RE: html to dsssl ?

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

Hi Sebastian,

Sebastian said:
hang on. you cannot have your cake and eat it. two reasons why DSSSL
has not taken off are

 * the lack of any implementation of the `advanced' layout features
 * a general distrust of unproven abstract formatting languages

Didier says:
I would say mainly marketing instead. W3C is doing a very good public
relation job. Its members too. So, it is not a question of language virtues
but more of marketing.

Sebastian said:
So
 1. if someone goes to the trouble of making a serious formatting
    engine do advanced layout from DSSSL/XSL FO input, are they going
    to retrofit it to OpenJade, or start from a clean base with XSL?

Didier says:
I won't teach you that any DSSSL document is a SGML document. Therefore, all
DSSSL documents are parsed, more particularly by SP. The module doing this
operation is "style.dll" and more specifically the class responding to SP
events is "DssslSpecEventHandler". So, nothing prevent somebody to implement
a XSLT engine to what's already there or simply to improve the actual DSSSL
to incorporate template constructs. We can even learn from the XSL
experience and debug some of its flaws (like XSLT did with some DSSSL
flaws). Why not?


Sebastian said:
 2. if the market fails to accept the abstract formatting objects
    concept, that both DSSSL and XSL FO are doomed together

Didier says:
I agree that if the general population has trouble learning abstract
formatting objects these two languages are in trouble. In this sense, CSS
has more chances because it only support a limited set of basic objects:
page, block, inline. Easier to remember :-)
For more info there is a series of articles I am writing (note that it is a
work in progress). These articles talk about XSL, CSS and DSSSL models
http://www.netfolder.com/CSS/CSSObjects.htm
http://www.netfolder.com/DSSSL/Paragraph.htm
I am still writing the XSL document because of the recent modification on
the spec. Something new that may help all style languages based on abstract
formatting objects: visual tools or dialog based tools. This may help a lot
people to use abstract formatting objects (at the first level, without any
complicated code)

Sebastian said:
I don't really mind whether XSL FO takes off, or OpenJade causes a
pure DSSSL renaissance. But as it stands today, both formatting
languages are very unproven and unimplemented. The fact is that you
cannot typeset more than fairly simple documents with any of the DSSSL
backends at present (yes, I know most documents _are_ simple!), and
until that situation changes, where are DSSSL/OpenJade/XSL FO going,
except down the drain? You might as well print from HTML and have done
with it.

Didier says:
Your point is well taken. We recently, in this list, got a request for a
more sophisticated page model for variable margins. This is a must for most
book publishing. OpenJade has to be improved on this side. Also, a PDF
backend would greatly help for printing and pre-viewing the result. This is
in the "to do" list :-)

Didier said:
 > Actually, here is what is happening in the XSL world. Most of the
 > implementations are made in Java except Keith Moore which is working on a
 > C++ one (and off course the microsoft implementation).
Sebastian said:
do I, as a user, care what the implementation language is? no....

Didier says:
but you may care about the speed no?

Didier said:
 > Don't you think that the OpenJade world could improve the actual product
to
 > add new features like templates, new script languages? Let's supose for a
 > minute that it would be the case. We then have a chance to have a better
 > tool than what's actually available.
Sebastian said:
sure, make OpenJade a tool which can parse either XML or SGML (as it
can), and apply either XSL or DSSSL stylesheets, and use a common
backend --- why then the world and all thats in it will be yours,
and what is more, my son, you'll be a man![1]

Sebastian

[1] from a probably misquoted poem by Kipling, just in case anyone
thinks I am being sexist :-}

Didier says:
Or have some concept form XSL imported into OpenJade. the goal is not
necessarily to copycat XSL and to copy its flaws (and good sides too) or
more simply, to include XSL as is without adding any more benefits to what's
already present in OpenJade. As you, I learned XSL, and used it enough to
spot certain flaws, certain limitations (and its good sides). There is two
features I appreciated in XSL: XPath and the template mechanism. As you
know, DSSSL has several flow objects construction rules (ref:
http://www.netfolder.com/DSSSL/MarkupRule.htm - for a complete list of them)
but do not have the concept of template based construction rules. I am
juggling with the idea to add this to the actual construction rules.
Actually, just on paper to test the model, later on with code to stress test
the idea.

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



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