Re: Saving the System i: Fight Rather Than Switch -- MIDRANGE-L

But in our
case, IBM refuses to consistently promote the "i" brand.  There's an
occasional, short lived ad campaign, but nothing long term.  Also, as
I've mentioned, the defection rate seems to be a lot higher than the
adoption rate.  This is not for technical reasons, as we all know.


I agree. You nailed it. In the early 80's for about an 8 year stretch
was when the beginning of the end occurred - We (a company) were 100%
true blue. While migrating from 100+  36's to AS/400's - the remaining
projects were starting to get the "is it possible to deploy on the
LAN" treatment - nowhere to go but defect or stay blue. OS/2 was not
exactly an attractive choice, but played a part in the belief system
as well as project cost. In 8 years I rarely saw good marketing to
hold my own mindshare or went to any free IBM/midrange training, the
local users group was always small and we paid for any info via
subscriptions while scarfing redbooks on CD's. The AS400 "welcome"
CD's were pathetic.

However, I went to promotional Microsoft seminars and downloaded tool
after tool - server after server, for free training. It became
impossible for any one human to keep up with what was good for any
single organization -at the detail level. Between the army at
Microsoft, freeware and "C-sources" scattered about the net, we all
knew the "mix" was here to stay.  Not to mention the reams of info
available online at MSDN & technet. MS has always whipped IBM redbooks
and the infocenter. From the consumer perspective, accessible online
information appears that the effort/expense ratio for documentation
and training was MS=100 IBM=1 in manpower to make information
available to customers. But the redbooks are an A+ effort for IBM.

Final note: When I loaded up the original version(s) of websphere on a
170 and saw the start up times and effort, my mindshare was history.
Most clients don't have strong systems or money for that garbage.
Linux and Apache was the cheapest safest choice, or MS if you like.
Websphere was a COLOSSAL strategic mistake any company that had a lick
of sense that did not need Websphere for the core mission - did not.
There strategy to mix other products just added to injury. Product
experts were confused about prerequisites and paths to the future.
Never had that problem with MS did we? We do not care if the JVM
needed x86 instructions or the 360 instruction set - it was slow and
IBM screwed up by making it THE choice! We already had extremely well
performing rock solid mission critical websites with a great socket
listener/controller. IBM should have snuggled up with Sun closely at
this point, payed them royalty loot or SOMETHING after the lab test to
get the JVM optimized for each server brand -OR- re-bet the farm on
another strategy.

I hope they read this thread.