RE: DB and App Agnosticism on the iSeries (thinkingabout:Questionabout UDB on iSeries)

> From: Hall, Philip
> 
> The Unix port probably helped, but me thinks the CEO - who is still on
the
> run? - and possibly the head of R&D - a nice guy from all accounts had
a
> little more to do with it...I was still there when it all went pear
> shaped.

Which CEO is that?  I hope you're not talking about Covey.  The man was
fanatical about the company and its employees.  Unfortunately, he was
blinded by a couple of executives, one who had great application vision
but poor technical acumen, and one who was simply an idiot.  I'll let
you guess which was which.  Larry Ford on the other hand had no business
running that company.

The demise occurred primarily while Roger was off chasing his passion
for Chinese culture and history.  And while that was a well-earned
pursuit for a man who had built a multi-million dollar company, the
problem was that the company WAS Roger, and without him it devolved into
furious infighting and some really disgusting politics at the end.  Had
Roger been in-house to PROPERLY handle the Unix cutover, we might have
managed it, but without his nearly savant-like capabilities, the project
was a picture perfect lesson of failure that, if not heeded, is doomed
to be repeated.

1. Use LOTS of outsourcing.  Make sure to do it overseas, with lots of
issues of language and time zones.  This is sure to waste lots of money
as well as derail proper development.

2. Hire lots of whiz kids with no application development experience,
and turn them loose without a real architectural design.  These days,
this is called "Extreme Programming", but back then it was still just
"stupid".  We got exactly what we should have gotten: C code that looked
like RPG with bad embedded SQL.  Not to mention the fact that the code
cowboys didn't have a clue what a "Job Description" did.

3. Run the project by committee.  Make sure upper management met twice a
day to argue technical details with senior staff.  This ensures that the
latest buzzwords HAVE to be incorporated in every design, even if they
MAKE NO SENSE.

Anyway, anybody who is looking to port an existing project from the
iSeries to "platform independent" ought to take a LONG look at what
happened at SSA.  The cost of re-architecting native iSeries code into
any of the lesser environments is an awfully big one to recoup.

Joe