RE: Green-screen versus browser -- MIDRANGE-L

> From: Trevor Perry
> 
> This is true from experience in the real world - saying it in CAPS
does
> not
> make it NOT TRUE. What I said was: "generally iSeries customers find
it
> requires heavy iSeries resource usage for a significantly slower user
> response time".

What you said "generally" is absolutely wrong.  I work in the real world
as well, Trevor, and my clients simply do not see the results you see.
Unless you have a poorly written system, JSP is nearly as fast as green
screen.  Sure, you can write bad JSP code.  You can also write bad CGI
code or use a bad tool.  But JSP is not inherently slow because it is
JSP.  Am I clear on this?


> Regarding "heavy iSeries resources", you agree with me when you say
> "require
> about 500MB of dedicated memory". Add to that the CPW requirements.
When
> you
> are being sold WAS from IBM, you are always told that it has a minimim
> requirement of 300CPW or 500CPW - depending on model. For a LOT of
iSeries
> customers, this is a ~significant~ resource requirement and can cost
tens
> of thousands of dollars - along with the additional memory.

I run WAS just fine on the smallest model 270, which has 370CPW.  This
is a very small machine.  What sort of machine are your clients running
on, Trevor?  The beauty of J2EE is the ability to offload it (something
that simply can't be done with RPG-CGI or Net.Data solutions).  Because
of this, I also have clients running J2EE in conjunction with beige
boxes, because they can offload the web serving requirements.


> I said "IBM has offered various graphical user interfaces" and then
"for a
> significantly slower user response time". Even IBM admits that
Webfacing
> will cause the response times to be reduced. Neither have they sold
HATS
> as a way to improve the response times. WAS has historically reduced
> performance so often, that many iSeries customers won't go near any
new
> version.

If you want to complain about Webfacing or HATS, be my guest.  I didn't
bring it up.  But don't try rewriting your words.  You did not tie bad
performance to WebFacing, you specifically stated that WAS was "in its
infancy" and that it was slow.  I disagree completely.

I am talking about simple, basic J2EE architecture, which performs very
well and negates all your arguments against WAS.  I don't even know what
you mean by WAS historically reducing performance.  The faster they make
the JVM, the faster WAS works, and believe me, the JVM is getting better
and better.  Ask anyone at IBM who works with it, like Blair Wyman.

And by the way, WAS is hardly in its infancy.  WAS is in its fourth
generation now: WAS 1/2, WAS 3, WAS 4 and now WAS 5/6.
 

> You may be able to improve response times with your technical hands on
> solution, however, most CxO will not see or hear about your solution
when
> they are talking to IBM.

"Technical hands-on solution"?  What in the world does THAT mean?  Is
this some kind of code word for "programming"?  Is programming now a bad
thing in your world, Trevor?

As to what IBM sells, that's not my point.  I'm simply addressing your
misstatements.  As long as you try to depict WAS and J2EE as slow, I
will continue to say that you're wrong, and provide proof to back it up.

Joe