Re: Minutes: MathML Intent meeting: 10 Dec, 2020

Hi Neil,

no worries. Everything is all right. I just wanted to avoid confusion.
Please find my slides attached.

Moritz

On Wed, Dec 16, 2020 at 10:22 PM Neil Soiffer <soiffer@alum.mit.edu> wrote:
>
> My apologies if I conveyed the idea that you endorsed schema.org as what we should do. It is simply an alternative approach which you have some knowledge about. I and I'm sure others appreciate your willingness to inform the CG about it so that we can discuss the pros and cons of it.
>
>     Neil
>
>
> On Wed, Dec 16, 2020 at 4:18 AM Moritz Schubotz <moritz@schubotz.org> wrote:
>>
>> Dear Neil, Dear Deyan, dear all,
>>
>> while I am happy to present existing RDFa/microdata technologies later
>> this week, I would like to use the chance to clarify my intentions
>> since I do not want to act as an ambassador for anything else as
>> content MathML.
>>
>> I think the list Deyan proposed is already now a great achievement. If
>> each content MathML element has a generally accepted default
>> presentation element, we already have achieved a great way in the task
>> of inferring content from presentation. However, default also implies
>> that there needs to be a mechanism to specify, this is not the default
>> case. How the orthogonal concept of intent attributes helps with this
>> job is not clear to me.
>>
>> In my little presentation, I will be covering this scope
>> https://schema.org/docs/gs.html
>>
>> With some additions to the link of schema.org and Wikidata, see
>> https://github.com/schemaorg/schemaorg/issues/1186
>> and
>> https://www.wikidata.org/wiki/Wikidata:Schema.org
>>
>> Deyan, maybe you can add your points afterwards?
>>
>> Moritz
>> https://schubotz.org | +49 1578 047 1397
>>
>> On Thu, Dec 10, 2020 at 9:25 PM Neil Soiffer <soiffer@alum.mit.edu> wrote:
>> >
>> > Notes thanks to Louis Maher. The discussion was detailed and involved screen sharing-- it is best to listen to/view the recording:
>> >
>> > https://us02web.zoom.us/rec/share/QEmF8jMfuBGT_1HID63xfgRDXOfPI6Hz17p0MGG4YupRUZeTfxTGvfStH2ZjZOyD.TzJ0yQeYPDJaKGYd
>> >
>> > Access Passcode: W56%ko#P
>> >
>> >
>> > Attendees:
>> >
>> > Neil Soiffer
>> >
>> > Louis Maher
>> >
>> > Bruce Miller
>> >
>> > David Carlisle
>> >
>> > Deyan Ginev
>> >
>> > Steve Noble
>> >
>> > David Farmer
>> >
>> > Sam Dooley
>> >
>> > Murray Sargent
>> >
>> >
>> >
>> > Agenda:
>> >
>> > 1. Any updates?
>> >
>> > NS: next week, Moritz will give a presentation on schema.org.
>> >
>> >
>> > SD: there will be an aria deep dive for Braille next week.
>> >
>> >
>> > DG: has completed his table (Level 3 sheet), 964 entries at this stage, as *preliminary* names with stable notations in encyclopedic resources:
>> >
>> > https://docs.google.com/spreadsheets/d/1EsWou1K5nxBdLPvQapdoA9h-s8lg_qjn8fJH64g9izQ/edit?usp=sharing
>> >
>> >
>> > ...Taking all of levels 1,2 and 3 we now have more than 1,000 “names”.
>> >
>> >
>> > 2. The group will (once again) attempt to finish an update on Sam's ideas for explicit and default values for intent (esp for mtable):  See
>> >
>> > https://mathml-refresh.github.io/mathml/docs/intent.html
>> >
>> >
>> > SD: talked about @append. Acts a bit like an unevaluated mrow.
>> >
>> > … some don’t like that
>> >
>> >
>> > SD: discussed bound variables.
>> >
>> >
>> >
>> > DG: discussed mover versus msup.
>> >
>> >
>> > SD: discussed equation arrays using tables. Mtd acts like an mrow.
>> >
>> >
>> > SD: discussed breaking an equation over multiple rows. @append and @extend. This is an example where you can’t eval the @append right away because the next row might extend it.
>> >
>> >
>> > Discussion/unease of unevaluated mrows came again. Really just want to concatenate the two rows together. Algorithm could use look ahead and not finish the append.
>> >
>> > NS: what about an idea analogous to rowspan/colspan where the first mtr says how many rows belong to it? No idea on what syntax for that is like.
>> >
>> > DG: Really mainly want to throw away table wrappers and glom stuff together. Maybe we should be thinking about a “next” feature. Maybe left/right/up/down/diagonal dir to say where to get the next element.
>> >
>> > NS: Need some way to stay “done”.
>> >
>> > DG: Almost always want natural top/down left/right reading order, so maybe start/stop is what is needed with default being reading order.
>> >
>> >
>> > SD: anything other people want clarified about my proposal?
>> >
>> > BM: still not clear on ‘@’ syntax.
>> >
>> > NS: would like to eliminate $named variables because you have to dive into the intent to find out what the $named variable means.
>> >
>> > DG: I have a couple of “awkward” examples. One is a subsup, where the sup is a ‘prime’. So one structure for deriv(index($1,$2),$3). See “awkward nesting (1)” in https://dginev.github.io/tiny-mathml-a11y-demo. The other is “awkward nesting(2)” -- an msub of an mover representing “an x with a bar over it”. In this case, the structure gets flipped inside out.
>> >
>> > NS: interesting cases, but neither require looking inside an intent.
>> >
>> >
>> > Challenge to everyone -- find examples where there is not a simple work around (as in the case for nested integrals) to diving inside of an ‘intent’ to reference a child.
>> >
>> >
>> > Action Item: DG and NS will flush out their suggestions to extend SD’s document.
>> >
>> >
>> > NS: next meeting will be Dec 17 where Moritz will talk about schema.org and RDFa/microdata approaches.
>> >
>> > NS: Then a break for the holidays followed by a meeting January 7, 2021. NS and DG to present their ideas then.
>> >
>> >
>> >