Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!apple!agate!pasteur!gaia.berkeley.edu!kemnitz From: kemnitz@gaia.berkeley.edu (Greg Kemnitz) Newsgroups: comp.edu Subject: Re: Language Use Keywords: C,Ada,Pascal,C++,Other? Message-ID: <12219@pasteur.Berkeley.EDU> Date: 21 Mar 91 21:01:57 GMT References: <1150@ra.MsState.Edu> Sender: news@pasteur.Berkeley.EDU Reply-To: kemnitz@gaia.berkeley.edu (Greg Kemnitz) Organization: University of California, Berkeley Lines: 37 My comments about industry are largely restricted to the US - other countries certainly can/do have different programming language usage patterns in their software industries. In particular, I would doubt that many other countries are hindered by the huge "installed base" of antiquarian languages as the US; a thing about being relative latecomers to the computer scene will be that many companies based in other countries will avoid the FORTRAN/COBOL black hole that many American companies have fallen into. Also, I intended to describe what I feel is actually happening in industry, rather than trying to say what can happen or should be happening - this is what academics are for :-) The thing that tends to determine which language will be used in industry would surprise many people - it has very little to do with the intrinsic quality of the language itself in most cases. It is whether people are available who know and are familiar with the language. In many companies the question on new projects is whether to use C or C++. No other languages are considered, primarily because training expenses are large and most good programmers in the US know C. Other factors include the cost and quality of compilers and debugging tools on the target platform for these new languages, etc. Also, software has a way of taking on a life of its own (at a couple of points in my career I have encountered FORTRAN II subroutines - that were in use - that are older than I am), so justifying the use of an unusual or new programming language for a product that is expected to have a long lifetime is difficult. Another factor is the ratio of completely new software ventures to improvements on existing software is rather low. One rarely has the chance to start completely from scratch - in many cases, rewriting is simply not feasable, even if the base is an ancient FORTRAN program using arithmatic gotos. ----------------------------------------------------------------------- Greg Kemnitz | "I ran out of the room - I Postgres Chief Programmer | didn't want to be killed by a pile 278 Cory Hall, UCB | of VMS manuals" :-) (415) 642-7520 | kemnitz@postgres.berkeley.edu | --A friend at DEC Palo Alto in the Quake