I see.
When I read your previous email I felt it was a bit odd because our systems
(bivio, extropia, openinteract) are web systems.
I think there are a couple why-nots though. I guess I am a bit of a
contrarian, so don't let that discourage you. Just points to address.
1) I don't like the idea of any one of our 3 systems mentioned becoming the
core framework because although it may be a short-cut it may also be quite
ummm political.
Each of those frameworks has a tie back to a different user community. By
adopting one as a P5EE standard, it will risk alienating the others. Think
of the Template community and how many template toolkits there are (even
one called Template Toolkit heh)...
If we were to choose one as "the" standard, it would alienate the
alternative communities.
So instead, the better thing to do (as far as templates are concerned) is
to choose an initial one to integrate with but leave enough of an open API
that allows them all to integrate and be advocated equally depending on the
requirements of the person.
Similarly, I am not sure that the 3 web app frameworks should be forced to
"compete" against each other for what may be a "coveted" P5EE recognition.
And then the other 2 are forever not part of P5EE because they never made
it in the core.
2) I think all three frameworks are actually unfortunately, in different
states of completion and/or support.
For whatever reason, many toolkits do ebb and flow. SmartWorker used to be
quite strongly advocated 3 years ago. Now it is latent because the company
behind it is essentially latent. The same may be true for any of the 3
toolkits. eg what if Chris Winters ends up being forced to do Java
programming because Perl jobs aren't as easy to come by -- causing more
stagnation. Similar pressures may hit eXtropia and bivio.
Yes, there are other people besides Rob or Gunther or Chris working on
these frameworks -- BUT most frameworks do require a leader and they do
tend to stagnate more than less even with others in the open source
community contributing.
Also, are these projects willing to hand over their source code trees to
P5EE? What about copyright? If the copyright to eXtropia toolkit is owned
by eXtropia, although published under Gnu, is this "enough" to base P5EE on
when P5EE may want to release what they do under Perl Artistic License?
Would p5EE expect the chosen toolkit to give the copyright to the public
domain p5ee escrow organization like yet another society?
p5ee may also "unfairly" tell people that one group's web app framework is
better than all the others and it will become the "standard" through
marketing of p5ee label. But then what happens if the company or person
who was originally behind the web app framework decides to leverage the
newfound popularity financially? eg stopping the production of completely
free releases...
3) It may be easier to choose as a code base something more base than a
full web app framework. For example, I would prefer what Brian Moseley did
to make Perl Servlets to become the "standard".
Then for a framework like Bivio or eXtropia or OpenInteract to be
considered for inclusion in P5EE, it should have wrappers to allow it to
work with Wombat... making the criteria for entry into P5EE the ability to
interact with other things in P5EE.
===
Anyway, these thoughts from me are quite random as I realized they needed
to be said, even though I normally would not have posted due to my schedule
(just coming back from a long vacation and catching up).
The architects behind Bivio and OpenInteract may obviously feel differently
than me. :)
Later,
Gunther
At 06:29 AM 6/2/2002, Stephen Adkins wrote:
>Hi,
>
>There was some confusion in the previous threads of discussion
>about whether the P5EE is a "web application framework" or
>"some kind of distributed systems thing".
>
>So let me define what my vision of P5EE is (in yet another way).
>
>I am primarily interested in building web applications.
>However, when I build a web application, I desire it to have the
>Attributes of an Enterprise System so that I don't find that
>my application has nowhere to grow.
> http://www.officevision.com/pub/p5ee/definitions.html
>
>For me, the P5EE must pass the following three tests.
>
> * P5EE facilitates best practices in web application development
> in such a way that those Web Applications are excellent
> examples of Enterprise Applications, exhibiting the full range
> of Attributes of an Enterprise System.
> * The application components (data access components
> and business object/logic components) and services (Security,
> Messaging, etc.) are equally useful in other types of
> Enterprise Systems.
> * The Web Applications integrate with and interoperate with
> the other Enterprise Systems so that they may act as in
> concert as a single integrated Enterprise System.
>
>Allow me to tie in some of the comments from the previous
>threads of discussion.
>
>------------------------------------------
>support webapps first priority
> Steve Lihn
> http://archive.develooper.com/p5ee%40perl.org/msg00950.html
>
>Steve Lihn thought it was a good idea to make web applications
>the first priority. I agree with this wholeheartedly.
>This is what I was doing because it is what I most need.
>
>------------------------------------------
>Start with and solve a problem.
>Do the simplest thing that could possibly work.
>Don't implement anything you don't need today.
> http://archive.develooper.com/p5ee%40perl.org/msg00951.html
>
>Rob encouraged us all not to get too Astronautically Architectural,
>but to focus on solving real problems.
>I agree with wholeheartedly.
>
>This is in fact what I have been doing.
>I have not always shared all of the details because
>they are confidential details with my clients for whom I am
>developing systems using P5EEx::Blue.
>
>------------------------------------------
>designed from the perspective of web development
> Matt Sergeant
> http://archive.develooper.com/p5ee%40perl.org/msg00945.html
> Dave Rolsky
> http://archive.develooper.com/p5ee%40perl.org/msg00952.html
>
>Matt and Dave felt the framework was too biased toward web
>applications. (This criticism puzzled me extensively.)
>I had in fact gone out of my way to envision many possible forms
>of applications, operated from many possible user interfaces,
>backed by many possible forms of data storage. I am guessing
>that this feedback was based on the fact that I have built out
>the classes for the web application user interface more than
>other user interfaces.
>
>This is as it should be. It will take someone interested keenly
>in other user interfaces or application contexts before those are
>fleshed out. However, the groundwork has been laid, and the
>P5EEx::Blue is hardly restricted to web applications.
>In fact, I am currently working on some command line programs
>and some forking-daemon programs which use P5EEx::Blue.
>
>I acknowledge that the general nature of P5EEx::Blue's architecture
>will not be proven until more of these application contexts and
>user interfaces have in fact been implemented and the kinks worked
>out. As a result, certain methods sometimes carelessly got placed
>into a class when it should have been in a more specialized subclass.
>
>------------------------------------------
>P5EE seems to be going after distributed systems like J2EE.
> Rob Nagler
> http://archive.develooper.com/p5ee%40perl.org/msg00961.html
>
>When I asked why bOP should not be considered as a candidate for
>P5EE, Rob seemed to think that P5EE was somehow not targeted
>at the web application developer.
>
>This is why I am seeking to clarify "What we are trying to do here".
>This is also why I see OpenInteract, bOP, and the Extropia Toolkit
>as viable candidates to at least be considered for the starting
>code base for P5EE.
>
>P5EEx::Blue is a prototype (albeit one that is in production in
>a couple of places, using subsets of its envisioned functionality).
>I have learned enough from doing it that I can weave its good
>qualities into a new code base if necessary.
>
>Perhaps these three other code bases are not architected
>perfectly to fulfill all of the Attributes of an Enterprise System.
>Fine. I simply wish to consider all options before we commit to
>a starting code base.
>
>As always, all feedback is welcomed.
>
>Stephen
__________________________________________________
Gunther Birznieks ([EMAIL PROTECTED])
eXtropia - The Open Web Technology Company
http://www.eXtropia.com/
Office: (65) 64791172 Mobile: (65) 96218290
