Meeting minutes
<AvneeshSingh> Discussion on grouping of the relevant accessibility metadata for user experience guide for accessibility metadata.
jean-benoit (leslibraires.ca): here to present the work we do in French Canada with funds from canadian governement. With Domenico and Christian.
DBauer: represent German libraries and German metadata group
Presentation from leslibraires.ca for their work on displaying accessibility metadata.
AvneeshSingh: background is we published a display guide and now we meet real book seller displaying this metadata.
Christian Roy: I'm a consultant working for leslibraires.fr, previously at DeMarque, been in ebook distribution for a while. Domenico will show our mockup, we are not a retailer with access to epub files, exclusively working with ONIX files. All of our work has to be mapped to ONIX provided by publisher's. We based our work on current data we find, should change when publisher's give more and better metatada. What Domenico will show is to evo[CUT]
Domenico: thank you for inviting us. I'm not a technician, my speciality is user's experience, my background is on research. I do mockups, and study the use made of them then redo mockups in a multiple iteration process.
Domenico: I work a lot in SEO and know a bit of accessibility but I'm not an expert. I'll show my screen and describe everything, please say if it's unclear.
<gpellegrino> 1+
Domenico: this is the book page, what appears on the first screen, what you immediately see is cover, resume, different versions of a book (paper, epub, pdf) and there is one line here saying the epub version has 6 of 9 accessibility criteria ; pdf has 0 of 8 accessibility criterias. If you clic on one of thoose mentions you're transported to the bottom of the page where a list indicates "the epub version is compatible with screen reader", etc.
Domenico: The list of criterias is not yet fixed. This list allows you to directly add to basket or choose another version recommanded as more accessible or audio version. List items of accessibility funcions included are green, missing ones are red and an orange says there's a particular point of attention
Domenico: The pdf list with 0 of 9 criteria ends with "we do not recommand this pdf version for accessibility".
Domenico: we also propose an accessibility page that combines only books with accessibility feature with a set of search filters dedicated (large print, audio, syncronized, navigation, etc.).
Domenico: we did not tested in real situation with blind users but had good feedback from others users.
gpellegrino: it's difficult to be sure something is missing because ONIX only allows to say if the feature exist. If it is not present it does not mean it's missing (exemple no alt text may be informed because there are no images).
Christian Roy: yes the hard part is interpretation of ONIX. We anticipate we'll have grey icons (meaning we don't know) and hope we'll have more precise metadata with time.
CharlesL: great presentation, icones are great what labels are attached? Watch out for repeated text, and WCAG AAA should be only AA as AAA can conflict for different disability groups. I like the approach.
gautier: maybe we can cross other lists informations to check if something is missing (the publication has images but not alt text code is represented means image descriptions are missing).
So if I have 81-19 but no 196-16 I have an accessibility problem
Naomi: yes, no mathML does not presume there is Maths in the publication.
Hadrien: working on this we realised onix is problematic in many ways (access mode and features missing). We started from the crosswalk document and iterated on that and identified missings. We need to engage a discussion with Editeur to obtain a better mapping from onix to schema.
AvneeshSingh: you are right we have to move this forward.
AvneeshSingh: Jean-benoit, Christian, Domenico, would one of you join this group on a regular basis.
Discussion on grouping of the relevant accessibility metadata for user experience guide for accessibility metadata.
<AvneeshSingh> @charles, can you please provide link to crosswalk
gpellegrino: we created this spreadsheet to prepare rewriting of the display guide. It's clear we have a large amount of information on accessibility, more than bibliographics ones. We have more than 40 informations, displaying everything to user will not help understanding. Sometimes it's difficult to understand what an information means, even for us experts on accessibility. The idea is to group metadata to be able to display them in a shor[CUT]
<AvneeshSingh> spread sheet: https://
gpellegrino: understandable. It's a starting point to be able to propose wordings.
gpellegrino: this work raised questions discussed in the mailing list.
<CharlesL> Here is the Schema.org / ONIX Crosswalk https://
<MiiaK> Sorry on the mobile. Thanks Wendy!
Miaak: do we understand correctly that accessModeSufficient:visual means a book with images not described
Madeleine: yes accessmodesufficient:visual only means there are lacks of accessibility
Madeleine: if it is the only accessmodesufficient present.
JF: we shold say expose or don't expose, we should not let it to devloppers asumptions.
<JF> +1 Naomi that was my concern as well
Naomi: we apply "visual" to mangas and this type of works, it's clear you can't consume them with access to text only
wendyreid: same point. reflowable with text sure don't necesary need this information but for fixed layout it is important to inform that text access is not sufficient.
<wendyreid> +1 to the display concerns
Hadrien: I go the same direction, comics, mangas, bd are access mode visual actually with no alternative. It's important to support it even if it looks difficult to find a way to display it to the users.
Madeleine: big difference is if another accessmodesufficient is present. Make sure the recommandation does not comes too complexe. A way is to use visual alone as a flag for accessibility difficulties
gpellegrino: a super edge case is only one alt text is missing, trying to determine accessibility by a missing information may create false positives.
CharlesL: in that case you use visual comma textual. I would not put only visual.
AvneeshSingh: next step is creating issues on the github so we get a larger feedback.