Meeting minutes
Further improvements in principles document
<AvneeshSingh> https://
Avneesh: we have been making many edits to the document with implementing feedback from multiple people.
… Any other improvements?
George: will add the word "claims to meet" instead of just "meets xyz" in the Conformance section.
Charles: agrees.
Charles: I have multiple editorial issues, but was thinking MathML should be called out in the Charts Graphs and Formula section.
Avneesh: if you have more feedback please include it in Github.
What approach should we use for techniques documents?
<AvneeshSingh> w3c/
Avneesh: we need to start working on the techniques document as the Principals document is getting close to being finalized.
… there are two techniques: suggested, XLST advantage implementers can just use it, or the other approach is to use Pseudo code. makes it more readable and you can have multiple implementations like JS XLST, etc.
… How should we proceed?
George: do you know if Gregorio make the pseudo code from the xlst file himself.
JonasLillqvist: I comment on the issue in favor of pseudo code. programming language agnostic.
gautierchomel: Last version of the techniques I don't see any pseudo code. I am not sure. We need to written form of what we want. This is what we should first offer.
… Then we can include examples, but we need to make sure we don't get overloaded. I think 2 voted for pseudo code.
… I still like XLST, we could have an implementation section.
… , showing what triggers in the example one XLST doing everything in the document. could be beneficial.
<AvneeshSingh> madel
Madeleine: No code just examples of the metadata itself. We may need to check multiple sources of metadata so implementation examples ie. schema.org and ONIX, and could be MARC. We will need multiple XLST for each of these implementations.
… pseudo code is more readable and what others have said but coders would love the XLST which they can just then grab and use. can this group maintain this and make sure it is bug free. Do we have the capacity to do this. If we have to prioritize one I would go with pseudo code.
… XSLT should be a stretch goal.
Avneesh: maintenance can be a big challenge.
… Pseudo code is first preference but XSLT could be an advantage and could maybe use XSLT to create the pseudo code via ChatGPT potentially.
Approach for supporting localization of descriptive sentences which retailers and distributors will display.
George: We have many strings that explain the accessibility of the document. That are generated from the metadata. We are trying to get it right in English. The meaning of these different strings, we need more than a translation. We must have it localized that the meaning is equivalent not just the translation. We would like to provide these translations available to everyone to reduce duplication, but we don't want to say its an
official translation but just a suggestion, and others may want to contribute to a localization. May end up with multiple for any given language. So we need to get the strings defined in English and then set up a mechanism to gather other language conversions.
Avneesh: table of multiple translations is not practical. We were thinking of a repository and set up a control system maintained by volunteers. Community group can review and approve / modify these before it gets included in the repository.
George: Rick has offered to translate into 37 different languages.
gautierchomel: 2 aspects: maintaining a list, not sure if we can be the judge of one over another.
… , asking for volunteer providing the translation, is it from one organization or done via a project like we have done in France.
… , have a table of translations we can see if it has been done from a company vs. an organization, so we don't make judgements.
… , first guide asking for feedback, which we have done in France, organizing workshops getting the feedback. What we got a true understanding from all the actors.
… , we now have an implementation in France,
<gautierchomel> implementation in France in 300 independant booksellers webpages.
Madeleine: if we are approving translations wouldn't we need an expert in that language. So Gautier's approach will be more scalable by providing who did the translation.
George: I agee. W3C has a formal process for WG's this is a Community Group, we don't have the resources / language expertise. Identifying the organization they did it and process they used will help people to decide if they want to use the translations.
… , Rick might say this was done by VitalSource and had localization experts do the translation. English should be used and go from that to another language always. that would be the recommended process.
Miiak: I agree, as a speaker of an European language
… , some of these languages may be hard to find. Complicated to get organization for smaller languages.
… , I have seen many translations into my language have multiple errors.
… , We can help with the Finnish language
Avneesh: we will finalize the process in the next Editors call.
Please comment on issue on search and filtering. We will discuss it thoroughly in our December call.
<AvneeshSingh> w3c/
Avneesh: this is important, having guidance on filtering and discovery. Filters are dependent on the audience.
… , we were trying to find out some least common denominators, and can expand for specific audiences.
… , looking pretty complex, we can discuss today if anyone has some ideas.
… , otherwise we can defer to Dec 14th.
George: Would a library / retailer provide a prepopulated type of search? Or if I did a google search of accessible books in EPUB format what would I get.
… , unclear on how this will work. But love the idea of finding accessible books.
gautierchomel: you can search for a word, but list of books. searching for "moby" then get a list of facets, then you can filter on "EPUB" to reduce the list then can further reduce with other facets like supports non-visual reading.
Avneesh: is the filtering criteria is prepopulated.
gautierchomel: the facets are pre-populated.
… , you may have 100 facets, but this could be unwieldily.
<gautierchomel> q§
cristina: What do you mean about filtering. I am not sure how easy this will be. Advance search could be very different depending on the bookstore etc.
… , will there be some kind of search for all math books that are accessible?
Avneesh: Yes that is what we are trying to solve.
<gautierchomel> w3c/
Madeleine: I think we need to say something about filtering, and encourage sites to do filtering, every site does filtering differently
… , current thought minimum filtering that is needed the most then expand. not all sites will have an accessible math filter. Some facets might not appear depending on the site and their audience.
… , I am in favor of a minimum filtering set then expand.
… , helpful to know what is the minimum is needed would be helpful.
Madeleine: Sure I will think about that.
George: if you can add a comment, this would be great.
Avneesh: Any other business? Thanks everyone. Next call Dec 14th last call of the year.