Re: How do we fix IBM's "eServer" mentality?

Bob wrote:
>With COMMON approaching in two month (www.common.org) we need to discuss
>the eServer situation and the downfall of the iSeries/AS/400 systems.
>
>Back in the 1980s a grass roots group of us decided IBM's end-all be-all
>"SAA" sucked.
>We helped kill SAA.
>
>In my view, we need to declare war on the "eServer" campaign, or start
>learning UNIX and Windows NT/2000 Server stuff.
>
>I designed a button back in the 1980s with the acronym "SAA" on it, with
>a red circle with a slash through it. "No SAA" it sad iconically.
>...

Bob:  Your memory of events 14 years ago differs somewhat from
my own.  I remember SAA well.  For a while, I was involved with
the effort to put RPG into the SAA arena.  (When I told
management I thought SAA RPG was a bad idea, I got shuffled
into the team charged with developing the SAA RPG compiler.
Another poor stiff was then assigned the SAA duties, and later
left the company.)

I remember the "/SAA" button.  I still have one or two buried
in my stuff somewhere.  Most of us here in IBM thought it was a
rather cute collectible.  But did it help bury SAA?  No, I don't
think so.  Although the idea of complete software portability is
a mother-hood issue among us computer geeks, it can't be achieved
by top-down fiat.  IBM must be a customer driven company, and the
SAA effort just had too little support both within the company
and among the customer base.  In other words, the initiative
was essentially still-born right from the beginning.

Now it's interesting to see what of the initiative has survived
into the present.  Out of all the component in the 3 broad
areas of SAA (user access, programming, and communications),
there are really only three survivors with broad cross-system
support:  C, COBOL, and SQL.  If you want to develop a truly
cross-platform application now, there are ways.  But these
techniques are all based on non-IBM systems and standards, such
as Posix and Java and TCP.

BTW, regarding RPG (this is an RPG mailing list after all), we
all knew full well that an SAA RPG compiler was a bad idea.
But the project eventually led to a number of other new RPG
products, like VA/RPG, CODE program verifier, and ultimately
RPG IV.  If it weren't for SAA, all of these tools may not have
come about.  So I suppose SAA wasn't a complete waste of time.

Also BTW, I don't think learning Unix is necessarily a bad
idea.  There are a lot of good tools in that arena.

Hans

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