From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham autolearn_force=no version=3.4.4 X-Google-Thread: 103376,6609c40f81b32989 X-Google-NewGroupId: yes X-Google-Thread: 1094ba,9bdec20bcc7f3687 X-Google-NewGroupId: yes X-Google-Attributes: gida07f3367d7,gid8d3408f8c3,domainid0,public,usenet X-Google-Language: ENGLISH,ASCII-7-bit Path: g2news2.google.com!news4.google.com!proxad.net!feeder1-2.proxad.net!feeder.erje.net!news.musoftware.de!wum.musoftware.de!news.karotte.org!uucp.gnuu.de!newsfeed.arcor.de!newsspool1.arcor-online.net!news.arcor.de.POSTED!not-for-mail Date: Tue, 06 Apr 2010 23:37:56 +0200 From: Georg Bauhaus User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1.9) Gecko/20100317 Thunderbird/3.0.4 MIME-Version: 1.0 Newsgroups: comp.lang.ada,comp.lang.fortran Subject: Re: Why is Ada considered "too specialized" for scientific use References: <4bb9c72c$0$6990$9b4e6d93@newsspool4.arcor-online.net> <4bbb3f22$0$7660$9b4e6d93@newsspool1.arcor-online.net> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Message-ID: <4bbba9b4$0$6987$9b4e6d93@newsspool4.arcor-online.net> Organization: Arcor NNTP-Posting-Date: 06 Apr 2010 23:37:56 CEST NNTP-Posting-Host: 402a4713.newsspool4.arcor-online.net X-Trace: DXC=:@HNcERDKg1n`gW2MTm]<34IUKejV8LTFfGUoDj\65ehhkjX@Ij4 X-Complaints-To: usenet-abuse@arcor.de Xref: g2news2.google.com comp.lang.ada:10872 comp.lang.fortran:24515 Date: 2010-04-06T23:37:56+02:00 List-Id: On 4/6/10 9:53 PM, J. Clarke wrote: > On 4/6/2010 10:03 AM, Georg Bauhaus wrote: >> So for scientific computing, MS C will be a less attractive choice >> than GNU C or Intel C, or Comeaucomputing's C on top of MS C adding >> C99 to MS C, or ... >> >> Or less attractive than compilers for one of the other >> languages such as Ada or Fortran or ... that support both fairly recent >> standards and computing with complex numbers. > > What is the objection to using the C++ complex library? (Or, in other circumstances, objections to using a library such as Leda maybe.) I'll speculate about two major reasons for not hoping for the C++ complex library to replace Fortran function libraries any time soon. At least in some domains... One reason would be successful tradition: a researcher has successfully written a scientific program using knowledge available with Fortran 77; moving to Fortran 90 has improved the solution. Why switch to non-Fortran? The post-hoc fallacy aside, if non-Fortran is C++, to use C++ effectively it takes learning a language integrating very many parts in far reaching and novel ways (from the researcher's perspective). Most parts need to be well understood in order to bridle the compiler. To him or her, what is the indisputable advantage of C++ in relation to, say, a modern subset of recent Fortran? Maybe the support of physical unit checks at compile time is an example. But the mechanisms behind template specialization based C++ computation are not that easy to grasp, are they? At least hardly easier than just moving to Fortran 95 or later and manually checking units by paying attention. Remembering professor Fitzpatrick's published remark that started this thread, a researcher's job is probably focused on computing scientific results rather than optimizing language use. So Fortran 90 it is, or C---until a new generation of researchers and research problems gives rise to a new tradition of similarly forced attire using another language. Technical arguments involving language properties beyond immediate necessity are subordinate, as ever. After all, we continue to pay them for this style scientific software! ;-) [end of speculation]