I'm trying to find guidelines for writing alt text for chemical structures and the Google Scholar search for "alt text chemical structure" is embarassingly bad.

[briney@glammr.us]

@belehaa Thanks for the lead. The math community seems to have some established guidance for equations, but I've not found similar resources for chemistry. Hopefully, I can connect with someone working on this issue.

[ardeninred@glammr.us]

@briney not chemistry specifically, but I have found this guide to be really helpful for writing alt-text (or creating alternative files that are text only) for many different complex image types: http://diagramcenter.org/table-of-contents-2.html

[carolski@mas.to]

@briney It was really fun! And during a very early time in the pandemic, it was something I could do that didn’t require me leaving home. I felt lucky to have the chance to do it.

[briney@glammr.us]

@carolski Writing a book is a big project but rewarding!

I'm actually writing a book on scientific data sharing and trying to include some accessibility principles in the book, hence the alt text question.

[briney@glammr.us]

@ardeninred I love this resource but I'm a bit disappointed that the only chemistry content they describe is the chemical element.

[quantumdot2@mas.to]

@briney I honestly thought there'd be some good guidelines out there, but I'm coming up with nothing useful from ACS (American Chemical Society) or RSC (Royal Society of Chemistry).

Julia Winter's company (Alchemie) is working on some sort of algorithmic alt-text generation. You can read some of the basic ideas here: https://pubs.acs.org/jceda8/article/101/7/2880/156131/The-Lewis-Structure-Explorer-Accessible-by-Design and more https://www.linkedin.com/pulse/accessibility-crisis-organic-chemistry-what-instructors-really-mc47c/ and a commercial product: https://www.alchem.ie/piph

ACS has a committee that I would have thought would have guidance (https://www.acs.org/about/governance/committees/chemists-with-disabilities.html) but all I could find was "Complex graphics and some other sections of online material may still require the use of a human reader or sufficiently written alt-text to be fully accessible to individuals with visual disabilities. This is due in part to the fact
that no standard requirements for figure creation and alt-text to describe figures are in place for most online publications.”

https://www.acs.org/content/dam/acsorg/about/governance/committees/cwd/teaching-chemistry-to-students-with-disabilities-5th-edition.pdf

Hey - if you find something good out there, please share!

[briney@glammr.us]

@QuantumDot2 thanks for sharing what you found. I too am suprised that there’s nothing from ASC or RSC. With the new U.S. accessibility rules going into effect next year, I really hope someone is working on this!

[srlevine@neuromatch.social]

@QuantumDot2 @briney @belehaa @carolski I found a handful of things (including that same ACS not particularly helpful guidance) last time I looked

limited help, but maybe something to start with (Lewis structures aren't really on the scale of what you want to describe): https://pubs.acs.org/jceda8/article/101/7/2880/156131/The-Lewis-Structure-Explorer-Accessible-by-Design

and this from libretexts for designing coursework: https://commons.libretexts.org/insight/organic-chemistry-accessibility

I think there are arguments to be made for InChI keys rather than (or in addition to) SMILES for things that have ambiguous SMILES, but I think libretexts is probably on the right track

[briney@glammr.us]

@bnut thank you so much for sharing the resource and your experience. This is really helpful, especially to know these strategies have been used in the classroom.

[briney@glammr.us]

@SRLevine @QuantumDot2 @belehaa @carolski oooh thanks. I have a feeling like guidance is going to be scattered across a bunch of different resources like this.

[idlestate@toot.cat]

@briney

One thing I get hung up on about this is that a systematic IUPAC name for any structure, in principal, should describe a structure fully.

This is a level of completeness that very few visual depictions of any kind have available, even if only in principle.

But as most of us who've worked with these things understand, that's cold comfort, because it transmutes the problem into one of being able to read and understand those names, which can contain highly nested components, punctuation, Greek letters, italics, and so on.

These are almost certainly very unhelpful when passed through a screen reader, for instance.

There's a reason we use diagrams so much!