RE: About multiple output documents from a single XML/SGML processed document

["Didier PH Martin" <martind@xxxxxxxxxxxxx>]

Hi Joerg,

Avi Kivity <Avi@xxxxxxxxxxxxx> wrote:

> I would rather not extend the dsssl model any more than necessary, as this
> would probably involve modifying all backends and losing compatibility to
> other dsssl engines. The simple solution exists today: run Jade once per
> output file, with parametrized stylesheets. It could be optimized to use a
> single invocation of the program, so as to minimize the parsing necessary,
> but no more.

Hoerg said:
I'd second that approach.  When I read the proposal, my feeling was
"this is clearly something to control from the command line".

Don't extend openjade even more beyond the dsssl spec, please.

Didier says:
You probably read the two previous messages and answers. But I'll repeat the
same message so that it can be understood.

There will be, in a near future a DSSSL-2 draft specification. We have to
think about it, prepare it, discuss it, propose it, modify it if necesary,
get it accepted by the international community. We already started the
process in a very democratic way and take the opportunity to ground it form
people's need. Obviously people's need differ and we have to be empathic
enough to understand the problem from their perspective not ours. If we can
ground this new spec on people's need we will have gained something. A major
mechanism that tremendously helped XSL is the XSL mulburry discussion group,
where people coul express their opinion, their views and have an open
discussion with the specs designers. I know, this is a good process because
I was also an active participant to these discussions and learned the
virtues and weaker side of democracy. However, from the discussion
interesting things emerged and XSL got improved.

We have here the same opportunity to build a good version 2 spec from the
experience of our user group and their needs. Like I said, the needs are
differents and we have to respect some constraints. I believe that we can
make this DSSSL-2 spec a success if we ground it on real problem
resolutions.



Joerg said:
If the majority would like to see such an extension, I'd suggest *not*
to extend existing FO's with new characteristics (which would drive
newcomes up the wall, when comparing extended examples to original
specs).  Rather add/extend the existing extensions of jade (the
sgml FO "entity" of jade comes to mind).  If you are to modify your
style sheets to create multiple output entites it would be easier to
simple wrap the old one by something like

(element whatever
  (make entity
    ...
    (next-match)))

than go into the old style sheets and add a new characteristic for
each simple-page-sequence etc.

Didier said:
several thing in your comment:
a) about the specs and documentation. We won't let the newcomers in the dark
with new features. Do we document OpenJade flow objects and some of its
feature? is it too slow? if to the second question the answer is yes, I am
deeply sorry not to be able to put more time and is even more sorry to be
alone to document OpenJade (but luckily not alone on the code part, Avi and
Matthias are precious colleagues). I could put more time, if a) I was
millionaire and would not need to work to pay my bills, b) people would pay
for OpenJade. Obviously, if monetary stuff would be resolved, we would put
zealous efforts to provide the needed documentation as fast as possible.
But, I am deeply sorry to say that I can put only limited time on this, and
that I should say thank you to Talva to sponsor some of my efforts (which
takes an average of 30 to 35 hours per week), and should also say thank you
to my spouse to accept that I put so much task on a labor of love but not of
money :-) This said, we won't add any feature without documentation, period.
so, newcomers won't get lost as long as they know where to get the OpenJade
doc and code (sorry Jade is a thing of the past and a beautiful gift James
gave us - we won't have enough thank you for the time he invested in it).

b) As a short term solution for the user's problem, this will work but we
didn't 1) increased his productivity, 2) provided him versatility with a
rule based solution, 3) did not simplified the process. So, as a short term
solution using the "SGML" mode may work, but I am no so sure that we
addressed the problem, maybe just skipped it.

c) Your solution is technically exact, we can do this as also suggested by
Brandon as a short term solution until, we, in the OPenJade team find a good
solution and we, in the DSSSL community help prepare a good draft
specification and we, the ISO community, work hard to write the spec,
comment it, modify it, approve it.

Thanks for your input, I'll be happy, when I'll see in your messages
constructive comments and suggestion for the DSSSL-2 draft specifications.
Joerg, we would greatly appreciate your help but focused on the future not
the past and the status quo. Help us build a better DSSSL-2 specification
and a brighter future for OpenJade. This said, your opinion is respected as
is an input to our reflection.

regards
Didier PH Martin
mailto:martind@xxxxxxxxxxxxx
http://www.netfolder.com


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