Re: Summary of MathML call meeting

On 15/03/2019 19:36, Neil Soiffer wrote:

> I'm not disagreeing with the "deprecated implies not in core", but the
> converse "not in core implies deprecated" is not true. My statement
> about duplicative elements/attrs was about the discussion of when
> something /can /be pulled from core. The msup example was meant to
> show that /can/ is not /should/.

I see.

> Probably not worth continuing this line, but just to be clear, we
> could define <msup> === <msubsup> <none/> sup </msubsup> and say that
> SubSuperscriptGapMin does not apply when only one non-none arg is
> given. This would remove four elements from core. In the early days of
> MathML, some people used msup, etc., as examples of MathML having too
> many tags and hence being too complex. However, I haven't heard you or
> other implementers complain about msup, mover, etc., being extra
> complexity.

I had opened https://github.com/mathml-refresh/mathml/issues/73 so that
we have some record of this discussion. But yes, basically reducing the
number of tags won't simplify the thing since you'd still have to have
special cases depending on which script is <none> + the usage is too
high to remove it, so I would just close the issue.

-- 
Frédéric Wang