Minutes: MathML intent meeting 6 Jan, 2022

 Attendees:

   - David Carlisle
   - Sam Dooley
   - David Farmer
   - Deyan Ginev
   - Patrick Ion
   - Louis Maher
   - Bruce Miller
   - Moritz Schubotz
   - Murray Sargent
   - Neil Soiffer
   - Steve Noble
   - Bert Bos
   - Paul Libbrecht
   - Cary Supalo

Regrets:

   - Stephen Watt

Announcements/updates

David: preliminary demo of content to presentation JavaScript (
https://w3c.github.io/mathml-docs/ctopintent/). the intent is to add intent
attributes but not done yet.
Progress on Intent (Neil demo)

NS, DC, and DG made progress over the holidays, but they are not done.
Should we rename Level 1 and Level 3?

NS: Most people were in favor of setting up two levels of intent. One level
would be called "core" and would be fixed names listed in the rec. The
second level would be called "open", and people could add items to this
level whenever they wished.
Should there be aliases for level 1 names?

NS: Should there be aliases in core?

NS: brought up several examples where things had aliases. One example might
the alias "log" for "logarithm"

DG: discussed the possible confusion between natural logarithm (called log)
and the common logarithm (called log base 10).

NS: When there are aliases, it is difficult for AT to know what name to
use. Aliases add program complexity.

MOS: What is the difference between core and open? He discussed a recent
failure with wikidata due to someone changing an entry in the open category
and this caused a functioning program to stop working.

NS: In the core level, every element should have an intent definition.

DF: Said that log in more advanced math meant natural logarithm and should
be pronounced log.

BM: People should not game the system. If you mean "a" you should speak
"a". He said that aliases would lead to confusion.

NS: If we have too many core level names, people will not implement them.
AT speech is not just a table lookup process. Speech may have to be
different for different audiences and languages.

BM: Different aliases correspond to speaking the same thing with different
words. This may not be the case in other languages.

MUS: The intent attribute would cover the case of speaking "squared" versus
"superscript 2".

DG: If the author wants to say natural logarithm, he should have a way to
force the speech to say that.

NS: In accessibility, blind users should have equivalent access whether it
be good or bad access.

PL: Does not see the use of aliases. Authors should be able to make the
system say what they mean.

DF: (a,b) should be read according to the intent of the author. (a,b) can
have several meanings such as the coordinate (a,b) or the interval (a,b).
Meanings must be guided by intent.

SD: said that we are discussing several questions at the same time. For
example, one symbol can have several meanings "(a,b)" and one thing can be
described in several ways with the same meaning (saying squared, or to the
power of two.).

MUS: This level of speech complexity may not be implemented for a long time.

BM: If we specify a subject area, this might help fill in defaults.

NS: asked DG: to open an issue on aliases. DG agreed.
Plan for unifying/disunifying level 1 name

(https://github.com/w3c/mathml/issues/254).

NS: reviewed an issue for future discussion: should we be unifying names or
breaking them apart. for example, how do we specify an open/closed
interval? Should there be 4 intents or should there by one and arguments
for open or closed interval. Should there be both a square-root and a
"root" intent or just one?

NS: Please feel free to add other issues at
https://github.com/w3c/mathml/issues