kent@xanth.UUCP (Kent Paul Dolan) (08/23/87)
One correspondent wrote: Ada is the COBOL of the 70's. This was probably meant as a slam, but really it is a compliment. Aside from the nit that Ada(tm) didn't really exist until January 1983 (the publication date of ANSI/MIL-STD-1815A-1983 American National Standard reference manual for the Ada programming language, affectionately known as the Ada LRM, so Ada is probably the COBOL of the 1980's, instead. Last time anybody bothered to count up, 80% of all operational commercial software, WORLDWIDE, was written in COBOL. (Reference si from a cComputerworld about three years back, and no, I can't find it again, and yes, it probably depends on how you define all those words.) COBOL is over 20 years old, and has survived from an era of machines whose internal arithmetic was really decimal numbers, millisecond gate speeds, and massive 10K's of words drum storage, to an era of parallel processors, 64 bit binary arithmetic, nanosecond and even picosecond gate delays, and casual specifications of gigabytes of secondary storage. In that period, the standard has required just two upgrades. The lady, CAPT Grace Murry Hopper, who gave us COBOL is, despite Countess Ada, probably the most respected female figure in the computing field (can Adele Goldberg be far behind?), which might indicate that the language still retains a bit of respect somewhere. Another correspondent says, roughly, that COBOL is an out of date antique, that nobody is writing COBOL code anymore, only a few (million) drudges are maintaining old COBOL code, that you can't do anything significant in COBOL anyway, et cetera ad nauseum. Amazing what ignorance can accomplish! Since, in most schools, CS and MIS split long ago, since COBOL shops and gee whiz bang programming shops don't usually mix, and since most CS departments leave the teaching of COBOL to the MIS folks, it is (just barely) possible to believe someone being this isolated from reality, but still! My _Mom_ writes COBOL (one of the little old white haired ladies programming for the Post Office department), my brother, as vice president of a programmer "body shop" can place all the COBOL programmers he can get his hands on, the folks I am consulting for, ditto, and even I (blush) have been known to COBOL a bit, when nobody's looking. I've seen parsers, language translators, graphics, even bit twiddling routines written in COBOL. After all, it _is_ a general purpose programming language. I've never seen an operating system written in COBOL, but with its excellent capabilities to define and redefine complex structures, it seems like a natural to do. Ada, if it is very, very lucky, may have as much success as COBOL. This still seems a bit iffy, since some features of Ada seem a poor match for the target problem set, in particular the restrictive rendezvous paradigm of intertask synchronization, the bizzare visibility rules and overly complex data range controls (Couldn't DOD at least have said: "all computer word sizes will be powers of two, and all arithmetic will be two's complement"? After all, they are specifying the hardware, too!), and the lack of programmer control over task spawning, to name a few. It is a shame the development of compilers was so slow that the language had to be frozen before there was any practical experience using it. (Yes, I am aware of all the work with subset compilers before Jan. 1983. I don't count that as practical experience.) In spite of all this grousing, I am an avid Ada supporter; Ada does too many things right to be ignored; I just don't like seeing COBOL take a bum rap. Kent, the man from xanth.