In search of a guiding theory for alt text for academic literature

Sterman, Leila

doi:10.5281/zenodo.17940862

In search of a guiding theory for alt text for academic literature

Sterman, Leila

Publisher: Zenodo
DOI: 10.5281/zenodo.17940862
Repository: zenodo.org
Language: en

Abstract

As we build and adopt best practices for repository compliance with Title II of the ADA in anticipation of the April 2026 deadline, many questions are left unanswered or unsatisfyingly answered. Setting aside for a moment the complications of Section 121 of US Copyright with Title II compliance and the practice of commercial academic publishers signing and then ignoring accessibility clauses in contracts, we can assume that many repository practitioners will be remediating PDFs of academic publications in the near future. A lot of the remediation (tagging, reading order, OCR) is straightforward, if time consuming. Alt text, however, seems to be more complicated. The general guidance for alt text is to provide alternative text that represents the content and function of images. This is often done as an attribute of the image element. How should we as non-subject experts create the right text (preferably under 140 characters) or check its verity? For some images, the alt text can be provided in body text nearthe image. Potentially for academic articles, the text in the cation combined with the body text of the manuscript is sufficient. How will we know if it is not? This presentation is a call to think about the usefulness of each case and how we, as non-subject experts, can create guidance for appropriately working within these parameters. Eventually, these discussions could lead to building a community of practice with both theoretical and practical guidance. There will be a mechanism for feedback, reports of current practices, and hopes for the future.

Full text

In search of a guiding theory for alt text for academic literature Leila Sterman Montana State University Alt text guidelines •Be brief but specific: Aim for one sentence that explains what’s important about the image. •Focus on meaning, not appearance: Say what the image shows, not just what it looks like. •Think about context: What does the image do in this specific place? Why is it there? •Alt-text doesn’t need to describe everything: just what someone needs to know to understand the content. •Do not repeat information available elsewhere, e.g. captions and axis labels for figures and charts. How to apply this for images in academic papers? A chart with a useful caption. Toric code lattice: On the Left, the toric code with spin-1/2 sites located on the links of a 2D square lattice. The star operator is the product of on the sites neighboring a vertex v, while the plaquette operator is the product of on the sites neighboring a plaquette p. at a site i flips both adjacent stars, creating two e excitations, which can then hop when acts on a neighboring site j. Likewise, at site k flips both adjacent plaquettes and creates two m excitations, which can hop by the action of on a neighboring site ℓ. We compute the response functions on a cylindrical geometry; the thick line at the Bottom represents the periodic boundary. On the Right, we model the toric code excitations as hard-core anyons hopping on a square lattice and its dual lattice; the e’s live on the lattice sites and are created/annihilated by the operators at each site, while the m’s live on the dual lattice and are created/annihilated by the operators . Each anyon carries with it a string (solid orange for e, dashed blue for m) which runs to the rightward then downward to terminate at the periodic boundary. Tracking how these strings sweep over each other accounts for the phase acquired when e’s encircle an m (or vice versa), as well as any phase acquired from incontractible loops around the periodic system (49 49). A graph with a bad caption. A graph with a useless caption. Current Practices? •Please share! Alt text in academic article remediation –Fill out form