Newsgroups: comp.lang.c++,comp.lang.smalltalk,comp.object
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news3.near.net!noc.near.net!paperboy.wellfleet.com!news-feed-1.peachnet.edu!insosf1.infonet.net!solaris.cc.vt.edu!news.mathworks.com!udel!gatech!howland.reston.ans.net!vixen.cso.uiuc.edu!news.uoregon.edu!engineer.mrg.uswest.com!cherokee!da_vinci!lookout!tblanch
From: tblanch@lookout (Todd Blanchard)
Subject: Re: C++ Productivity
Message-ID: <D3L8AC.2uL@da_vinci.ecte.uswc.uswest.com>
Followup-To: comp.lang.c++,comp.lang.smalltalk,comp.object
Sender: news@da_vinci.ecte.uswc.uswest.com (IT Netnews)
Nntp-Posting-Host: lookout
Organization: US WEST Information Technologies
X-Newsreader: TIN [version 1.2 PL2]
References: <1995Jan23.193745.7044@boole.com> <jim.fleming.84.00133AB6@bytes.com> <D3FuGq.Kwv@da_vinci.ecte.uswc.uswest.com> <1995Feb4.210947.718@mole-end.matawan.nj.us>
Date: Mon, 6 Feb 1995 16:54:11 GMT
Lines: 52
Xref: glinda.oz.cs.cmu.edu comp.lang.c++:111229 comp.lang.smalltalk:20427 comp.object:26240

mat@mole-end.matawan.nj.us wrote:
: In article <D3FuGq.Kwv@da_vinci.ecte.uswc.uswest.com>, tblanch@lookout (Todd Blanchard) writes:
: > Robert Martin (rmartin@rcmcon.com) wrote:
: > : tblanch@lookout (Todd Blanchard) writes:
:  
: > : >Here, perhaps is the point.  The C++ language is entirely too large and
: > : >complex.  
:  
: > : Certainly it is large and complex.  But not "too large" nor "too
: > : complex".  If it were, then it would not be selling as well as it is.
: > : You can claim that the purchasers don't really know what they are
: > : buying, but that is a pretty weak argument.
:  
: > I could argue that if Windows wasn't a technically elegant architecture
: > then no-one would buy it either.  ...

: And I could argue that if poodles didn't have wings, they'd make poorer
: pets ... but that's besides the point.

Clearly you lack the ability to generalize from analogy.

: > And I submit that C++ is at least an order of magnitude larger and more
: > complex than C, Pascal, Cobol, FORTRAN, or whatever.

: But unlike Windows, it provides rewards for mastering the complexity.
: Look at _Scientific and Engineering C++_ (Barton and Nackman, AW).
: Their templates capture--and need!--the concepts of higher algebra
: (abelian, semi-abelian, non-abelian groups, moniods, etc.).  They
: allow strict checking of dimensional correctness at (virtually) no
: runtime cost.

Yes, but these are small programs.  C++ is fine for small programs.  Its
not bad for medium programs.  It totally bites for huge programs.  

: There is a revolution brewing in programming, and it's C++'s templates
: that will bring it about.

Oh yes, automated code bloat with extra long compile times just for good
measure.  Using the RogueWave libraries I managed
to produce a library that was 4/5ths template instantiations for a
pointer list class that could have just as easily been written as
void*.  Templates are of less value than it would at first appear.  And
templates are no better than C macros in many cases.

: The question is not `which language will replace C++?' but `what
: preprocessing will people invent to make writing templates less clumsy?'

Actually, the real question is whether direct to SOM compilers will
appear soon enough and provide robust enough capabilities to isolate
dependencies so you can build in a reasonable amount of time.

Todd Blanchard
