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

Re: SVG is not MathML

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Wed, 30 Jul 2008 23:54:07 -0700
Message-ID: <4891618F.70108@mit.edu>
To: Doug Schepers <schepers@w3.org>
CC: HTML WG <public-html@w3.org>

Doug Schepers wrote:
> Henri Sivonen wrote (on 7/30/08 7:04 AM):
>> One would think that asking them to expand what for MathML is an 
>> existing feature in Firefox would be easier than asking them to deal 
>> with the complexity of the SVG WG's parsing proposal.
> In several messages, some people are quick to equate MathML with the 
> proposal for SVG-in-HTML.


It seems to me that all Henri was saying is that Firefox has an existing 
"view XML serialization of this MathML DOM fragment" feature, and that 
it would presumably not be difficult to remove the "MathML" qualifier, 
thus producing a general "view XML serialization of this DOM fragment" 
feature.  It's not quite as simple as I infer he believes it to be, 
because there are some UI issues in terms of denoting the relevant DOM 
fragment, but in broad strokes I think that such a feature is eminently 
implementable, at least in Firefox.

This was relevant insofar as certain parties had expressed interest in 
such a feature in browsers as part of the SVG-in-HTML discussion.  But 
at no point was there an implication that MathML and SVG are in any way 
equivalent, other than in the obvious way: to work in a browser, both 
have a DOM that can then be serialized as XML.

> This comparison seems to be made for purely rhetorical goals, not for 
> technical reasons.

I didn't see a comparison.  I did see a statement of fact: a behavior 
that some constituencies want for SVG-in-HTML already exists for another 
XML vocabulary in some UAs (happens to be MathML and Firefox, but that's 
not actually all that relevant), and hence is presumably not too 
difficult to also implement for SVG, or indeed any other XML-based 
language that ends up in DOM form in said UAs.  There was the further 
conjecture that it should be easy to implement in other UAs as well, and 
while I happen to suspect that's true, there is no obvious evidence in 
support of this conjecture.

Received on Thursday, 31 July 2008 06:54:52 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:44:34 UTC