft"]]
Witness all the debate on the net where people have "template" code that
works on one vendor's compiler but not the other.  For any vendor, having
"complete" implementations of templates and/or exceptions is going to require
a "complete" spec of what these words actually mean.  Until then, people's
template and/or exceptions code isn't going to successfully port between
compilers nohow.  Which is the whole point of standards, I think.