on this page...
- purpose
- related urls
- files here and in the dev-server-dir
- chatGPT image-accessibility exchange
- next steps
- TODOs
- to explore what we can/should do to improve the Studio's image-accessibility.
(I ended up focusing pretty-quickly on IIIF-viewer-accessibility; we need to broaden that scope. But I think the things that can improve IIIF-viewer-accessibility will be core to improving the accessibility of collection and search-results pages.)
- relevant bdr item: https://repository.library.brown.edu/studio/item/bdr:247087/
- relevant bdr mirador viewer: https://repository.library.brown.edu/viewers/mirador/bdr:247087/
h_current.html-- (theh_indicates html) the current html for our mirador viewer that you'd see if you did a view-source on the viewer url, above.h_more_accessible.html-- chatgpt-suggested html that loads more-accessible manifest-json. Note that (likecurrent.html) viewing this file doesn't show much image information -- the js needs to be loaded. Seefull_more_accessible_generated_htmlbelow, or right-click and inspect the generated html from a browser.dummy_manifest.json-- chatgpt example minimal version-2 manifest-json loaded-manually bymore_accessible.html.full_current_generated_html.txt-- the html produced by javascript executing when loadingcurrent.htmlin a browser.full_more_accessible_generated_html.txt-- the html produced by javascript executing when loadingmore_accessible.htmlin a browser.
https://chatgpt.com/share/68d17932-7c10-8006-b51e-1518829d8eaa
Summary...
- I asked it to analyze info from https://repository.library.brown.edu/studio/item/bdr:247087/ for image-accessibility. It had a bunch of good suggestions we can implement, like adding aria-labels to buttons and toolbar-icons, just as one example.
- It mentioned "Provide meaningful alternative text and long description", for that item-page -- which made me wonder how the image-viewer itself would fare.
- I shared the html from the stand-alone mirador-viewer at https://repository.library.brown.edu/viewers/mirador/bdr:247087/ and asked for an analysis.
- It noted improvments and output suggested updated html (
more_accessible.html). - I was interested in this piece because it noted that the source-data should come from the manifest.json file, which the page's js would cause to be converted into more accessible html.
- It showed a sample v3 manifest.json file (with this image's data) -- but our viewer's older, so I had it make a v2 version.
- It mentioned a workflow of data coming from MODS to populate the manifest.json.
- Curious about that, I asked it where image-description data should go in the MODS. It recommended mods:abstract and mods:note, with a few additional attributes -- primarily
@type="image-description"and@type="image-description-long"-- to distinguish between abstract/note entries that do and don't have useful image-description info. It also offered other MODS suggestions.
I have some thoughts; will update this more tomorrow.
- The above focus was on an item-page. We also need to review collection and search-results pages, which show multiple images.