Re: Semantic Markup: mrow's are needed

On 26/06/2020 16:39, Neil Soiffer wrote:
>
> More directly actionable though is what we should say in a MathML 4
> specification. Specifically,
>
>   * Should we discourage or maybe even deprecate implicit mrows?
>
For MathML Core, mpadded and msqrt are currently the two remaining
elements that relies on the mrow layout for part of their layout. It
would be more convenient to force explicit mrow, but it seems that would
be a serious backward compatibility issue. In Chromium, we plan to
implement these using "anonymous mrow" (well-defined css concept), so
maybe that's what we want to say in the core spec too.

In any case, that does not really help your semantic use case. But if
you decide to deprecate them in MathML full, we should also warn about
this in MathML Core as otherwise people will keep using it.

Frédéric Wang