Re: [MathML 4] Add rules to map from non-combining to combining accents

On 09/03/2018 17:31, Frédéric Wang wrote:
> Hello,
> 
> Another idea for a future MathML spec.
> 
> MathML suggests to use non-combining characters for operators and most 
> of the operators in its dictionary are non-combining. However, TeX seems 
> to rely on combining characters for accents and hence in practice 
> OpenType fonts with a MATH table only provide size variants or glyph 
> assemblies for these combining characters ; and nothing for the 
> non-combining equivalents. That means non-combining characters are not 
> stretchable with these fonts and users may be tempted to use the 
> combining versions instead.
> 
> I think it would be good if MathML specifies how web engines can 
> fallback from non-combining to combining accents in order to stretch 
> accents. Ideally, we should have a list of such mappings (maybe data in 
> the XML Entity Definitions for Characters). For example to stretch 
> U+00AFMACRON, WebKit will also try and find size variants or glyph 
> assemblies associated to that character inthe OpenType MATH table and 
> otherwise fallback to data associated to U+0304 (COMBINING MACRON) or 
> U+0305 (COMBINING OVERLINE).
> 
> Having such a list will allow to improve consistencies between web 
> engines and to write WPT tests for that feature (currently it's 
> non-standard so only tested in WebKit repo).
> 
> WebKit bug: https://bugs.webkit.org/show_bug.cgi?id=159513
> 
> Mozilla bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1285277
> 
> -- 
> Frédéric Wang - frederic-wang.fr
> 


Yes we should say something (as it happens TeX also has difficulties 
with combining characters)

somewhere around

https://w3c.github.io/mathml/chapter7.html#chars.comb-chars

I guess is the place to add something.

As you hint I may need to add some extra data to unicode.xml to specify 
which characters are related in this way, I don't think the existing 
Unicode data reliably says which are equivalent combining/non combining 
forms although obviously taking the character name and deleting 
"Combining" gives a first approximation of the mapping.



David

Disclaimer

The Numerical Algorithms Group Ltd is a company registered in England and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses and malware, and may have been automatically archived by Mimecast Ltd, an innovator in Software as a Service (SaaS) for business.