W3C home > Mailing lists > Public > public-html@w3.org > March 2008

Re: Exploring new vocabularies for HTML

From: William F Hammond <hammond@csc.albany.edu>
Date: Mon, 31 Mar 2008 11:10:33 -0400
To: public-html@w3.org, www-math@w3.org
Message-ID: <i7d4pboro6.fsf@hilbert.math.albany.edu>

James Graham <jg307@cam.ac.uk> writes:

>> With a small amount of text editor support such as closing open
>> elements and less perhaps small amount of text editor support such as
>> auto-completion of element names based on language knowledge, it doesn't
>> really take many more key strokes to type
>>  \frac{\partial \mathbf{D}}{\partial t}
>> as
>> <mfrac> <mo>&PartialD;</mo><mi mathvariant="bold">D</mi></mfrac>

Actually there's no denominator in this MathML.  (The pace of this
discussion has become a bit frenzied, hasn't it.)

<mfrac>
  <mrow><mo>&PartialD;</mo><mi mathvariant="bold">D</mi></mrow>
  <mrow><mo>&PartialD;</mo><mi>t</mi></mrow>
</mfrac>

The original LaTeX is much closer to this author-level XML:

<frac>
  <numr>&PartialD;<mathbf>D</mathbf></numr>
  <denm>&PartialD;t</denm>
</frac>

Less cpu is burned going from the LaTeX to the author-level XML than
is burned going from the author-level XML to the MathML.  Most of that
is involved in accurately constructing, with a recursive descent, the
tree of the mathematical expression -- something that has an abstract
mathematical reality independent of the design of MathML (but not
invariant from author to author).

If math is included in html5, will html5-browsers be 'required'
to support it?  (Have any so committed?)

Even if html5 browsers are going to provide such authoring services that
really belong in the author's shop, I fear that html5-math will wind
up with the reputation of being unacceptably slow.  (And I fear that
could be its downfall.)

The following xhtml+mathml item has 1086 mathml nodes, out of 3250
elements in all, and is still a bit slow for a document of that size:

      http://math.albany.edu/math/demos/nyjm/9-8.xhtml

I don't want to think about slowing down such an article significantly
more with authoring services in the browser.

Math aside, ordinary HTML is not a convenient format for direct authoring.
So why all the attention to the idea of direct authoring of math?

                                    -- Bill
Received on Monday, 31 March 2008 15:11:17 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:53 UTC