Re: More from the past.. old operating systems

Yup,

I started on the S/38 about the same time.  It's stabilty was very good as
long as you didn't go more than a couple months between IMPLs.  By the later
releases, the need to IMPL was more due to the inevitable need to regenerate
addresses.  Early versions required setting the dials for this operation.

R03 introduced DSPACTJOB!  Before that you had to use DSPSBS to figure out
what was going on.  I also recall DSPFs had to be recompiled for R03.   The
model 3 could have as little as 512KB,  the 5 maxed out at 2MB,  the 7 at
4MB and the 40 & 700 at 32MB.   I think a minimal R06 install only required
about 100MB of disk.

Of intesest to the MI programmer, CPF programs were observable.  Along with
the availability of OS internals documentation,  you could argue that CPF
was more "open" than OS/400.


Here's some S/38 pictures...
http://www.corestore.org/s38.htm
Mike's IBM page is nice, but I think he needs a S/34.
http://www.corestore.org/ibm.html



Keith



From: "jt" <jt@ee.net>
> Ahhhhhh, SO...!
>
> Them WERE "the good ole days"...
>
> I don't recall if I worked on these models, but I may just have a
> compile-listing from CPF 2.x around here.  When I actually started the
job,
> Jan. '82, I think Liebert was on CPF 3.0, mebbe...?  We had to work the
> programmers 3 shifts, for a time, because 4 meg memory doesn't go as far
as
> You would think...  (Or at least as far as I would-a thought, because I
was
> coming from a 64KB Univac mini, that didn't have virtual memory!)
>
> The OS was ROCK-SOLID, of course, way back in "the good ole days".  (Not
too
> shabby these days, imv...;-)