Re: Aint Life Grand

Hello, Frank:

What is old is new again. "In computer science, there are no new ideas."

Many years ago, when I was fresh out of college, one of my first bosses taught me that, in Computer Science, old ideas get recycled, over and over again, in slightly different forms. Apparently, he was correct.

In the late 1970s, I had to support CICS applications on the IBM mainframes. They worked in a manner very similar to your CCP applications that you described on the System/3. CICS transactions consisted of short, run-to-completion programs that performed a single task. To make a more complex application, you would string together several of these (or as many as you needed), where one transaction would end with what CICS calls "RETURN with a TransID" that indicated the transaction ID of the next screen/program to run, in the sequence. This arrangement allowed CICS to support hundreds of terminals and hundreds of on-line users, all simultaneously doing work, running various transactions, on a machine that was only barely capable of executing 1 MIPS (one million 370 instructions per second). This was less powerful than an Intel 386, to give you some idea of how much (or how little) computing power we are talking about here.

IBM's Airline Control Program (ACP) worked in a similar manner, but with even more rigid constraints on the resources that any given transaction could consume, before being summarily terminated.

As it turns out, this paradigm is very appropriate for modern web applications running on a web server, with the client using a web browser ...
Funny how the really "good ideas" just keep coming back around in slightly different forms.

Regards,

Mark S. Waterbury

> Frank.Kolmann@xxxxxxxxxx wrote:

I looked up all my posts on WDSCi to this list and it reads like
a soap opera. Thing is just when I get to grips with WDSCi the
world takes off in another direction. RDi SOA GNW(God Knows What)
and I need to try to get to grips with that.

A long time ago I coded PRUF on a S/3 (no subfiles, program active
for a very short time, do it yourself record locks, keeping track
of what was sent to the display somewhere else, etc etc.) all
this stuff the OS/400 does automatically for you.

It seems to me that all the new fangaled stuff is simply
Deja Vu back to the future just like I did 30 years ago on the S/3.
(no subfiles, program active for a very short time, do it yourself record locks, keeping track of what was sent
to the display somewhere else, etc etc.)
Oh its a bit more sophiscated in that you can have more that one
PRUF (Program Request Under Format) at a time on the display but
essentially the concept is the same.
I hope someone can show me where I am wrong.

Frank Kolmann