Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: nyu notesfiles V1.1 4/1/84; site ur-univax.UUCP Path: utzoo!linus!philabs!cmcl2!seismo!rochester!ur-univax!ning From: ning@ur-univax.UUCP Newsgroups: net.micro.apple Subject: Re: Macprogramming. Message-ID: <5400002@ur-univax.UUCP> Date: Mon, 8-Oct-84 05:20:00 EDT Article-I.D.: ur-univa.5400002 Posted: Mon Oct 8 05:20:00 1984 Date-Received: Tue, 9-Oct-84 19:38:05 EDT References: <131@qumix.UUCP> Organization: University of Rochester: Computing Center Lines: 58 Nf-ID: #R:qumix:-13100:ur-univax:5400002:000:3028 Nf-From: ur-univax!ning Oct 8 05:20:00 1984 > Does anyone have any insights on what's the best program language > software available or soon to be available for the Mac... ... > What decides which is the "best" programmming language for the Mac? > My thoughts- price, availability, no bugs, transportability, ease to > learn it, execution speed, development time, ease of understanding to > other programmers, program structure, I/O capability, and documentation > either built into the SW or in the manuals. Some years ago, just as Pascal became popular, we read and heard much about using the "correct" programming language. Arguments abounded about using the correct level of abstraction in a program versus arguments the speed and efficieny of it's resulting code. You couldn't beat the Fortran jock by writing faster code, but you couldn't beat the his Pascal opponent by implementing more elegant and sophisticated algorithms. Doesn't the same argument hold true when choosing the correct langauge for the Macintosh? Who won? Well, absolutely no one starts a new project using Fortran; however, not many use Pascal, either. It seems that no one won. Many programmers prefer C; critical applications demand assembly langauge; and a good deal of business gets done over APL and some coffee. The point that I am trying to make is that the right langauge should provide the right level abstraction for the task, and more. Sometimes it's nearly impossible to reach a creative frame of mind with a langauge that's obtuse to your thoughts. [See Alan Kay's and Lawrence Tesler's articles (you know who they are) in September's single- topic issue of _Scientific_American_.] With the limited choice of compilers available, I will agree that most of them are barfworthy implementations; they won't survive their infancy. It is tempting to use something that produces beautiful code or gets the quickie program done. But then you will be stuck with an assembly-langauge recipe-finder, or a Basic, smart-terminal program that runs 50-baud serial-ports, or the chore of translating them. All because the right compiler implementation didn't exist at the time. No one could say anything truer of the Macintosh. Software houses have come out with about 20 different C compilers for the IBM PC, and some of them generate straight-jacket-tight code. Why--competition and a growing market of demanding PC developers. (All this nagging is not in vain.) However, let's not limit ourselves to just C compilers. Here at the University of Rochester on a BSD 4.2 Vax exists a Modula-2 compiler (from DEC Western Research Labs) that blows away the Unix C compiler. (And we all know that Modula's syntax is similar to Pascal's.) While I am not trying to be as dogmatic as Niklaus Wirth, the point is that you should choose a langauge that you like and is suitable for the job, and then get the right implementation/s as they become available. Ning____ ..!{allegra|siesmo}!rochester!ur-univax!ning about time someone--anyone--put the U of R on the netland map.