RE: Switching to Java on OS/400 - got a few questions.

> From: StokesP@xxxxxxxxxxxxxxxx
> 
> > And there's simply no way you get the same reliability on a Windows
> > machine (or a Linux machine, or even a Solaris machine).
> 
> I would differ on that point.

Based on what?


> > I see.  How about an APAR number?  Typically when a problem is
reported
> > to IBM, there's an APAR number associated with it.
> 
> "I can't advise on any specifics, Sorry."

Oh stop.  You originally publicly reported this problem back in 2004
running on WebSphere 4 (a particularly horrid release).  This was a
third party application using log4j for something, and you were getting
22 second response time as opposed to just under a second for Tomcat and
WebSphere running on Windows.  Peter Lin suggested a GC problem, and you
never posted again.


> That's fine, no problems there. This doesn't encourage any community
> support either.

Community support for what?  An unknown problem that you can't tell us
about on a third party application you can't name runs poorly on an
iSeries that doesn't exist.

I'm sure you're having a problem, Peter, but without a little more input
we can't identify whether this is just a one-off problem or a real
issue.   For example, what did your third party provider say about the
issue?  Did you run benchmarks to pinpoint the problem?


> I'll sign off now as I don't want to get involved in any petty
platform /
> OS wars - I'll leave that to people who have time to waste.

This isn't petty.  Just trying to pin down facts, is all.  To tell
someone they should test performance is all fine and good; everybody
should do that.  But it would be nice to have more information about the
type of application that gets a 40-to-1 performance boost moving from an
8-way i5 to a Wintel machine.

Joe