RE: MathML-Presentation specs criticized.

> Er, what about 5.3.2 "Fine-grained Parallel Markup" and 5.3.3 "Parallel Markup
> via Cross-References: id and xref"? They allow you to associate presentation
> and semantics not only for the whole expression but for each subexpression,
> which seems to be what you're after here.

I forgot about those sections!  Thank you very much. :)

<read>sections.....</read>

Although the use of semantics elements does address my criticisms, the solution
just doesn't "feel" right.  Content markup and presentation markup can be linked
using these elements, but this linking is only a footnote to the original
encoding because anything within annotation elements are ignored.  Furthermore,
a parser would have to use heuristics to figure out what to use the annotation
for, if anything (especially with other annotations in the same semantic tag).
Granted these are weak arguments at best.

There are three more substantial points against the current specification of the
semantic elements that I have to make.

1) Although annotation and annotation-xml do the same thing (with different data
types), they are separate tags.  Shouldn't there be a type="text/xml" or some
other attribute to encode the type of content (note: not the same as the poorly
named "encoding" attribute)?  This (with an optional "src" attribute) extends
the annotation element to provide additional information about the primary
markup (such as an image or MathSpeak rendition).  However, this nullifies the
"annotation-xml" tag as unnecessary.

2) Annotation should be used to provide additional information about the markup
and not specify a secondary rendering option (such as an image or text) although
I find that could happen without some warnings.

3) Not really against Annotation, but errors in the PDF (that I have, at least)
and the (x)HTML.  Shouldn't there be a namespace in the following so that the
XML-grammar is explicitly defined:

<semantics>
  <apply>
    <plus/>
    <apply><sin/>
      <ci> x </ci>
    </apply>
    <cn> 5 </cn>
  </apply>
  <annotation-xml encoding="OpenMath">
    <OMA><OMS name="plus" cd="arith1"/>
      <OMA><OMS name="sin" cd="transc1"/>
        <OMV name="x"/>
      </OMA>
      <OMI>5</OMI>
    </OMA>
  </annotation-xml>
</semantics>

should be:

<semantics>
  <apply>
    <plus/>
    <apply><sin/>
      <ci> x </ci>
    </apply>
    <cn> 5 </cn>
  </apply>
  <annotation-xml encoding="OpenMath">
    <OMA xmlns="http://www.openmath.org/OpenMath">
      <OMS name="plus" cd="arith1"/>
      <OMA>
        <OMS name="sin" cd="transc1"/>
        <OMV name="x"/>
      </OMA>
      <OMI>5</OMI>
    </OMA>
  </annotation-xml>
</semantics>

Or, is annotation-xml depreciated and I didn't notice that.  (I make too many
mistakes, don't I? :)

---
Jimmy Cerra





-----Original Message-----
From: www-math-request@w3.org [mailto:www-math-request@w3.org] On Behalf Of Tim
Bagot
Sent: Wednesday, April 17, 2002 3:38 AM
To: www-math@w3.org
Subject: RE: MathML-Presentation specs criticized.

At 2002-04-17T00:15-0400, Jimmy Cerra wrote:-

[...]
> I feel there should be a better way to combine presentation and
> semantic markup  I still don't know why a generic scheme for either combining
> generic presentation** markup and semantics or extending presentation markup
so
> authors can specify the exact meaning of their presentation markup isn't
> necessary or in the specs(did I use a double negative!?!?).

Er, what about 5.3.2 "Fine-grained Parallel Markup" and 5.3.3 "Parallel
Markup via Cross-References: id and xref"? They allow you to associate
presentation and semantics not only for the whole expression but for each
subexpression, which seems to be what you're after here.


Tim Bagot

Received on Wednesday, 17 April 2002 17:19:53 UTC