(dsssl) Re: OpenJade minutes

[Javier Farreres de la Morena <spanish@xxxxxxxxx>]

T. Shipley sent some days ago the minutes of the discussion about the future
development of OpenJade. They were supposed to organize the discussion,
but there has been a bit of silence since then.
First of all, I want to thank Shipley for his effort to organize the different
topics that were discussed so randomly along the days and putting them in
such a readable form.
In my oppinion there are some topics to talk about. To follow this reply,
you will need consult concurrently the minutes written by Shipley.
Under the topic Mission, Shipley talks about two possible goals. He has
very correctly evaluated the different flows of oppinion. About this theme
I think there are two points. The first is the interest to spread OpenJade.
The development of page models would be a good way to do it. But there is
also the need to reorganize the overall class architecture, what would guide
towards version 2. What I am considering is to do both things at the same
time. One of my students wants to develop page models. This would be done
under the current design of OpenJade, directed to generate a 1.5 or 1.4
 version. It could be studied it this were a good point to introduce two
wat communication, or if this would mean a complete reorganization of the
style process. I have two other students interested in the project, and
I would dedicate them to the work of generating a version 2 of the tool.
On one side, one would develop a complete grove model (as complete as he
can, I don't know if it will be complicated or simple), and with the other
I would develop an UML model of the existing tool and a reingeneering process
oriented to a good separation of classes and modules.

Under topic Philosopy, Shipley points out the various oppinions shown along
the discussion confronting XML and SGML. In my oppinion, while assuring
the implementation of the proper standards, extendibility should be assured.
First, because standards can evolve. Second, because we can provide more
than the standard sais. There should be no problem in offering XML oriented
stuff. The problem is, I think, the lack of propsets in the XML world. Anyway,
I consider it is good to advance in small steps, and I think it is a good
short term aim to just implement the standard as it is.

Under topic Management issues, I must say the number of students volunteered
will vary with time and will be discontinue. The students carry on a project,
that lasts usually 6 months, and then the fly to the real world. My students
usually have studied 3 years of computer science, and have a medium degree
(no MIT :) ). About other professors volunteering, I doubt. Few people in
the world know about DSSSL. And professors in my department are more interested
in research projects. And this one is not. It is totally practical. Apart,
it should be possible to adapt external help from any volunteer who wants
to help. I think SourceForge and CVS already helps in this. And about the
code, my students will keep historical coherence. The bunch of the code
in OpenJade is c++ and we are not going to change it now. There would be
any reason for changing it and recoding everything that I am not aware of?

Under the topic Engineering and computer science issues some comments are
made about FOT. As long as I know, in the DSSSL standard FOT is part of
the style process. It should be maintained. But it is something internal
to the tool or the process, and style rules affect to the construction of
the FOT and also rule the formatting of the Flow Objects. In my oppinion,
we cannot rely on Xpath/Xquery for the reason that they are not grove aware.
We could though offer them as extensions. I don't have a clear idea what
would suppose accepting Schemas. As I remember they rest atop the DOM layer.
I can't follow many of the HyTime comments written by Shipley. I have tried
to read the HyTime standard but it is clear I didn't learn so much from
it as Shipley did.

Many things are still in the minutes to be discussed but they are more concrete,
and first I want to make clear these general concepts.

I wait for replies on everything I have discussed.

Javi





-------------------------------------------------------------
http://correo.inicia.es
-------------------------------------------------------------





 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist