Re: XS: Ports --> math

[James Clark <jjc@xxxxxxxxxx>]

At 10:05 23/05/97 +0100, Sebastian Rahtz wrote:
> > I found ports the hardest thing to implement in Jade.  I also think they're
> > hard to explain/understand.
> > 
> > Ports are currently essential for two flow objects:
> > 
>They are used in the math flow objects; do you mean that they are not
>essential there?

I mean ports aren't essential for any flow object class in XS other than the
two I mentioned.

>Which brings me to math. Jon started to say that basing XS on what
>Jade had implemented was not a bad idea; but James _has_ implemented
>the math objects, and I would suggest they deserve a place in
>XS. Since the W3C has in parallel with XML beavered away on defining
>MathML as an application of XML, it would seem perverse for XS not to
>deal with it.
>
>James raised the question of whether MathML can be dealt with by DSSSL
>math flow objects; I haven't looked in any detail at all, but my
>initial impression is that for the presentation math tags, it should
>not be a problem. For the semantic math, other questions arise,
>perhaps. One of the arguments in favour of semantic math markup is
>that it provides a rational way of working out equation breaks. Since
>DSSSL math is presentation-based, the wrapping DSSSL code will have to
>do the understanding of semantic math, right? But its hard to do
>equation breaking without font metrics, so how will DSSSL communicate
>to the formatter about breaking?

I would handle this by adding an atomic math-breakpoint flow object that is
used to indicate that an equation can be broken (with some specified degree
of desirability) at this point. It would then be up to the DSSSL style-sheet
in generating DSSSL's presentation-oriented math flow objects from the
semantic math to insert the math-breakpoint flow objects at the appropriate
points.  (I don't know if math-breakpoint is the right way to handle it: in
general what is needed some sort of flow objects that allow the effect of
the semantics on breaking behaviour to be specified by the style-sheet.)

James



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