remove broader mathml discussion

Author: dginev

gap-analysis/index.md

@@ -468,45 +468,6 @@ For example, a publisher might add a subject area to all the math tags in a geom
 
 The Math WG has not had detailed discussions about using subject areas yet. If the idea works out, it is likely that at least initially, the number of known subject areas would be limited to perhaps as few as 5–15 subject areas covering basic math and science areas. The number of changes to defaults each subject area would is very dependent on the subject area. As with `intent`, `subject` area` can be implemented without changes to other web standards other than needing to be part of the accessibility tree. At least one AT tool (MathPlayer) makes use of user-specified subject areas to override defaults on what is spoken, so there is a little experience with this concept. Classifying mathematics and other sciences is difficult. It is unknown if a broad brush categorization approach to K-14 topics as currently envisioned is feasible; it remains to be investigated.
 
-## Other target applications of MathML
-
-MathML is widely used by assistive technology (AT) to generate speech, braille, and allow navigation of math. It is also accepted and generated by a significant number of computation systems. A number of specialized programs have been developed to search for mathematical expressions that take advantage of MathML, but none of the major search engines make use of MathML.
-
-### Search
-As discussed above, in some situations a very literal vocalization of the displayed formula can be concise and understandable, while in others, or for other users, a more semantically oriented reading is more useful. In the same way, how users will search for mathematical information may be closely tied to the usual way of writing it, or more tied to its semantics, and thus affected by ambiguities. The task of indexing and search of mathematical formulas must therefore deal with some of the same issues as accessibility.
-
-* Examples of what someone might search for
-* Pain points -- when is ambiguity a problem?
-* Ecosystem pain points -- what is missing for major search engine vendors to provide a high-quality search for math syntax queries? Is there a good query language? Is there a need not being met from students/educators/researchers? Example or two.
-
-### Structured data via schema.org RDFa annotations
-Schema.org is an ongoing effort developing vocabularies for aiding “Rich Results” in information retrieval, endorsed by most major search engines. We did a basic evaluation, with the goal of exposing our rich “intent” information to search engines.
-
-* There are no targeted vocabulary entries for the kind of information we want to enrich in presentation MathML. Loosely related entries were “disambiguatingDescription” and “name”. A prototype of how they could be added to MathML 3 via RDFa can be seen [here](https://github.com/dginev/tiny-mathml-a11y-demo/blob/master/schema_experiment_2.html)
-* There are vocabulary items explicitly referring to mathematics, which are specific to applications - “mathSolver”, “solveMathAction”, “mathExpression”, and can not be repurposed as augmentations over MathML.
-    * These developments are [recent](https://github.com/schemaorg/schemaorg/issues/2740), and there is some possibility we could request new vocabulary entries for MathML-specific search needs.
-* There was no viable outcome in using the existing schema.org vocabulary. Pages with well-indexed structured data did not benefit - querying for the new information we introduced in the annotations returned no results. It is likely that while this data can be accessed by web crawlers, it may not be added to search engines before creating an explicit “Rich result” application.
-    * This leads us to believe this avenue will only be viable in close collaboration with browser vendors, even if existing vocabulary entries can be repurposed for our needs.
-* The theoretical upside of adopting a structured data annotation approach is to avoid introducing additional complexity inside MathML itself, and delegate all search-related markup to existing technologies. However, it is currently unclear if there is any path to reusing these approaches over math expressions, especially when considering adoption in major search engines.
-
-
-### Discoverability
-
-
-To aid in Search Engines' ability to find pages containing MathML, these pages with MathML should be included in the webpage’s Header Schema.org accessibility metadata.  Currently there is one property which would be appropriate for this which is ‘accessibilityFeature’.  Here is a current [list of the values for accessibilityFeature](https://www.w3.org/wiki/WebSchemas/Accessibility) of which “MathML” and “describedMath” would be appropriate to use.  There are plans to move this work from a W3C wiki page into something more formal, where new values are voted upon by the community but since accessibility metadata is broader than digital publishing it hasn’t yet been decided which group will take on this work.  When this group either gets created or this work falls into another W3C group any new values the MathML group would like them to consider adding can be done at a future meeting.
-
-### Editing/Selecting Mathematical (Sub)Expressions
-Applications to edit mathematics need to have a notion of a insertion cursor and a selection. This information may sometimes be useful to preserve during copy/paste, so some additional attributes may want to be added. See [MathML and OMML User Selection Attributes \| Math in Office (microsoft.com)](https://devblogs.microsoft.com/math-in-office/mathml-and-omml-user-selection-attributes/). Note that the insertion cursor or selection may want to be displayed on a refreshable braille display (typically by using dots 7 and 8).
-
-### Conversion, Computation and Transfer
-Mathematical formulæ are often the subject of more than just reading and writing.
-Formulæ are, in written form, the lever used to request a computation: the resolution of an equation or evaluating an expression made of numbers can easily be made by computing engines which sometimes understand presentation-MathML (but are limited by ambiguities and context): An enrichment such as intents can lower these ambiguities.
-Closely connected to computability, the use of formulæ input in a test to assess mathematical competencies can benefit from lowered ambiguities.
-
-
-Finally, formulæ displayed in a web-page are subject to being transferred to other places where further work can be done applying mechanisms such as copy-and-paste or drag-and-drop. Since MathML can be the content of the transfer (e.g. using the MathML clipboard flavours), receiving applications can benefit of a standardised way of lowering ambiguities.
-
-
 ## Authors/Acknowledgements