This document outlines the workflow used by the Norwegian Braille Authority to publish our braille standards online.
Background
For many years, we have authored our standards in Microsoft Word. From these Word documents, we generated PDFs for publication on our website and for producing printed copies. For embossing 6-dot braille on paper, we have relied on an older software package called CX for Word, still used in Nordic countries.
Our standards include many braille examples. To achieve visually appealing braille in Word and PDF formats, we’ve used a legacy font called BLOCK, based on the Eurobraille system. This means that characters like a–z map to standard braille symbols, and numbers 1–9 are formed by adding dot 6 to the letters a–i. However, when such documents are read with a braille display using a different braille table, the results can be confusing.
Motivation for Change
About five years ago, we initiated a project to improve how our standards are published online. The project had three main goals:
- Ensure accessibility and readability, both visually and on braille displays
- Use HTML as the file format
- Still use the old Word documents as source
To ensure accessibility and readability regardless of braille table we use characters from the Unicode braille block to present braille.
Although some fonts exist that visually represent braille, we couldn’t find one that supported the Unicode braille block with appropriate shadowing for unraised dots. As part of the project, we therefore developed our own set of braille fonts.
It’s important to note that the Unicode braille block (U+2800-U+28FF) is formally designed for eight-dot braille, not six-dot. While the first 64 characters in this block correspond cleanly to six-dot braille patterns, dots 7 and 8 are always blank in these cases. This causes the rendered characters to appear too high relative to the baseline in typical fonts, since font rendering engines align braille cells based on the full eight-dot height, even when only the top six dots are used. This can create visual misalignment in both print and digital formats.
Custom Braille Fonts
We commissioned programmer Jan Martin Kvile to develop the required fonts. These include both shadowed and unshadowed versions, and support for 6- and 8-dot braille. The fonts are freely available in our GitHub repository, under the fonts directory.
The folder contains the fonts in TTF and WOFF/WOFF2 formats, packaged into two ZIP files. Each archive includes a README with detailed information. The NorBraille_woff.zip file also contains a style.css stylesheet needed for web integration.
The font is licensed under SIL Open Font License (OFL), and can be used, modified, and distributed freely.
Conversion Workflow
The conversion process is performed in a Linux environment using shell access and has not been tested on Windows.
To replicate the workflow, you will need:
- Pandoc
- LibreOffice
- Two custom Perl scripts: Found in the utils/scripts directory of our repository
- Pandoc template: Located in utils/pandoc_template
To view the resulting HTML, place the file on a web server and include the appropriate fonts. These can be extracted from the NorBraille_woff.zip archive.
Future Goals
We have yet to establish a unified workflow where a single source document can seamlessly be converted to embossed braille, HTML, and PDF. Our ideal solution would use Unicode braille and our custom fonts directly in the source file. However, it is unclear whether braille software like Duxbury or CX for Word supports this approach. We plan to explore this further.