On 10 September 2015 at 02:44, Timothy Sipples <[email protected]> wrote:
> One of my pet peeves is that some people -- even a few "experts" -- in the > IT industry persist in using the term "open system" to refer to whatever is > not > a mainframe. I believe words ought to have meaning, including the word "open." I agree. But historically the notion of "open systems" was used in reference to the APIs, and effectively meant UNIX. Open systems in theory allowed programs to be moved from one system to another with merely a recompile to match the hardware architecture. Standards efforts such as POSIX evolved to certify such systems, and of course z/OS is POSIX certified, and therefore "open" in this sense. More recently, "open systems" has been confused with "open source", which z/OS and Windows are manifestly not. Linux is in a funny position, being mostly POSIX compliant but not actually POSIX certified. In my experience, though, Windows was not generally included in what people meant by "open systems"; they meant UNIX, and if they failed to include z/OS (or OS/390) UNIX, it's because they were unaware of its existence. If they wanted to include Windows in a term meaning "not mainframes", they'd say "distributed systems". I hear very few people these days use the term "open systems" at all. But that's about where my agreement stops. > 1. z/OS 2.2 has no activation keys, no Digital Rights Management > "wrappers," or any other such nastiness. Never has. And "never will"? Yeah, I know - not up to you. > 2. Moot. z/OS 2.2 has no activation keys. There are no technical > restrictions on moving z/OS installations from machine to machine. Thank > goodness, since enterprises shouldn't have to worry about such barriers and > complexities in disaster recovery scenarios, for example. Thank goodness. Unless, of course, you want to run z/OS on a zPDT, in which case you need not merely an activation key, but a hardware key/token/dongle. The current version of z/OS (ADCD) available to run on that hardware has been modified to require the matching dongle, so you can't move z/OS from one machine to another unless the dongle follows. Doubtless IBM perceives that there are good reasons for this, but to say there are no activation keys is just not correct. OK - maybe narrowly correct, in that base z/OS doesn't have them. But they are there for some z/OS customers, and I see no reason IBM couldn't implement them for real iron any time it wants. > 3. ... > Importantly, z/OS does not have artificial hardware limitations. Timothy, sometimes you're completely over the top. I understand that sales and marketing people have to be relentlessly positive, but really... Have you tried IPLing z/OS on an IFL lately? Did it run well, taking advantage of that fast CPU, since there are no artificial hardware limitations? How exactly does an IFL differ from a regular CPU? Just in the millicode? Is that difference much more than a single bit? How about running z/OS on a non-IBM, compatible CPU? Oh, wait - there are none anymore? Not because of any artificial limitations, I'm sure. Gimme a break. Tony H. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
