On Wed, 22 Aug 2001 07:01, Leo Simons wrote:
> ...aren't well-defined at all. Methinks.
nope.
> ---------------------------------------------
> Part of this is already true. But there is a
> large duplication of code and effort across
> the various projects.
>
> There are three server frameworks (Avalon,
> Tomcat, Turbine), they have three pooling
> systems, two application expansion systems
> (catalina and camelot).
> They all provide services. Avalon does
> deployment, so does tomcat. Turbine, Avalon
> and tomcat all do logging using one of the
> two toolkits.
>
> Just by looking at class names and pattern
> descriptions, this is obvious.
>
> More concrete: doing a search across apache
> sources for "Service.java" gives 6 results.
> ---------------------------------------------
yep - in many cases, the evolution of systems also follows a very specific
path. Initially everything is available via static methods and has flat
architecture and much of config info is flat.
Slowly over time the system will evolve to allow multiple instances per JVM
(ie don't rely on static access so much) and layered approach. TC4 and
Phoenix are 3-layer systems (Kernel-App-component) while turbine is still two
layer system (App-component).
There is also remarkable similarities in evolution of other aspects
(container-component contracts, config formats, lifecycle management,
services architecture etc). This is not restricted to Apache though and
similar evolutions can be seen in OpenEJB and jBoss. You will note that in
almost all cases the "top" layer (ie kernel) is configured by flat properties
while lower layers (if any) are almost always configured via hierarchial
config (like XML).
I find these sorts of trends fairly interesting ...
> Now the question is: do we wish to remove
> all that duplication of efforts, create a
> logical "separation of concerns" between
> projects?
>
> Probably: yes.
>
> I was just wondering how we're going to do
> that.
The problem is not code but people.
Most developers are arrogant and stubbon. If they "discover" X they often
think they are first to know about it and everyone who doesn't use X is
foolish. Some even go as far as to force their vision on others. However it
is very rare that X is anything new and forcing their vision can sometimes be
bad.
Now the above statement can be seen as both a good thing and a bad thing. If
you look back through the archives it describes some of the things I did
quite well ;) While I think in most cases, X was a good thing, there is also
some mistakes I made (arrogance blinded me at the time).
however there is also a very bad side-effect of that. Sometimes some people
will not bother to evaluate anything based on technical merits (ie how good
is it, how widespread is it, how likely to be supported into future etc).
They instead base it on semi-religious beliefs, moral arguements and other
trite. Ironically they are also the same people who slam GNU ... kettle ...
black ... pot ... black ...
There is also a fairly big anti-standards here. Sometimes this is good
because we get fab tools out of it (ie Velocity) but othertimes it is less
than good. Even now we have esssentially 3 different build-system styles at
jakarta, some differences are needed while others are just there to satisfy
egos.
Anyways I would love to see Avalon Framework + parts of Excalibur more
universal but I guess I have been burned a few times trying to get universal
acceptance of anything .. let alone something as controvertial as a
framework. Especially considering most developers seem to disdain security,
reusability or any quality control. Changing developers is as easy as herding
cats.
Sam Ruby has been the only person inside apache to get this sort of
cooperation ... not sure how we could do it. I plan to just use stuff in my
own projects. If others use it then great, if they choose not to use it then
thats fine aswell.
--
Cheers,
Pete
*-----------------------------------------------------*
* "Faced with the choice between changing one's mind, *
* and proving that there is no need to do so - almost *
* everyone gets busy on the proof." *
* - John Kenneth Galbraith *
*-----------------------------------------------------*
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
