[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

RE: Formatting Objects considered harmful

Subject: RE: Formatting Objects considered harmful
From: "Didier PH Martin" <martind@xxxxxxxxxxxxx>
Date: Sun, 18 Apr 1999 10:24:32 -0400

Hi Liam,

I am opposed to splitting XSL into two parts because I think Microsoft
will implement one part and Netscape the other.

unfortunately the reality is more complicated than that. Actually, Mozilla
has no XSL. If a XSL processor is included, it will come from a third party
(not developed by Netscape). To implement the FO part is not easy. Here is

a) Mozilla is re-implemented as a set of components
b) It includes a new engine with HTML objects with CSS properties.
c) To implement XSL Fo require a mapping from XSL FO to HTML+CSS Fo. Again
Why? because if it is different, we'll have to re-write the layout engine.
Forget it, we cannot afford this. So, all XSL FO has to be mapped to
HTML+CSS Fo. For example, do we map a block object to a DIV and a SPAN to an
inline object? Also, the actual XSL FO model has objects difficult to map to
HTML+CSS FO. Bottom line, to rewrite a new layout engine may need again 1 to
2 years, forget it! we cannot afford it.
d) This is also the same problem with Microsoft. They invested a lot in a
new engine named Form+. This engine is based on HTML+CSS object. Will they
invest again in a new engine with new FOs? forget it, they have to deliver
something (i.e. windows 2000) ASAP.

However, a subset could be implemented but this bring some problems with
specs compliance.

This is realities we have to deal with.

To really understand the issues I am talking about:
a) Thing how to map HTML+CSS objects to XSL FO objects. If it match, yes
this is an object that could be implemented in the XSL interpreter. If not,
it means that this object has 1) to be removed form the spec 2) cannot be
implemented without investments.

But actually, I can tell you that it is a lot easier to parse HTML+CSS
object contained in a template and have the XSL interpreter use the HTMLDOM
to create the layout. Here are the advantages:

a) monetary. The investments are already made, we can then have a ROI on our
work. Even if it is free, have it code complete mean an economy because we
have no more expenses (time + money to organize the group + tests programs,
b) The set of objects are implemented now with an interface: the HTMLDOM. So
an XSL interpreter encountering a HTML tag can use the HTMLDOM to place the
object on the screen layout. However, if the XSL FO object could be mapped
to a HTML+CSS object, then it could be done quite easily.
c) People invested to learn HTML FO, we re-use their knowledge.
d) a lot more to say but it is sunday morning, its sunshine and I'll go with
my bicycle forget about these issues and enjoy life :-)))

Didier PH Martin

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list

Current Thread