On Thu, Aug 20, 2009 at 1:18 PM, Edward Jaffe<[email protected]> wrote: > As I understand it, ONE useless IPL-time instruction program checks on an > IFL (and possibly other specialty engines). This conveniently prevents z/OS > from IPLing on a specialty engine. > > I know one ISV that wanted their software to come up stand-alone on an IFL. > They had copied their IPL procedure from MVS code back in the old days > before OCO. They discovered which ONE instruction caused them grief and they > simply removed it. Now, everything works perfectly.
Right, and it's not a big secret (in fact, it's been named on this list fairly recently). The key is that all of these checks mean that a site cannot "accidentally" run production on an IFL. This protects both IBM and the customers (IBM from loss of revenue, customers from getting sued by IBM). Note that a similar argument has been applied to source code for vendor products: I worked for a vendor whose products were all sourced (so no CPUIDs or other keys). The one and only time we found a company using a product on more CPUs than they were licensed for, it was a bluebird for the sales rep. Full price, too. Customer couldn't exactly argue, since the alternative was to get sued for damages in a fairly slam-dunk case. Before you say, "How do you know there weren't dozens of others?" I guess we don't. But (a) real companies don't deliberately steal enterprise software, and (b) we were interactive enough with our customers that we would *probably* have known. At a minimum, the "cost" of given them source code paid off in good will, as they knew what they were dealing with and could identify issues themselves. Of course, this was long ago and far away; not clear how many of today's so-called sysprogs (this august group excluded) would know what to do with assembler source. SAS/C (and perhaps SAS itself) used to use a CPUID that didn't disable the product if it didn't match, but always wrote a banner saying the company name to whom the software was licensed. Thus if you borrowed a copy of SAS/C from Merrill Lynch (as we did, legitimately and with SAS's agreement), we'd run it and it would say "LICENSED TO MERRILL LYNCH". Again, the theory was that in a real company, someone would eventually ask "Why does our compiler claim it's licensed to someone else?" and a manager would take action. Dunno how that worked out, but it made sense to me. ---------------------------------------------------------------------- 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
