Accessibility Remediation Metadata (ARM)

Accessibility Remediation Metadata (ARM) refers to a metadata model, currently under development by a National Information Standards Organization (NISO) working group, for describing remediation work: the process of taking a publication or other content resource in an inaccessible format and altering it into a format usable with assistive technology (screen readers, refreshable braille displays, text-to-speech tools, and similar). ARM metadata records what remediation was performed, to what accessibility standard, in what format, and where the resulting accessible copy can be found — so that institutions and disability services offices are not duplicating the same remediation work on the same title.

As of this writing, ARM is a working group in progress rather than a ratified standard. NISO’s Information Discovery and Interchange Topic Committee approved the ARM working group’s charge in early 2024, and the group’s stated goal was to develop a machine-processable metadata schema and supporting documentation with an eye toward eventual ANSI/NISO standardization. Readers evaluating ARM for practical implementation should check the current status directly on NISO’s site (niso.org/standards-committees/arm) rather than assume it has already been finalized — this page describes the model as publicly documented, not a completed, ratified standard.

What ARM builds on

ARM extends and generalizes an earlier metadata model developed for the FRAME project (Federating Repositories of Accessible Materials for Education), a Mellon Foundation-funded effort that built a metadata approach to support discovery and deposit of remediated educational materials, most visibly implemented through the EMMA (Educational Materials Made Accessible) shared repository. ARM’s working group set out to broaden that model beyond the education sector to serve a wider range of stakeholders, including publishers, libraries, and disability services offices working with scholarly and general-interest content, not only textbooks.

Operational definition: what makes something ARM metadata

• It describes a remediation event, not original publication. ARM metadata is attached to an accessible alternative produced from an existing source resource — it documents that a transformation happened, not the accessibility features of a file as it was first published.
• It records the target accessibility standard. Typically a WCAG conformance level or an EPUB Accessibility specification target, so a requester or repository can tell what the remediated file actually achieves.
• It supports discovery and/or deposit. The point of standardizing this metadata is to let institutions search a shared repository to check whether a given title has already been remediated before duplicating the work, and to deposit newly remediated copies for others to reuse.
• It identifies who performed the remediation and in what format. This includes the remediating organization/individual and the resulting file format (for example, a tagged, text-layer PDF or an accessible EPUB).

Worked example

A university disability services office receives a request from a student for an accessible version of a course reading that exists only as a scanned, image-only PDF chapter. The office remediates it — adding a text layer, correct reading order, and image descriptions — to a stated WCAG conformance target, and deposits the resulting accessible PDF in a shared repository along with metadata describing: the source resource, the remediation standard targeted, the format produced, and the remediating institution. A second institution with a student needing the same chapter can search that metadata, find the already-remediated copy, and reuse it instead of independently remediating the same source material — this reuse-to-avoid-duplication scenario is the core use case ARM metadata is designed to support.

Counter-example

A publisher marking up an EPUB at the time of original production with schema.org accessibility properties (such as accessibilityFeature, accessibilityHazard, and accessibilityAPI) to declare that the born-accessible file already conforms to a given WCAG or EPUB Accessibility level is a related but distinct practice: it describes the accessibility of the resource as originally published, not a remediation event performed on an inaccessible source afterward. That schema.org accessibility vocabulary is separately established and already in use in EPUB and ONIX metadata; it is not the same thing as ARM, though the two are complementary parts of a broader accessibility-metadata landscape and a resource could in principle have both born-accessible metadata and, later, ARM remediation metadata if it is subsequently altered again.

Why it matters for scholarly publishing

Scholarly and educational publishers, libraries, and disability services offices face legal obligations (under laws such as the ADA and Section 504/508 in the US, and equivalent accessibility regulations elsewhere) to provide accessible versions of content on request, but remediation work is frequently duplicated across institutions because there has been no standard, shared way to record and discover what has already been remediated. A finalized ARM standard is intended to reduce that duplication, make accessible alternatives easier to find and trust, and give publishers and repositories a consistent way to exchange remediation metadata.

Related terms

See also PDF/A (ISO 19005) for a related, already-ratified format standard relevant to long-term accessible archiving of documents.