From psmith@mozart.convex.com Wed Mar 18 18:43:21 1992 Received: from convex.convex.com by dkuug.dk via EUnet with SMTP (5.64+/8+bit/IDA-1.2.8) id AA15258; Wed, 18 Mar 92 18:43:21 +0100 Received: from mozart.convex.com by convex.convex.com (5.64/1.35) id AA01223; Wed, 18 Mar 92 11:42:57 -0600 Received: by mozart.convex.com (5.64/1.28) id AA29269; Wed, 18 Mar 92 11:42:56 -0600 Date: Wed, 18 Mar 92 11:42:56 -0600 From: psmith@mozart.convex.com (Presley Smith) Message-Id: <9203181742.AA29269@mozart.convex.com> To: meissner@lynx.cs.usfca.edu, sc22wg5@dkuug.dk Subject: Re: (SC22WG5.67) Another tangent Cc: psmith@mozart.convex.com X-Charset: ASCII X-Char-Esc: 29 > Just to toss another irrelevant fly into the ointment: > what about "the Fortran user community"? > Clearly the F.U.C. has changed A LOT since "77". Then (as since 1960, say) > F was the predominant general purpose language. Clearly Pascal has taken over >one big segment of that market, and C has taken over another. Maybe some >others like ADA have picked up a share. I include C++ with C. And I have >been hearing about the "revival" of Basic. >What is left to Fortran in this decade is the BIG scientific market. > In that light, the (pre)"dominance" of the Energy community does not seem >so strange. > Do we count "the majority of Fortran users" or do we count machine cycles or >MIPS-hours (whatever the units should be)? Let every Fortran machine-second >have a vote, and Energy's representation on X3J3 might not seem so biased. >At the other end, if we count "users", is every student in my class a "user" >--- even one who writes a total of 6 - 100 line nearly trivial programs in >the semester? Should we have one X3J3 member for every N Fortran programming >students? > Something to think about: if X3J3 should be "representative", what should it >represent? Some people want an X3J3 member for every billion dollars of >"investment" in Fortran (dusty decks? new codes since NAGWARE F90?) >??? >--Loren Great question Loren... This is the kind of thing we need to resolve. What is the target audience for a new Fortran standard. And, how many users or potential users are there in that target audience. And what are their requirements vs wishes. If a business is to be successful, you've got to know what your customer requires vs what he might just like to have. If your business trys to be all things to all people, then it will fail. Same thing is true with standards. If we want Fortran to overtake COBOL, then we'd better add ISAM file constructs... Obviously, that's silly so we will not do that. We really should try to get a picture of what the Fortran community looks like today and what we project it to look like in 10 years and then decide what things need to be done based on that knowledge. --------------------------------- I'll start the process. My Fortran base is made up of customers who's prime concern is PERFORMANCE! The code has got to run very fast on the machines they buy. We live or die on performance. Customers many times are willing to change their code or even rewrite parts of their code to made it run fast. They are also interested in libraries and standard 3rd party products that are written in such a way that they run fast. ------------------------------ If we make modifications to Fortran that slow performance down, then we'll just send those customers to other langauges... How big a group is this? It's public information that we have about 1000 computers with 500 customers. If we take 50 programmers per customer, which is probably low, then that equals about 25,000 Fortran users who's main concern is performance. If we don't want to be the language for performance computing, then we can write off those 25,000 users. Loren, are you going to be the collector of information in the survey you started??? FYI. Presley