Minutes: MathML intent meeting 13 Jan, 2022


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


   - Cary Supalo
   - Patrick Ion

Continued thanks to Louis for taking notes.

NS: We need to start editing the MathML specification.
Progress on Intent (Neil demo)

NS showed his MathCAT demo where MathCAT means "Math Capable Assistive


NS: discussed a name having a prefix or postfix in its intent.

BM: Should there be defaults for the pronunciation of math operators like
plus and factorial?

BM: suggested that ways to speak superscripts should be in the dictionary.

NS: Asks people to report bugs in the demo.

NS: Discussed the order of execution of the rules in his example.

NS: Was showing examples using Clearspeak in his demo.

NS: wants to experiment with using subjects to guide intent expression
Should there be aliases for level 1 names? --

DG: We have a common problem of synonymous and near-synonymous names when
dealing with mathematical concepts. The same exact construct, often using
the same notation, can be narrated using different words, while understood
by professionals to have the same meaning. This makes it difficult to
maintain a list of Intent values where we have a single name for each
listed concept.

DG: If an author hears a term spoken in a manner he does not like, he must
go through the table of aliases to see if he can find a term that would be
pronounced in a way the author agrees with. This would be tedious.

DG: This problem may be increased if the author uses a different language
than is used by the system defaults. We will need a mechanism to deal with
different languages.

NS: How does an author find the alias that would allow an expression to be
spoken in a way that the author wants.

DG: The system should speak a term the way the author wants the term to be

NS: Does not know how to implement this open strategy.

DG: We should define a baseline behavior that authors can rely upon.

DC: We cannot ban the open part. We cannot ban aliases.

NS: Is common log an alias for log?

DG: The author knows what the author is writing. The author will provide
the proper meaning of the terms in their subject area.

DF: We want to provide intent that describes the expression. It should not
give pronunciation guidance.

We then had a discussion on how to speak "log" when "log" can mean "log
base 10" or "natural log". Acoustics expects that "log" means "log base 10"
while other subjects require "log" to mean "natural log".

MOS: When there was one AT, there was only a presentation layer and one way
to speak things. Now we want intent to have the ability to dictate what
words are spoken.

NS: You can force speech to say things with intent.

BM: It is up to the listener to distinguish natural log from common log.

BM: If you are giving the intent for a natural log say natural log and not
just log.

PL: Are we trying to converge on using site-based definitions?

BM: The pronunciation should depend on the listener and not on intent.

PL: Often a school would have a policy on how things should be pronounced.

DG: Error on the side of usability. If authors must look through a list of
aliases to determine how a term is to be described, then this would be a
large burden.

MOS: We are going too far. It might not be wise for us to use intent to
control speech.

BM: is not sure about the purpose of aliases. He does not think aliases are
the solution to guiding the system's pronunciation.

DG: If we discard aliases, people will argue about what the right word to
describe a term should be.

NS: completely agrees. Names are important.

Received on Wednesday, 19 January 2022 04:22:16 UTC