Clem,

I am very curious about this UNIX for OS/400 work it sounds either different or much earlier than what I am familiar with. I am familiar with the PASE environment that shipped around OS/400 V4R4 (1999?). After AS/400s started running PowerPCish CPUs (there is a bit of history there I won't dive into) PASE was like WINE for Linux.. same CPU arch, do some library and linker/loader tricks to hoist a different (AIX 4.3 first) environment within OS/400s understanding of the universe. A year or so later, some very bright group figured to use the OS/400's Single Level Store as the device model/device virtualization for the CPU virtualization (LPAR) in later POWER CPUs. You could run Linux or OS/400 or AIX this way. That work was then somewhat inverted, and pHyp was born from the OS/400's SILC idea of machine dependent code as a light weight firmware hypervisor in the converging iSeries and pSeries POWER systems.. they switched the device model/device virtualization to AIX called APV or PowerVM.

It was nicknamed "Fortress Rochester" for a reason. They did some very nice work. But yeah IBM was running 4 large and extremely different computing businesses in the 1990s and probably some smaller ones too. They were very different but the systems did interoperate pretty well given the stakes.

Regards,

Kevin

On Tue, Nov 5, 2019 at 1:10 PM Clem Cole <clemc@ccc.com> wrote:

On Tue, Nov 5, 2019 at 2:03 PM Christopher Browne <cbbrowne@gmail.com> wrote:
The dates seem to decently explain the invisibility; introduction in 1992
The formal introduction of the PS/2 was April 87 (by the crew of Mash for the TV ads IIRC). Again, if my memory serves LCC started working on UNIX for the 370 in the mid'86s, actually before the PS/2 was announced, which would be announced as AIX/370. ISC had done the original 386 port for Intel, IBM, and AT&T - but that was for an ISA based systems originally [Phil Shevrin pulled one of the best sales jobs I ever knew -- he got paid three times for the same basic work].

At some point (and I would have to ask someone like Bruce Walker or Greg Thiel for the better info), the contract got widen AIX to include the PS/2 - a.k.a. create AIX/386. How much of the original ISC work was that build upon, I never knew.

LCC worked for a number of years and the two AIX's were available for customers, probably under a special University license. The formal introduction was later (and '92 sounds right). But there were sites that had one or both before that time. I want to say, LCC worked with IBM for about 8-10 years starting in the mid-80s. BTW: They also did a UNIX port to AS/400 (on top of the native OS - similar to Eunice for the VAX or today's MingWin and Dave Korn's UWIN stuff). I've forgotten the dates on that, I want to say 93/94 time frame.

Enough time has gone by, I think I can safely tell another story, WRT the AS/400. When that happened, IBM and LCC had a number oif years under their bridge and the LCC management team thought we knew how to work with IBM. Since we had a base IBM contract, we all figured that could be added/amended to as needed. When the folks from Rochester called asking about a quote for the AS/400 work, our sales folks trotted out the existing contract for AIX and figured - ok write a new SOW and we are done. Nope -- different division/different set of lawyers. Something was said to us in the form of 'Rochester Won the Baldridge Award.' I remember our CEO groaning - and saying something like 'Here we go again.' It was then I realized IBM was N different companies, each competing with each other.

Clem
ᐧ