Why I'm Interested in H/390

kiyoinc@...

Feb 15, 2000

** Reply to note from hercules-390@onelist.com 14 Feb 2000 19:32:24 -0000

I was there during OCO and it was a big mistake. I think a people's S/390
and especially a people's MVS could save MVS and S/390 from IBM marketing.
You do realize that MVS and S/390 are going down? Who did it? IBM.

Here are a few things I'd do if I had all the source.

1. I'd rewrite 3270 so that the host could sound a quiet audible alarm when
it unlocked the keyboard. 25 years of watching users press enter, only to
hear the keyboard-is-locked sound is long enough. How mornful.

2. I'd add either fine lines or a light shading to 3270s that ISPF or TSO
FSE could modify. This way, I'd *know* where column 10 and 16 are by the
light shading.

3. I'd add a key to tab-to-next-shaded area. What column do you start
your comments in?

Maybe these 3270 things are passe now. Maybe it's all PeeCee editors. Oh
well.

4. I'd define a sysout option called queue. It would work like this, the
JCL sets the Q-size in records. When an abend occurs, only the last
N-records are printed from the queue.

//SYSDMP DD SYSOUT=(Q,500) prints the last 500 lines only.

The implementation, an in-virtual memory circular queue that step-term knows
to trim to the last N-records and pass to JES2.

Only the last N records are ever printed from the Q.

5. I'd define a shared memory area, using the above Q-mechanism, that I
could view from a TSO session. Jobs could write processing status to a
file.

//SYSTRAC DD SYSOUT=(SHARED,24) Limit this to 24 lines?

From a TSO session, I could:

Inspect dmb34a(j00456) ddname(systrac)

and I would see the current 24 lines update as the job writes to the file.
It wouldn't scroll, the 25th line would overwrite the 1st line at the top of
the screen. If the TSO session doesn't request the ddname, the records are
still written but since it's only a few bytes, it's at nominal overhead.

Great for job status and answering the age-old question, what the Hell is
that job doing now?

6. I'd tighten the connection between the PeeCee and MVS, I'm pretty tired
of IND$FILE and FTP, aren't you? Why isn't there a dataset type called
PCAM, PeeCeeAccessMethod. ASCII sequential would be first. I create a
dataset on MVS or the PeeCee. I could only write printable ASCII into it. It
exists on both my PeeCee disk and MVS as a shared resource. I don't move it
or copy it. It exists in both places.

Possible in H/390 because the PeeCee disk is the mainframe disk. Let EMC
salivate.

Maybe someone smart will figure out how to do numeric and binary data.

Other comments.

I understand the appeal of a pure S/360 mode but those days are over. There
will never be a need for S/360's unaligned operand exception again.

I hope, by mentioning the steps for ordering MVS 3.8, I haven't unleashed a
denial of service attack on PID. As of 1995 (or 6) it was still possible to
order MVS 3.8 and 21.8F. I saw it done.

Hopefully someone will take the lead and talk to IBM about ordering one copy
with a few hundred or thousand licenses. All we need is a letter that says,
to Joe, Bob, or whomever, here are a thousand licenses to do with as you
please. Source would be nice. I'd love to have PL/I (Optimiser) too.

Then we can press it onto CD. I don't know about you but I don't have a
3480 tape drive in my living room. Neither do I have a DS3 to download the
DLIBs.

A CD would be very nice.

By the way, you do know that MVS 3.8 lives in the 20th century only. We
have to IPL back in the past. VSAM Catalogs don't work after December
31, 1999. Anyone want to pick a date? How about following the rules of Y2K
and making this year (2000-28) 1972. That way, in a few years when we're
NJE'ing jobs across virtual CTC's (encapsulated in TCP/IP), our systems will
be synchronized.

1972 isn't a bad year to be running MVS on a 1 MIPS S/370 with 16 meg.
National governments barely had that much power. Even with the current
state of H/390, it'd take a Linux on a 4-way Pentium Pro 200 to deliver
that. I make that about $1500 of hardware at the local used iron store.

Take a look at the processing power of the FLEX, the other software S/390,
http://www.funsoft.com. Interesting possibilities?

The P/390, P/390E, P30, and H70 are also interesting.

These are very fun times.

-cory

Cory Hamasaki

3:00 am

Re: Why I'm Interested in Hercules 390

Roger Bowler

Feb 15, 2000

Cory Hamasaki said:
>I think everyone on this list has the technical skills to build a S/390
>product. You're talking about putting up Linux, an experimental
>simulator, and gen'ing MVT and MVS, and mod'ing all the internals?
>C, asm370, ??? Sounds like serious skillsets to me.

What puzzles me is why it should be so darned difficult to find a job
with these skillsets at the moment. Anyone with an interesting and
attractive job offer is welcome to email me privately.

Regards, Roger.

9:44 pm

Date: Wed, 16 Feb 2000 06:17:23 -0500
From: Don Higgins <Don_Higgins@...>
Subject: Re: Why I'm Interested in Hercules 390

All

If the AS/400 is not to your liking, another new beginning for MVS system
programmers is to get into the eCommerce integration of MVS systems with
the Internet by porting some or all of MVS application systems to PC lan
server and client architecture. The Fortune 1000 still has tons of MVS
cobol, assembler, etc. that they cannot just throw away but want to
repackage as eCommerce enabled systems.
The companies in this growing integration market need MVS skills as well as
Unix and PC skills.

Don Higgins

Re: Why I'm Interested in Hercules 390

Roger Bowler

Feb 16, 2000

Don Higgins:
>The companies in this growing integration market need MVS skills
>as well as Unix and PC skills.

They seem to be keeping pretty quiet about it.
Got any names of companies/managers doing the recruiting?

Roger.

11:29 pm