[Zope] Advocacy FAQ?

Bill Anderson bill@libc.org
Tue, 23 May 2000 19:39:27 -0600


srl wrote:
> 
> So, I was hanging out on IRC last night, and mentioned Zope to
> someone who's fairly active in the Python community. I ended up
> receiving a rant about how "bloated" and slow Zope is, and anecdotal
> evidence about Zope crashing Linux servers repeatedly.

Those conviced by 'anecdotal' evidence are rarely worth the effort,
IME.
Perhasp a rant about how bloated and slow python in compared to C...
nah not worth the ffort either. Basically, slow is a relative term. if
you considered it worht your effort, I would ask him for specific
details. Why _exactly_ is Zope 'bloated', slow in _what_, and ocmpared
to _what_? When pressed for these, this type person tend to just shut
down, and begin waving their hands.

> 
> I know there was/is a memory leak that might be to blame in this

There was more talk about one due to many people misunderstanding the
reporting tools of their OS. That isn't to say there are/were none,
but that most of them turned out to be misunderstandings.

> case, but the fact remains: at least one person who might have been
> convinced that Zope is cool has been *seriously* turned off to it,
> for reasons of speed and memory hogging in a particular environment.

No, for _perceived_ slowness, and _alleged_ memory problems. On top of
that, since we cannot possibly please everyone, this ounds like
something of little consequences. Sure, Zope can be 'sped up', but
then, so can python. Zope is written in python with a few things
written in C.

All things are slow, and all things are fast, it all depends on your
point of view. Slow compared to what? Consider whether the comparison
is valid. For example, serving static, unchanging html files. Zope
loses there, but it was not designed to do that. Zope was designed as
a coherent environment, and as such will incure some overhead.

In addition to that, you would need to take into account the
environmentof the comparison. Want to compare Zope's speed to other
servers? Test against ZServer, and not Zope behind another. I can show
conclusively, and I doubt any here would doubt it anyway, that ZServer
is faster than any other method of serving up Zope (that we have found
anyway).
 
> Is there a Zope Advocacy FAQ somewhere I can use when people throw
> things like this at me?

Mostly just apply principles of reason and logic. Ask for specifics,
ask for sources, then carefully, and accurately deal with what is, or
is not, provided. Often times, there will be none provided. Anectdotal
evidence should be taken lightly, with little import when considering
advocacy arguments. Otherwise, you wind up with an anecdote war, wich
is neither helpful, nor useful.

I would take claims of 'taking down linux' with a very hefty dose of
salt. ATM, I don't recall any reports of ths happening.


Stay objective and rational; and if that doesn't work, conclude with
'Not all apps are all things to all people. Feel free to use your
own.'. 

:-)