Re: default for ma@symmetric

> All the boolean attributes other than symmetric default to false, so in
> the operator dictionary interpreting the lack of the property name as
> the default or as false comes to the same thing.

That's why I thought it made sense to have symmetric default to false as well.  But I'm happy to do it either way.

> For symmetric, now it's clarified as defaulting to true, as written an
> operator dictionary _could_ set it to false but the compressed notation
> used in appendix C can not express that and so in fact they are all true
> and "symmetric" in the table no effect other than a visual hint in the
> sample dictionary.

Right, that's how I read it now as well.  So "symmetric" in the properties column really makes no difference.  Got it.

> An alternative would be to say that an omitted boolean property name
> means false,

That seemed to be Neil's reading.

> but then we would have to check whether it needs adding
> to more entries eg the vertical arrows which are the majority of
> vertically stretchy characters for which currently symmetric isn't
> explicit in the table.

Right.  I wasn't sure what the reasoning behind the ones not marked symmetric was.

MathJax currently has mo@symmetric = "true" as the default, but I noticed it listed as false in the spec.  MathJax's values were set back in 2008 or 2009 from an earlier version of the draft, and I wonder if it was true then and was changed later?  I don't have any comment indicating why I would have changed the value from the one given in the spec.

In any case, thanks for clarifying the situation.

Davide

Received on Sunday, 9 February 2014 10:40:58 UTC