Re: [MathOnWeb] call for comments -- directions for 2018

Hi Liam,

As Arno, I'm also not arguing that MathML isn't useful outside of the
browser rendering.

But I want to comment on a few things you mentioned.

> I don't agree that this obsoletes mathml - especially the semantic
> markup - since done well it could actually *enable* the use of mathml.

I'm guessing with "enable the use of MathML" you mean that the use of
Content MathML could enable the native implementations of Presentation
MathML.

While that might theoretically be true, I see no indication that this could
happen. In particular, I don't see how the use of Content MathML would
reduce the problem of Presentation MathML being at odds with the rest of
the web platform.

But more importantly...

> You might as well say that having display:block and size:larger
> obsoletes the h1 element, but that'd be ignoring the needs of search
> engines and content management systems.

I find this comparison lacking because the h1 element is actually in use.

After 20 years, there's no significant corpus of Content MathML; people are
either unwilling or unable to create it. There's also not a single notable
application of Content MathML. In particular, search engines around MathML
remain vaporware (both academic and commercial).

I'm very tired of promises that if only people tried, one day MathML will
solve all the problems.

Best,
Peter.

PS: If you wrote this because you're building an application around Content
MathML, then I apologize; it would be great to hear more about it.




2018-01-15 20:01 GMT+01:00 Liam R. E. Quin <liam@w3.org>:

> On Mon, 2018-01-15 at 18:34 +0000, Arno Gourdol wrote:
> > On Mon, Jan 15, 2018 at 5:57 PM, Liam R. E. Quin <liam@w3.org> wrote:
> >
> > > On Mon, 2018-01-15 at 13:44 +0000, Arno Gourdol wrote:
> > > > Personally, I don't think that MathML is the solution. I would
> > > > rather
> > > > see CSS and ARIA improved. This would be a less significant
> > > > effort
> > > > from a standard and implementation point of view, while providing
> > > > a
> > > > more flexible solutions.
> > > >
> > > > Specifically, I would like to see support in CSS for stretchable
> > > > fences and notations, features which are currently
> > > > difficult/impossible to implement well.
> > >
> > > I agree that building mathematical formatting on CSS (and extending
> > > CSS
> > > where needed) makes sense in the Web platform.
> > >
> > > I don't agree that this obsoletes mathml - especially the semantic
> > > markup - since done well it could actually *enable* the use of
> > > mathml.
> > > You might as well say that having display:block and size:larger
> > > obsoletes the h1 element, but that'd be ignoring the needs of
> > > search
> > > engines and content management systems.
> > >
> >
> > I should have been clearer since you seem to disagree with something
> > I didn't say. I'm not arguing that MathML should be obsoleted.
>
> oops sorry!
>
> > I agree that MathML could have a role to play as a machine-readable
> > format.
> > However, that doesn't mean that browsers should have anything to do
> > with displaying MathML. There are a number of solutions today that
> > display MathML in browsers, even if those browsers don't support
> > MathML. Browsers do not need to support MathML for MathML to be
> > useful.
>
> Well, that's for sure true although it would be helpful for adoption if
> they did!
>
> Thanks for explaining what you meant. I thnk we agree that browsers
> need to support the necessary building blocks, though, and that w'd
> like them to do that through CSS rather than hard-wiring behaviour to
> specific element names, although we may diverge on the importance of
> having those element names also shared by being e.g. part of the user
> agent stylesheet.
> >
> --
> Liam Quin, W3C, http://www.w3.org/People/Quin/
> Staff contact for Verifiable Claims WG, SVG WG, XQuery WG
>
> Web slave for http://www.fromoldbooks.org/
>