- From: Neil Soiffer <soiffer@alum.mit.edu>
- Date: Wed, 18 Dec 2024 14:20:32 -0800
- To: "www-math@w3.org" <www-math@w3.org>
- Message-ID: <CAESRWkCoy1xHDBqVZQuoAoRGH+CA0SLLoeKm-5b9HOL_koUhvQ@mail.gmail.com>
Attendees: - Neil Soiffer - Louis Maher - Deyan Ginev - Bruce Miller - Moritz Schubotz - Paul Libbrecht - Murray Sargent - Bert Bos - Stella Vouthsina <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-regrets> Regrets - David Carlisle - Patrick Ion <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-action-items>Action Items Our last meeting of the year will occur on Thursday, December 19, 2024. <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-2-a-href-https-github-com-w3c-mathml-issues-181-181-mathml-4-extensions-for-alignment-and-possible-deprecation-of-maligngroup-and-malignmark-a->2. #181: MathML 4 extensions for alignment and possible deprecation of (maligngroup/) and (malignmark/) <https://github.com/w3c/mathml/issues/181> *ACTION* MuS will try to shorten the descriptions of maligngroup and malignmark. He needs to find references to the alternating ampersand method. <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-3-a-href-https-github-com-w3c-mathml-issues-449-449-intent-properties-ordering-amp-references-a->3. #449: Intent Properties: ordering & references <https://github.com/w3c/mathml/issues/449> (lots of new discussion) *ACTION* NS: I think we have agreement on what should happen. We need to generate language to describe our agreement. <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-4-a-href-https-github-com-w3c-mathml-issues-433-433-comment-include-or-not-a-sample-set-of-default-conversion-from-plain-mathml-to-mathml-with-intent-a->4. #433 (comment) include (or not) a sample-set of default conversion from plain-MathML to MathML-with-intent <https://github.com/w3c/mathml/issues/433> (this is a shortened version of this doc <https://github.com/w3c/mathml-docs/blob/main/minimal-intent-core.md>) *ACTION* NS wants people to comment on his literal (formerly called structure) list shown by expanding the details button. Then NS wants to consider the common category of defaults. <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-agenda> Agenda <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-1-announcements-updates-progress-reports>1. Announcements/Updates/Progress reports Our last meeting of the year will occur on Thursday, December 19, 2024. Stella Vouthsina (SV), from Google, has joined our group. We introduced ourselves to Stella. PL: Stella knows Greek and Spanish. NS: MathCAT works in Spanish. <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-1-2-a-href-https-github-com-w3c-mathml-issues-181-181-mathml-4-extensions-for-alignment-and-possible-deprecation-of-maligngroup-and-malignmark-a->2. #181: MathML 4 extensions for alignment and possible deprecation of (maligngroup/) and (malignmark/) <https://github.com/w3c/mathml/issues/181> (updates?) MuS was going to rewrite this section. He wanted to describe how it worked in the Microsoft applications. He could not find references, to the use of the alternating ampersand in the LaTeX community, that he needed. He did find references to aligning equal signs by this method, and this is the most trivial case. DG came up with the idea of aligning columns by using padding. MuS knows that this alignment application is aimed at browsers and not Microsoft applications. NS to MuS: So, are you able to take advantage of intense properties on tables so that essentially the columns go away when it's read by the screen reader? MuS: Well, in my speech engine, it is not doing this correctly. It reads equation arrays without saying where the align marks are. I'm working on that. *ACTION* MuS will try to shorten the descriptions of maligngroup and malignmark. He needs to find references to the alternating ampersand method. <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-0-3-a-href-https-github-com-w3c-mathml-issues-449-449-intent-properties-ordering-amp-references-a-lots-of-new-discussion->3. #449: Intent Properties: ordering & references <https://github.com/w3c/mathml/issues/449> (lots of new discussion) NS: Last week we decided that the last property on the list should take priority. NS: There was the issue of the multiple conflicting properties, and we did agree last time. The last one wins. NS: When you have conflicting properties, does the one on inside win or the one on the outside win? Clearly the one on the outside is going to win if it wipes out the inner ones. NS: There's this issue where you have a reference that essentially brings the reference back up to the top level. DG: So, if we have a clear order and having the last one win, that's where the last meeting ended. DG: It gets a bit more confusing with the up and down axis. So, if a reference has a property we need to dereference. And the question is, what is the algorithm to dereference? If there are other properties adjacent to the dereferenced property, does the dereference property override the other adjacent properties? DG: The outer property overwrites the inner properties. NS: There is another issue that when you convert a property to speech and bring it up, do you just have a symbol, or do you have the concept name. BM: If you have a minus sign, if you say unary minus or negative is decided by the AT. It is a reader preference and not fundamental. NS: If you have vertical bars, do you have a determinant or cardinality? This is determined by intent. If the author wants the vertical bars to be disambiguated, they are going to have to give a concept somewhere. NS If we do the reference and the reference does not have an intent on it, then whatever happens happens. DG: Intent is a partial annotation attribute over presentation MathML trees Which means many of the nodes may not have it. DG: We need to have both situations addressed where nodes may or may not have the intent. DG: If you want the entire reference resolution to be inside the spec of the intent attribute, every node that has an arg attribute has to have an intent attribute that is kind of self-sufficient. BM: Most of the discussion is with the assumption that when you have a reference, mostly what you do is find the intent that's on the referred-to node. Pull it up and substitute it in. DG: You take the reference up and you merge down all properties associated with it at the point of referencing. You take the prefix property, and you merge it down at the nodes which contains the arg. We have to decide if you merge it at the front or the back position. I suspect we've agreed it's the back position, and then you can instantiate that node recursively as necessary. NS: We need to figure out some language to make it clear that's the intent of how intent works. BM: A pure reference doesn't have any properties or arguments. So, if you have an intent where there's a reference that does have properties and arguments, you could remove those properties and arguments and append them to the reference node, and then the intent at the upper levels has the reference. *ACTION* NS: I think we have agreement on what should happen. We need to generate language to describe our agreement. <https://sandbox.cryptpad.info/code/inner.html?ver=2024.12.0-0#cp-md-1-4-a-href-https-github-com-w3c-mathml-issues-433-433-comment-include-or-not-a-sample-set-of-default-conversion-from-plain-mathml-to-mathml-with-intent-a->4. #433 (comment) include (or not) a sample-set of default conversion from plain-MathML to MathML-with-intent <https://github.com/w3c/mathml/issues/433> (this is a shortened version of this doc <https://github.com/w3c/mathml-docs/blob/main/minimal-intent-core.md>) NS: This issue considers the situation when converting from plain MathML to MathML with intent. Ns believes there should be fallbacks, essentially having something predefined. There are three possibilities that he outlined for a property. One is what we originally called structure, but now call literal. Another is what he called common, and the last one was called legacy. NS: So, the only thing we've agreed on so far is to have something called literal, And so we need to have some notion of what literal does. NS: By expanding the "details" button, NS reviewed the "details" of the items he covered as literals. He invited individuals to make comments on these items. Items that promoted discussion were: mfrac which could be a fraction or something else like the binomial coefficient. You could say argument one over argument two. MuS and DG: You could use character navigation and just describe what is there. DG: To give an example to Murray, MFRAC with line thickness equals zero is very often used for a binomial coefficient with the coefficients above each other. So that's not a fraction and it's not a denominator. You would need to indicate when a term begins and ends; that is, you must indicate when a numerator or denominator begins and ends. BM: Literal is exactly for the case where there is no intent. NS: Menclose should indicate the notation attributes along with the contents. so, you could say a box around such and such or a diagonal strike over the following contents. NS: You would never say Msup as a power. It would just be spoken as a superscript, but I think that you want to have exceptions for pseudo script characters like a prime or a degree sign. *ACTION* NS wants people to comment on his literal (formerly called structure) list shown by expanding the details button. Then NS wants to consider the common category of defaults. DG: When in doubt, you should just read what is there and not interpret it, for there is always another special case.
Received on Wednesday, 18 December 2024 22:20:50 UTC