Daejeon, July 27–31
Founded in 1889 and recognized as a public-interest organization in 1891, the association Valentin Haüy (AVH), https://www.avh.asso.fr/ (Accessed March 24, 2026) created by and for blind people, promotes independence in learning for blind and visually impaired people (BVIP), both within the association and through its library services (Roy 2019, Miklas 2022). Its heritage library is particularly notable as one of the first in the world to assemble an extensive body of knowledge on blindness. It houses a wide range of documents produced by newspapers, authors, and institutions, reflecting the rich history of the blind community.
This project proposes a methodology for producing accessibility-oriented, TEI-based scholarly editions, validated through a co-design process with BVIP participants. Rather than introducing a new platform, it focuses on adapting encoding, enrichment, and publication workflows to ensure accessibility throughout the editorial process.
This three-year project is based on a close collaboration between Sorbonne Université and the AVH. The two partners combine their expertise on the AVH heritage collections: selection, digitization, encoding, and the development of access tools are carried out jointly, drawing at every stage on both academic research and the association’s in-depth understanding of the needs of BVIP. This complementarity forms a partnership model that integrates heritage preservation, technological innovation, and accessibility.
For blind readers, access to written content relies either on braille, a tactile writing system, https://en.wikipedia.org/wiki/Braille (Accessed March 24, 2026) or on screen readers, which convert text into speech or braille output. By contrast, standard print texts remain inaccessible without braille transcription or machine-readable formats (Chottin / Roussel 2023).
The AVH heritage library holds a large number of such documents, representing over one million pages of diverse materials. These range from journals dedicated to the dissemination of knowledge about blindness, particularly aimed at typhlophiles A person who loves or/and is interested in helping blind people (“typhlòs” = “blind” and “philos” = “loving”). to collections focusing on blind veterans, particularly reflecting the significant role played by the AVH during World War I (Chottin / Roussel 2023). Although researchers from around the world consult these documents, they remain poorly catalogued and are often fragile.
These factors led to the launch of a project aimed at producing digital scholarly editions of these materials. However, given the diversity of the target audience, the methodology must be adapted to address a wide range of accessibility needs.
A DSE, as presented by (Sahle 2016), is the result of a process combining transcription, encoding, enrichment, and publication, aimed at disseminating corpora and facilitating access to resources for researchers. In recent years, the development of DSEs has expanded, with the emergence of new publication platforms offering rich functionalities, such as those supported by TEI Publisher, https://teipublisher.com (Accessed March 24, 2026) including DiScholEd. https://discholed.huma-num.fr/exist/apps/discholed/index.html (Accessed March 24, 2026) However, while these features enhance access for researchers, they are not fully compatible with accessibility for blind users, as they rely primarily on visual interaction.
At the same time, libraries have developed strategies to improve access to their collections, providing resources such as braille materials, audiobooks, and screen reader software (Srivastava / Kumari 2025). However, these approaches often target blind users specifically, whereas our objective is to create a scholarly digital edition that is accessible to all audiences, whether academic or non-academic, blind or sighted. This requires addressing the challenges posed by diverse target audiences while leveraging opportunities to share knowledge as broadly as possible. The aim is to provide sufficient and meaningful access for all users without compromising the richness of the data.
This requires careful consideration of several stages of the process. First, in accordance with Text Encoding Initiative guidelines (TEI Consortium, eds 2026), encoding must be handled with particular attention in specific cases. Its richness allows for the tagging of textual elements that may initially seem unnecessary but become essential in contexts such as text-to-speech output, for example when a character functions as a mathematical symbol (Figure 1).
Figure 1: Example showing the encoding of a mathematical symbol from the October 1924 issue of Bulletin mensuel de l’Union des Aveugles de Guerre et Journal des Soldats Blessés aux Yeux.
Encoding typically adheres closely to the source document. However, adding contextual and structural elements can be necessary to improve navigability, particularly in lengthy texts. This may include the systematic insertion of a table of contents or the introduction of headers for tables and lists when they are absent, thereby enhancing readability and user comprehension.
Corpus enrichment may focus on named entities, domain-specific terminology, or critical annotations, depending on the priorities defined by the BVIP committee. It should be implemented through structured navigation, explicit textual labels, and WAI-ARIA-compatible markers (W3C 2023), rather than relying solely on colour or pop-up-based visual cues.
Finally, an online platform will be developed to host the documents in multiple forms: digitized facsimiles, transcriptions, and enriched versions. It will offer intuitive navigation across the corpus, along with semantic, analytical, and thematic search tools to support exploration. A first prototype, built with Omeka, a CMS with a range of useful plugins available, has already been developed (Figures 2a and 2b). However, it currently provides only basic visualisation, with accessibility features planned but not yet implemented (Figure 3). Further development will proceed as encoding and enrichment advance and project requirements are clarified.
Through our work on the creation of DSEs for both researchers and blind users, we aim not only to improve access to documents relating to the history of blindness, but also to contribute to an ongoing movement toward greater openness, web accessibility, and inclusivity. This movement was notably highlighted by (Martinez et al 2019) in 2019 through a survey on open access and accessibility, and further developed in 2024 by (Cappellotto / Cioffi 2024) within the DIA-DSE project.
While we do not propose fundamentally new methods, we seek to introduce a shift in perspective in the design and production of DSEs, in which accessibility is integrated at every stage of the process.
Figure 2a: Homepage of the website prototype
Figure 2b: Document page of the website prototype
Figure 3: Functionalities under development for text-to-speech, Braille, and print