2008/12/31 John McKown <[email protected]>: > On Tue, 30 Dec 2008 16:16:25 -0600, Don Higgins <[email protected]> wrote: > >>Here is something new and something old for 2009. Come to SHARE session >>8194 in Austin TX on March 3, 2009 at 8 AM for the first live demonstration of >>zcobol portable mainframe COBOL compiler which is written in z390 structured >>conditional macro assembler. zcobol supports multiple dialects of mainframe >>COBOL with the option to generate source and executable program code in >>any one of the following target languages: >> >>1. HLASM compatible mainframe assembler for z390/z9/z10 >>2. Java for any J2SE platform > > Does this mean that it might be possible to use zcobol to compile some of > our current z/OS COBOL programs into Java classes and run them on a zAAP?
Heh heh heh... The specialty engine can o' worms arises yet again. I have no doubt that IBM has applied people much smarter than I am and equipped with all the best tools, to this situation, because it has been inevitable that someone would produce a COBOL to Java compiler (whether the compiler output is Java source, or JVM bytecodes) just to exploit this situation. I believe there are a couple of existing commercial offerings along these lines, but they are probably geared to running COBOL on non-mainframe platforms, and don't really provide any threat of moving workload from regular (expensive) z engines to specialty (cheap) ones. But here comes a compiler that is written by mainframe knowledgable people, with a clear understanding of what's needed for such a transition, including the supporting environment items like CICS. My guess is that IBM won't <do> anything at the moment, because zcobol is probably too difficult for production shops to use, and isn't (yet?) supported in the way that big commercial users would require. And of course there's the performance question. But clearly IBM does have a problem if it becomes fairly easy to move existing (dare I say "legacy"?) workload to a cheaper and maybe even faster engine. All of the specialty engines types except zAAP have technological and contractual means of restricting their use to the IBM-desired "new" workloads. But an engine that runs Java by design in order to attract new work (hopefully from Sun or HP boxes) can't really look at your JVM bytecodes and say "this <looks> like something that was originally written in COBOL so we won't run it", any more than an airline can say "you <look> like a business traveller, so we'll charge you five times the rate your seatmates are paying for this flight". My wild guess is that if zcobol looks like a real threat, IBM will just buy it. And maybe that wouldn't be such a bad thing from its developer's point of view... Tony H. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
