[Zope] Benchmarks vs Jakarta.

[J. Atwood]

> So, my first question is what version of 1.2.2 was used.  Second, was
> there a JIT involved?  If so, which one.

It is the JavaTM 2 SDK, Standard Edition from Sun.

> What webserver was used in either case?

The webserver is Tomcat. Kinda like Zope in that it can either be behind
Apache or stand alone.

> Anyone want to try it on the new 1.3 IBM just dropped a pre-release of?
> :)

Nope. It actually takes a lot to set this thing up. It took me about 8 hours
to get Apache + JServe. I would say Tomcat was just a few.

> One thing that I think may be causing problems for Jakarta is the inital
> load of the JSP.  Zope has a huge advantage in that it cahces most
> objects in memory and the JSP will have to hit the disk on it's first
> run.  Not to mention that it'll compile it into the Java class the first
> go 'round.  I'd be interested in seeing more complex JSPs and servlets
> and more requests to even out the startup time.

This is true. It does not look like Tomcat has any ability to cache things.
It reads through the JSP and executes it. Zope on the other hand can cache
requests. This really comes in handy when you are dealing with RDBMS.

J