home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.lang.c:20084 comp.lang.c++:19776
- Path: sparky!uunet!pipex!warwick!uknet!brunel!cs89rdb
- From: cs89rdb@brunel.ac.uk (Roger D Binns)
- Newsgroups: comp.lang.c,comp.lang.c++
- Subject: Re: C/C++ Correctness (was: Re: C/C++ Speed)
- Message-ID: <C19qw8.1z3@brunel.ac.uk>
- Date: 22 Jan 93 18:48:07 GMT
- References: <TMB.93Jan14143610@arolla.idiap.ch>
- Followup-To: comp.lang.c,comp.lang.c++
- Organization: Brunel University, Uxbridge, UK
- Lines: 28
- X-Newsreader: TIN [version 1.1 PL8]
-
- Thomas M. Breuel (tmb@arolla.idiap.ch) wrote:
- : But, in my experience, you can't do both: once you start taking full
- : advantage of C++'s convenience and safety features (memory management,
- : virtual functions, etc.) you pay a hefty price in terms of
- : performance. In fact, I find that the price I pay for these features
- : in C++ is often larger than in other languages.
-
- You will only take full advantage of features if they give you some gain.
- Usually, it is less debugging time, and quicker development. The question
- is what would you rather have:
-
- - A product early to market that is correct, and runs at a performance
- 'penalty'
- - A product early to market that is incorrect, and has no penalty
- - A product late to market that is correct and has no performance penalty
-
- I suspect the case is the first one. The second one seems to be common
- practice. The last one will give you the warm fuzzy feeling you desire
- when having no performance penalty.
-
- Roger
- --
- +=============================================================================+
- | cs89rdb@brunel.ac.uk Roger Binns Brunel University - UK |
- |:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::|
- | Do your bit to destroy the American economy by overwhelming their legal |
- | system - buy an Intel chip to run your Lotus software TODAY. |
- +=============================================================================+
-