Minutes: MathML Full meeting, 28 March, 2024

 Attendees:

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

<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0-regrets>
Regrets

   - Patrick Ion

<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0-action-items>Action
Items

NS reminded DC, MoS, and himself that they have a paper to write.

DG: should close issue 489 <https://github.com/w3c/mathml/issues/489>
entitled "Update minsize/maxsize percentage values to match MathML Core",
as completed.

There was no resolution on core issue 142
<https://github.com/w3c/mathml-core/issues/142> Entitled "Should all MathML
elements really be potential hyperlinks".

PL: Does MathML have a safe list core issue 227
<https://github.com/w3c/mathml-core/issues/227>

*ACTION* DG will put in a link to the database of security flaws.

*ACTION* NS asked PL to find the link to the ISO official list of
currencies and create an issue for currency. And NS will populate it with
what he found from math player, which NS thinks comes from the Unicode
list, and we'll compare those. Okay.

PL: The ISO standards are behind a pay-wall, and I do not have access to it.

NS to PL: Can you look to see if somebody has published the ISO list in
some other form?

PL: Yes.

From David Carlisle to Everyone: https://en.wikipedia.org/wiki/ISO_4217

*ACTION* NS: revisit degrees and see if it needs a concept. K C and F may
need to be made concepts or properties.

*ACTION* NS: that it is an open question what to do with the directions
north, south, east, and west, but they should not be units if we're using
properties.

*ACTION* NS: That the units discussion wherever it ends up being should
have some mention that division, and raising to a negative power, the
equivalent of a division, should consider using the word per.

*ACTION* DG: should look up Appendix B: table of 24 unit conversions, and
Appendix C: physical constants.

*ACTION* NS: fermi fm needs to be added to the list.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0-agenda>
Agenda
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0-1-announcements-updates-progress-reports>1.
Announcements/Updates/Progress reports

NS: DG passed me some data from Archive. We were just looking at various
things, and just for assisted technology, I have a style of speech that I
developed, probably 20 years ago, called Simple speak, and I've never
written up the basic idea that if you have something that's simple, such as
a fraction A over B, you just say over B, and you don't bother putting
beginning and end words around it. But if you had a over b plus one and the
B plus one was in the denominator, Then you add the words fraction A over B
plus one end-fraction, so add some extra words. But at least it's unique.
So, the data that DG came up with basically said, around 70% of the time
you can use the shortened form which made me feel good, because that means
that my idea actually is useful, and now I have some data to back it up.

PL: I saw the other day a mathematical statement of some sort, and it was
linking to the archive page. And why not directly link to the archive HTML
page, and indeed, this is what they did. So, we have direct links to our
active papers that can be read by just looking at them, and so this is an
argument to convert the entire thing to MathML.

PL: We had a few discussions in the paper about citing literature on what
is AT doing, what is its role, and what is a typical user's expectations.

PL has not found a paper that gives this type of information.

MoS: We have now the new MathML rendering mode enabled in all Wikipedia
languages, so users which have an account can now sign in and try it out.
There has not been any feedback up to now.

NS: It's always frustrating when you have some great new accomplishment,
and then there's dead silence. That's either a sign that people think it's
great, and there's no issues, or that people don't care. I'm never sure
which.

DC:
From David Carlisle to Everyone:
https://github.com/latex3/tagging-project/discussions/72

There's a project discussion page mostly for tagged Pdf. Some of them are
mathematical. There's quite a lot of MathML within the pdf for people
interested in MathML that's not in a web page. There's some documents you
might recognize that don't have any math at all, just to show they think
isn't just math.

NS: We have done the things we need to do for the horizontal review, so
hopefully it will move forward by the end of April.

DG: Asked for group agreement to close issue 489
<https://github.com/w3c/mathml/issues/489> entitled "Update minsize/maxsize
percentage values to match MathML Core ", as completed.

MathML core will match what MathML 3 does, and the full spec needs no
changes. We will have a chromium patch to make it act like Firefox has for
years, and use the unstretched size for the percentage values.

*ACTION* From Deyan Ginev to Everyone: From
https://github.com/w3c/mathml-core/issues/103 entitled "percentage of
minsize/maxsize"

Chromium (dev + patch)

See
https://chromium-review.googlesource.com/c/chromium/src/+/5402478/1/third_party/blink/renderer/core/layout/mathml/math_operator_layout_algorithm.cc

DG will close issue 489.

NS: There was a discussion on links to whether we're going to change MathML
4 to HTML5 to take the ability to link out of every element.

DC: MathML 3 allows href on everything. This causes more problems than
you'd expect in a web browser where you think linking wasn't that
difficult? But anyway, MathML 4 Core doesn't allow linking at all. You have
to embed a HTML an element inside a token element which means you can't put
a link around one plus 2, for example. We have pushed back on this.

There is a suggestion to allow an (a) element, or href on an mrow.

DC: There's some restrictions on elements with Link attribute property in
shadow Dom.

DC: There are restrictions on elements with href, see MathML core issue 142
<https://github.com/w3c/mathml-core/issues/142> entitled" Should all MathML
elements really be potential hyperlinks".

NS: Do we want to take a stand on this issue?

DC: BK is just saying that it's difficult to add things to shadow Dom if
they've got link behavior.

DG: svg and HTML have an (a) element, and we are expected to follow the
precedent.

BM: Whatever works. He leans towards mrow. What happens in the transition
period with MathML with (a) in it. What would the html parser do with that?
The parser will do something random. If we use a new element or (a), at
what point can you generate MathML in browsers that have not implemented
that behavior.

DC: The (a) will not upset the parser. The parser will make it an mrow.

BM: This is a reasonable approach.

DC: Adding (a) to MathML would be safe.

DC: I expect that if we push back (a) we'll get nothing.

NS: (a) does not break the parser, but it does not do anything with Chrome.

DC wants to double check this. It shows up under Firefox, but not chrome.

NS: We prefer an attribute on mrow but can accept the (a).

NS: Firefox and Chrome seem to behave differently with a ?? On the other
hand, Firefox works with an HREF or link on the mrow.

DG: We prefer having the capability for links for non-leaf elements. We
want complex structures to be linkable.

There was no resolution on core issue 142
<https://github.com/w3c/mathml-core/issues/142> entitled" Should all MathML
elements really be potential hyperlinks". .

PL: Does MathML have a safe list core issue 227
<https://github.com/w3c/mathml-core/issues/227> entitled "{ MathML support
in the HTML Sanitizer API"? Do we want to spend time on this today?

PL: Do we have a safe list, a list of safe elements that a sanitizer could
use? The safe list will say which elements will be safe to copy.

PL: My question is, can you name elements and attributes to be possibly
rubbed out?

NS: Getting rid of links might be safer. Is there anything else that might
make MathML unsafe in any way?

NS: Well, embedding HTML in MathML is unsafe.

PL: So is putting SVG inside of MathML unsafe.

PL: Potentially, if you copy TeX, then a macro could be doing something
funky with your computer.

PL: There was this feeling that annotation should go away. But I kind of
disagreed because annotations really are useful.

DG: Discussed exploited issues.

DG: There are historical precedents of this. If you look at the
vulnerability exploits that MathML has been involved with historically,
these come from the parser changing namespaces.

DG: The issue is that the notation Xml may need to be parsed as a tree, and
in that process you open a can of worms for all kinds of exploits.

DG: They want a safelist to screen the parser.

PL: The core meeting did not discuss this.

*ACTION* DG will put in a link to the database of security flaws.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0-2-units->2.
Units:
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0-potentially-a-full-list-of-units-with-all-their-prefixes-as-intent-concept-names>Potentially
a full list of units with all their prefixes as intent concept names

NS: Discuss next meeting.

NS: We are thinking about if we have a property name or concept, or both.

NS: I do not think having both makes any sense.

NS: Let us look up the units on the other lists and see if they are
mentioned on my MathPlayer list. First, let us compare issue 475 (from
MathPlayer) with DG's physics list.
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0--a-href-https-github-com-w3c-mathml-issues-475-list-of-units-475-a->List
of Units (#475) <https://github.com/w3c/mathml/issues/475>
<https://sandbox.cryptpad.info/code/inner.html?ver=2024.3.0-13#cp-md-0--a-href-https-gist-github-com-dginev-825078ae316c32c312436f42061b3d05-deyan-s-physics-list-a->Deyan’s
Physics list
<https://gist.github.com/dginev/825078ae316c32c312436f42061b3d05>

We covered: foot ft pound lbs radian rad meter m kilometer km centimeter cm
hour hr, h year yr minute min , ′ second s, ′′ dollar $

NS: did not put in an entry for dollars.

*ACTION* NS asked PL to find the link to the ISO official list of
currencies and create an issue for currency. And NS will populate it with
what he found from math player, which NS thinks comes from the Unicode
list, and we'll compare those.

PL: The ISO standards are behind a paywall and I do not have access to it.

NS to PL: Can you look to see if somebody has published the ISO list in
some other form?

PL: Yes.

From David Carlisle to Everyone: https://en.wikipedia.org/wiki/ISO_4217

NS: We considered degree ° Kelvin K Celsius C Fahrenheit F

NS: are K, C, and F units?

NS: From the context you would know what degrees means.

PL: Degrees can be for angles.

NS: When you say "degrees" the word is ambiguous because you do not know
what kind of degrees it is.

*ACTION* NS: revisit degrees and see if it needs a concept. K C and F may
need to be made concepts or properties.

NS: directions: east E, west W, north N, south S, and their combination s
like NE for northeast.

From Deyan Ginev to Everyone:
https://en.wikipedia.org/wiki/Points_of_the_compass#Compass_point_names

east E west W north N south S

NS: If we go with concepts for units, we go with concepts for these, and if
we use properties for units, then these have properties of direction.

BM: While they are not units, they are often combined with units like 90
degrees west.

DG: N can be north or newton.

BM: Maybe authors should say what they mean.

DC: Do you have to say all the ways these things can be combined like north
by northwest?

NS to DF: have you encountered this in pretext? No.

*ACTION* NS: that it is an open question what to do with the directions
north, south, east, and west, but they should not be units if we're using
properties.

NS: What do we do with the word "per"?

DG: Instead of "per" you can say -1 power or slash.

NS to DG: What do you want to do with this?

DG: I was worried that it would be missed with a property unit. It's
possible some implementers will not notice raising to the minus first
power, which could be spoken as per.

DG: The reason I added it is to somehow raise awareness documented
somewhere that this is part of the units parsing coverage.

*ACTION* NS: That the units discussion wherever it ends up being should
have some mention that division and raising to a negative power the
equivalent of a division should consider using the word per.

*ACTION* DG: should look up Appendix B: table of 24 unit conversions, and
Appendix C: physical constants.

NS: What do we do with constants like the speed of light "c"? Typically,
you call it "c".

NS: Planck's constant has its own Unicode symbol "h bar".

NS: So, do we need to say anything about these constants?

NS: Does anybody remember a constant that you don't say the letter? But you
say the name for the letter in typical speech?

NS: They are variables and we do not need to do anything.

*ACTION* NS: fermi fm needs to be added to the list.

We completed DG's physics list.

NS: Europe is on summertime next week.