RE: Time taken to impliment FOs?....

Subject: RE: Time taken to impliment FOs?....
From: Guy_Murphy@xxxxxxxxxx
Date: Thu, 11 Feb 1999 17:19:24 +0000

I was under the (possibly wrong) impression that as rendering of XML with
CSS is possible there must be a high degree of seperation between the
rendering engine and the interpreter, where did I go wrong in this notion?
I confess to not knowing enough about how a browser rendering engine works
under the hood.

I agree with your further comments that the transformative part of XSL can
be nailed down and declared complete before the FOs (and have previously
advocated that this should happen), just that this should b seen as
completion of one half of XSL rather than a seperate part :)


xsl-list@xxxxxxxxxxxxxxxx on 02/11/99 05:59:13 PM

To:   xsl-list@xxxxxxxxxxxxxxxx
cc:    (bcc: Guy Murphy/UK/MAID)
Subject:  RE: Time taken to impliment FOs?....

I cannot speak for Microsoft. But to implement FOs you need to separate
display engine from the HTML interpreter. The browsers (if you look at the
code) still have a big dependency between the rendering engine and HTML
interpretation. On its side, Microsoft is working since a certain time on a
technology called Form+ that was called internally: Form3. That engine is a
generic display layout engine.
 The surest path is to first provide an API to the display engine like for
instance DOM2 even if it is unnecessarily verbose. An other API based on
composite pattern would have been better (and the notion of property set
like, for instance the ADSI API has). It remains that the first step is to
create an object model for the display objects, CSS2 propose such a model
and is a lot more advanced with its concepts than XSL formatting object
In fact XSL FOs have to wait for CSS2 (in actual review) objects. When you
have this model in place and the corresponding API, you have independence
the display engine toward any interpreter. So, a) the first goal for
engines is to de-couple the display engine from any interpreter b) have an
object model for each rendering object (independent of any language). Form+
can fulfill this for Microsoft and NGLayout for Mozilla. But on both side,
this is work in progress and it will take a certain time before having
something stable enough to be used. However, Form+ has the advantage that
the effort started 4 years ago and the team can leverage the knowledge they
gained from VB, VBA (FORM2 is included in VBA). NGLayout concepts are also
the reflect of good knowledge. So Mozilla needs volunteer to move faster on
this side. Volunteers? not a lot? it will take more time then :-)

 XSL-List info and archive:

Current Thread