Chemistry in EPUB -- 06 Feb 2019

<scribe> scribenick: dauwhe

George: I'm george kersher
... I've been asked to pull this group together for best practices for both EPUB and the Web
... we can start with introductions

Bill_Kasdorf: I'm a consultant, participated in PWG and EPUB work in IDPF. I do modeling and workflows and a11y.

Steve Noble: I'm with Pearson

scribe: work with school assessment
... interested in chemistry on state assessments

dauwhe: Dave Cramer, chair of EPUB 3 CG and on CSSWG

John Gardner: founded a company that makes embossers for tactile graphics

scribe: I work on access to graphics

Avneesh: DAISY consortium and chair of EPUB a11y group

Janina: I'm hearing some familiar voices. I chair a11y platforms wg
... I remember a discussion about doing a better job with domain-specific knowledge

judy brewer: w3c, improving domain knowledge representation, and computer models of metabolic disorders

Jason: working on edu a11y at educational testing service
... involved with diagram center

erin lucas: redshelf, a reading platform

???: Work at edu testing service, do research in chemistry education

liam: garble garble garble
... garble Prague garble W3C

Makoto: was member of original XML working group

<liam> if you can't hear me, i'm Liam Quin from Delightful Computng (independent consultant), been doing XML & digital typography for ever :)

Makoto: I know Peter who did ChemML
... now member of APL at Keio University
... I work with Hiroshi of Japanese DAISY consortium

SueAnn: Director of Diagram center

George: department of education is interested in STEM

<Makoto> Including IUPAC?

George: symbolic representation of information is important
... we'd like to identify best practices on how to present information that is visually appealing and accessible to people with disabilities

Volker: I've build a workflow for a11y molecule diagrams
... and I've done a best practice around this in the Netherlands

George: I linked to that announcement in the resources in the meeting announcement
... the goal is a best practice of what publishers can do today both in EPUB, and the web setting
... publishing @w3c is working on web publications
... there are two different specs, one to be consumed on the web and one packaged and delivered through EPUB 3
... there may be a scope problem
... people might want to take Volker?'s work or ChemML and take it further
... MathML 4 work is happening in a CG now
... at this point I'd like to open it to discussion

???: I'm chairing a committee on chemistry for braille

scribe: standardizing braille code for chemistry
... we've discovered that there's non-uniformity in what symbols textbooks are using
... so there was braille for each variant in published materials
... so we're partnering with the ACS to work on standardized symbols
... and we want the textbook industry to adopt the standard symbols
... this will make our work in braille easier

George: we're working in the standards domain
... harmonization of standards is an objective
... I'd like to see one standard
... so what we're doing could dovetail
... I don't know how publishers are getting symbols into the documents
... are they using MathML? CML?

???: I don't know either, but I have friends who do

???: At Macmillan we use MathML in our EPUBs

George: I've also heard that AT can't tell if something is math or chem

Volker?: what do we mean by MathML?

Steve: if they're using markup at all, they're using MathML.

<janina> Telling one knowledge domain's symbology from another's may be a job for the knowledge domain abstraction layer we hope to define in W3C

Steve: some AT might try to identify a chemical forumula
... do you want to hear "Sodium Chloride" or "NaCl"?
... there's nothing like it in JAWS, but it is in NVDA.

George: I hear there's no identifier that points to whether an equation is math or chem

???: It would be easier if they were left in LaTeX

janina: this might be solved by knowledge domain discussion in w3c
... with an abstraction layer
... the superstructure needs to handle that. I'm working on a writeup

George: this is a CG?

<janina> https://github.com/w3c/apa/issues/9

janina: it will be soon

Judy: that reference should be in george's reference link

George: this means that there are two things
... one, there's a short term issue of what publishers are doing today
... how that can be standardized and developed as a best practive
... and there's a longer term thing with symbology coming out of w3c, as Janina mentioned

CharlesL: what's the difference between ChemML and CML?

Volker: i think it was a renaming
... it represents the visual layout of the formula
... having worked with many of these
... the two useful ones are ChemML and ?mol?
... I wanted to clarify one thing
... we don't want to propose novel rendering solutions
... we have SVG, mathml, etc
... we should focus on what chemical annotations we need

Avneesh: rendering comes in with EPUB
... which is why mathml solutions are attractivve, because MathJax is available
... so how would we render ChemML in EPUB?

Volker: prerender as SVG, and have ChemML as hidden annotation

George: at the base level you have SVG, with alt text in natural language
... in addition, have the CML that could be available to AT that knows how to process it

<liam> [i wonder what you do about stereochemistry if you fall back to SVG]

George: is there any ability for the SVG to be decorated with ARIA so that it could present itself

Volker: depends on the presentation
... if you just want to tap through the molecules, you could have invisible groupings
... most chemicals are graph structures, you need a graph for navigation with JS
... we've also embedded some things in SVG

Avneesh: this can be svg based

CharlesL: then you can zoom in
... for a 3d model that might be out of scope?

George: it would be great to have a collection of 3d models
... in the image share collection
... so the school could print them out

janina: you were also on the VR call yesterday
... that might be low-hanging fruit for that tech
... it's simpler than modeling the real world for games
... having an equation you can walk around

Jessica: you say the molecule doesn't change, but sometimes pedagogy demands having lots of different representations

Judy: once we start enabling graphical representations, we might think of supporting many layers/stages
... graphical annotations, or crystalline models
... being able to link all that data in is good

George: one of the principles we outlined in our preso of math at AHG
... we're looking at trying to meet the needs of a comprehensive group of people with disabilities
... audio, tactile, braille... but we need to support low-vision, which SVG should serve
... and highlighting or exploring in conjunction with speech would be good
... and audio presentation for people with dyslexia or learning disabilities

Bill_Kasdorf: I'm forgetting where you've published your AHG presentation on math. Could you provide a link?

George: we're testing our math stuff
... when I crack it open and use the HTML, it's working in a few environments but not all environments
... and we're not using mathjax
... we have the ability to test these solutions

dog: woof woof woof

Ron: We've been working on a project to transcribe math
... one problem we have is interleaved mathml and ascii text
... some chunks are just text, and didn't end up as an image

George: that brings up a good question of best practices
... if you have a + b = c, sometimes publishers won't use MathML or an image, but then you have inconsistency

Jessica: Or do you chemML for H20

Ron: You might here two minus three and then two dash three

George: is this an issue with your OCR?

Ron: the problem is the source. If it's not complex, leave it as text, if it is, make it mathml/image

Steve: we did research in Kentucky
... if it was on the keyboard they did ascii
... they only did an equation if they couldn't do it with their keyboard

George: Neil helped us with the sample test book on epubtest.org
... we have ten minutes left
... where do we want to go? Do we want a small group to create samples?
... do we want regular meetings?
... how do we want to move forward

Avneesh: do we want key objectives
... in short term (6mo or 1y) and long term?

<mateus> +1 to Avneesh

Judy: that's a good starting approach, Avneesh
... we might want to consider a set of objectives in parallel
... we hope tehre will be a central effort in w3c for different knowledge domains
... in chemistry, we need a community to define the specific challenges
... and an approach for bringing the relevant communities into a harmonized approach
... the comment about the interleaving problem
... ideally we could address a whole set of issues like that

Bill_Kasdorf: I have a scope question
... the mention of NIMAS... are there issues between math for K-12 vs HigherEd vs researchers?
... is chemistry the same across all those?

volker: no
... in lower grade, you have very explicit diagrams
... for higher grades, you have skeletal depictions of molecules
... with lots of omissions

George: Judy, you're suggesting that the community work on knowledge domains
... already exists...

Judy: there's not a CG yet, but there will be soon
... there are discussions in the Research Questions Task Force (RQTF) of the Accessible Platform Architectures (APA) Working Group
... and we need stuff outside of accessiblitiy
... there's discussion in the TAG
... a Community Group (CG) focus could be the umbrella discussion
... but then within each domain there's another discussion
... what objectives do we need to address within chemistry?

George: There's a clear separation between the EPUB CG developing a best practice, vs the long-term symbolic knowledge domain in w3c
... the focus of this group would be to develop best practices around SVG display of chemistry
... with alt text to support that, and a notation like CML or MathML with some chemistry flag for presentation
... we can test different approaches and see what works in the real world

janina: we're likely looking declaring a container
... I think we'll need them

Jessica__: with your best practices, you also need to talk about the cutoff between ascii and markup

Bill_Kasdorf: I have another scope question
... there's presentation mathml and content mathml

Judy: we should avoid that problem

Steve: at pearson, we use content mathml in the testing environment for auto-scoring
... using mathml for chem is suboptimal

<liam> [decorated presentation is a better strategy than two different languages ala MathML]

Volker: if we have CML or MOL in the page, we have all the info we need

George: this is the short-term group, many of us will be happy to join the other CG
... should we meet in two weeks?

Avneesh: let's meet in two weeks
... and then work on Github

<Bill_Kasdorf> +1 to two weeks

George: same time? on the 20th?

<liam> [hard to keep momentum monthly, +1 to 2 weeks]

Avneesh: we should start a new issue to document things

George: thanks everyone, I should have a recording of this. We'll talk in two weeks

RRSAgent: draft minutes

RRSAgent: make logs public

<CharlesL> RRSAgent: draft minutes

- DRAFT -

Chemistry in EPUB

06 Feb 2019

Attendees

Contents

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output