Re: Stretchy equal sign for commutative diagrams

Thank you David and Daniel. My question was not really about how to 
implement the stretching but about the markup. The choice <mo 
stretch="true">=</mo> seems the most natural (and preserves the 
semantics of the diagram). It is used by LaTeXML and the MathJax 
extension. Since I don't see constructions for stretchy equal sign in 
Gecko or the STIX Open Type Math table, I wanted to be sure that 
everybody agree about that choice, and font designers and MathML 
implementers do the necessary to make the equal sign extendible.

On 25/03/2013 10:10, Daniel Marques wrote:
> I agree with David that using the equal sign has some implications. For
> example, the distance of the two horizontal lines might not be the desired
> one. Using the stretchy part of the double arrow or the box drawing have
> the extra advantage that the parallel vertical bars (and maybe diagonal)
> are also available and consistently drawn.
>
> Dani from WIRIS
>
>
> -----Original Message-----
> From: David Carlisle [mailto:davidc@nag.co.uk]
> Sent: sábado, 23 de marzo de 2013 17:57
> To: www-math@w3.org
> Subject: Re: Stretchy equal sign for commutative diagrams
>
> On 23/03/2013 15:02, Frédéric WANG wrote:
>> Hi all,
>>
>> We are implementing the AMScd extension in MathJax:
>>
>> https://github.com/mathjax/MathJax/issues/420
>>
>> and one horizontal stretchy operator that is needed is the character
>> for two parallel horizontal bars. We use the equal sign "=" and it
>> seems that LaTeXML does the same.
>>
>> I'm wondering if others have considered this operator and whether they
>> used a different Unicode code point for that. The equal sign is not
>> stretchy by default (that makes sense) and I don't see constructions
>> in the Open Type Math table of e.g. the STIX fonts.
>> This is not available in Gecko either. That's probably not hard to
>> implement (one could repeat the equal sign or use the middle glyphs
>> available for the "Double Right Arrow"), but I just want to coordinate
>> with other projects and keep things consistent.
>>
>
> I suppose U+2550 (from the box drawing block) is another possibility but
> using an equals or an arrow part is more likely to blend in with other
> arrows used I'd guess. AMScd itself I think doesn't use a font glyph at
> all for this but uses a rule (I'm not sure that's relevant here though, I
> think you'd want a character and stretchy attribute for MathML.
>
> David
>
>


-- 
Frédéric Wang
maths-informatique-jeux.com/blog/frederic

Received on Monday, 25 March 2013 09:30:01 UTC