RE:SAA Historical perspective

Bob wrote:
>I agree with your comments about there being some good coming out of
>SAA.
>Getting SQL and C on most platforms was goodness. (Although I gotta
>believe virtually no AS/400 programmers care about C, while a growing
>number are at least using SQL to query data, an a minority are using it
>for lots of cool things.

I'm becoming a big fan of SQL myself.  There's a lot of powerful
stuff in there, provided you have a proper understanding of
relational database theory.  (Lately, I've been playing with
PostGreSQL on my home computer, which is arguably the best of the
free open-source RDBMS's available.  Some would argue it's the
*only* free open-source RDBMS available!)

Understanding relational DB theory is important for many reasons.
For instance, some vendors claim to have products that improve on
the "old" relational model.  But for products I've looked into,
I've found the claims to be dubious.  In one case, after I dug
through the hype, the technical details of the product still
looked suspiciously like the relational model.  Even the concept
of "object-oriented database" seems more hype than reality.

One database expert believes that the general understanding of the
relational model amongst programmers is actually poorer today than
when relational was invented 30 years ago.  He argues that back
then, most programmers understood the problems associated with the
existing database models (such as hierarchical and network), and
so most readily understood the benefits of relational.  Today, few
working programmers have any knowledge of the older DB models.

Most AS/400 programmers shouldn't have to worry about C.  C is a
systems programming language.  Most AS/400 programmers are
applications programmers.

>I actually think all languages (except COBOL <vbg> & TIC) would be on
>all platforms. So where's my PL/I? :)

To a large extent today, a large number of programming languages
in common use can run on most hardware platforms.  I would argue
that SAA's dream of hardware independence has largely been
achieved since Linux and C have been ported to most hardware
platforms.

For example, consider my favorite programming language, Python.
Since the interpreter (and many run-time packages) for Python was
implemented in C, and since C runs practically everywhere,
Python programs likewise can run practically anywhere. Someone's
even gone to the trouble of porting Python to the iSeries!

>I also realize that most of the engineering/development staff at the IBM
>labs weren't too interested in SAA, but some where. Weren't the
>mainframers proponents?

As I'm sure you remember, the initial release of SAA looked
sort of like a client-server design where the client was OS/2
and the server was a mainframe OS.  Which is one reason why the
midrangers were up in arms.  Personally, I thought the intention
of SAA was not unreasonable, but yeah, no one here treated it
very seriously.

Just for fun, here are a couple more names pulled out of the old
"SAA Overview" book.  Remember these?  OfficeVision.  AD/Cycle.
CSP.  Easel.  That book offers some fun reading, eh?

Cheers!  Hans

Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com