W3C home > Mailing lists > Public > www-math@w3.org > November 2004

Re: why is declare invisible? how to say "Let X be ..."?

From: Stan Devitt <stan_devitt@stratumtek.com>
Date: Mon, 22 Nov 2004 07:43:51 -0500
To: Dan Connolly <connolly@w3.org>
Cc: www-math@w3.org
Message-Id: <1101127431.9328.23.camel@radical.stratumtek.ca>

Dan,

Actually, the answer to your post is partly in your post.  You already
use two quite different notations which might possibly map onto a
declare:
both

> 	RDF_T ::= RDF_U \U RDF_L \U RDF_B.
> 
and
> 
> Or is there some other preferred mathml syntax for "Let X be..."?
> 

For the most part, presentations of content constructs were included
when there was a notation that we were aware of that was in broad common
use and could be agreed upon.

This presentational difficulty was further compounded by the fact that
declare can be used to provide, for example, default attribute values
and other less easily visualized bindings of names to meanings (e.g. by
assigning definitionURL's.). 

It was originally envisioned as having a semantic role separating
definition from use, somewhat akin to the way CSS separates
presentational definitions from use, and as a macro definition facility.

While its was original conception was as a semantic style sheet
mechanism for the entire document, we did not see any easy mechanism to
complete that separation with existing W3C technologies, or even to make
the definitions apply to an entire document the way we wanted it to
behave and much of that benefit is lost when it is limited to a single
math element.  In the end we decided to wait for a general construct to
emerge and tackle this at MathML-n   rather than invent a construct that
would need explicit support in browsers to ever be used.

Of course, while it has no default presentation, the spec is still quite
tolerant of user specified translations from content to presentation.  

Hope this bit of "history" helps...

Stan Devitt


On Sat, 2004-11-20 at 14:13, Dan Connolly wrote:
> I'm working on formalizing the definitions in the SPARQL
> spec[1], and since my larch tools[2] are getting crufty and
> I just saw a bunch of nifty MathML products at XML 2004,
> I'm trying to use MathML. I'm copying from the nifty
> MathML test materials
>   http://www.w3.org/Math/testsuite/static_toc.html
> and using the stylesheet there
>   http://www.w3.org/Math/testsuite/style/mathml.xsl
> which seems to work pretty well in Mozilla firefox.
> 
> My work-in-progress is
>   http://www.w3.org/2001/sw/DataAccess/mathml-rules.xml
> 
> For example, I encoded
> 
>   The set of RDF Terms, RDF-T, is RDF-U union RDF-L union RDF-B.
> 
> as
> 
>   <math xmlns="http://www.w3.org/1998/Math/MathML"> 
>     <declare type="set">
>       <ci> RDF_T </ci>
>       <apply> 
> 	<union/> 
> 	<ci> RDF_U </ci> 
> 	<ci> RDF_L </ci> 
> 	<ci> RDF_B </ci> 
>       </apply> 
>     </declare>
>   </math>
> 
> expecting to get something like
> 
> 	RDF_T ::= RDF_U \U RDF_L \U RDF_B.
> 
> but the test materials say:
> 
> "Description:
>           	declare test (no visible rendering)"
>   --
> http://www.w3.org/Math/testsuite/testsuite/Content/BasicContentElements/declare/rec-declare3.xml
> 
> 
> Am I just using the wrong stylesheet to lay out my MatML?
> Or is there some other preferred mathml syntax for "Let X be..."?
> 
> 
> [1] http://www.w3.org/2001/sw/DataAccess/rq23/
> [2] http://www.w3.org/XML/9711theory/#about-larch
Received on Monday, 22 November 2004 11:42:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:56 GMT