Blind Muslims who read the Quran with Braille software or a screen reader are still let down by the tools in several ways. Our desk check across six readings of the Quran found that Braille software handles none of the Quranic marks beyond the eight basic vowel marks, that nobody has published a Braille key for two of the North African readings, and that one official digital file of the Hafs text stores a single mark in a way that would make screen readers misread it in 3,988 places. None of these problems is hard to understand, and each could be put right with care and with the help of blind readers themselves.
Why this matters
Braille lets a blind person read by touch. Each Braille character is a small cell of up to six raised dots, and different patterns of dots stand for different letters and signs. Arabic Braille uses the same six-dot cell, and it is read from left to right like other Braille, even though printed Arabic runs from right to left.
Many blind readers also use screen readers, which speak the text on a screen aloud, and refreshable Braille displays, which raise and lower small pins to show the text under the reader’s fingers. Both depend entirely on the digital text they are given. If a mark is missing, or is stored as the wrong character, the reader hears or feels the wrong thing and has no way of knowing it.
For most books, a small error of this kind is an inconvenience. For the Quran, every mark can matter. The marks in a printed mushaf guide the reader on vowels, on how letters are joined or lengthened, on where to pause and where not to, and on many finer points of correct recitation. A sighted reader takes them in at a glance. A blind reader can only receive what the software passes on.
There is also a question of dignity. Blind Muslims have the same wish as anyone else to read the Quran for themselves, carefully and correctly, without having to rely on another person to check every line. The tools should serve that wish properly.
What software handles today
Arabic Braille has well-established signs for the basic vowel marks, and the software we examined in our desk check handles these. Beyond them, our finding was that Braille software handles none of the Quranic marks beyond the eight basic vowel marks.
In practice, this means that much of what a sighted reader sees on the page of a mushaf does not reach a blind reader through the software. Signs that help with correct recitation and with pausing are passed over, so the reader receives a plainer text than the one in front of a sighted reader, with no indication of what has been left out.
We want to be careful about the limits of this finding. It came from a desk check, which means we studied the readings, the files and what the software is designed to handle, and we have not yet tested the software with blind readers. We would expect a test with real readers to add detail, and possibly to show workarounds that experienced readers already use.
Readings with no published key
A Braille key is the agreed table that says which pattern of dots stands for each letter and mark. Without one, two people can write the same text in different ways, and software has nothing to follow.
Our desk check covered six readings of the Quran, and for two of the readings used in North Africa we found that nobody has published a key at all. A blind student in those countries who learns one of these readings therefore has no agreed way to read its particular marks in Braille, and anyone who wishes to build software for them has no standard to build on.
This is one part of a wider pattern that we describe in our article on what is missing from open Quran data. Readers of the readings beyond Hafs are already served by far fewer tools than readers of Hafs, and blind readers of those readings are served least of all.
One small mark, 3,988 places
Our most specific finding concerns a single mark. In one official digital file of the Hafs text, the small round zero is stored as the character used for the sukun (سُكون), the mark that shows a letter carries no vowel. In the printed mushaf these are two different marks, and they tell the reader two different things. On a screen they may look much the same, which is how a difference like this can pass unnoticed.
A screen reader does not look at the shape of a mark. It reads the character that is stored and announces the name of that character. So wherever the round zero has been stored as a sukun, a screen reader would announce a sukun, and a blind reader would be told something different from what the printed mushaf shows. By our count this would happen in 3,988 places in that file.
The digital standards that computers use to store text give these marks their own separate characters, so the problem lies in how this one file was encoded. A sighted reader of the same file would see nothing wrong, which is exactly why it matters to check digital text for the people who cannot see it.
We have not named the file here, because our aim is to see the problem put right, and naming the file would add nothing to that. The people who prepare these files do careful and valuable work, and a mistake in encoding is easy to make when the marks look alike.
What would help
Several things would make a real difference, and none of them needs new technology.
The first is for Braille software to support the full set of Quranic marks, so that a blind reader receives everything that a sighted reader sees on the page. That needs agreed Braille signs for each mark, which in turn needs the knowledge of teachers of the Quran and of experienced blind readers.
The second is for keys to be published for every reading, including the two North African readings that have none, so that students of those readings can read their marks in Braille and software makers have a standard to follow.
The third is for digital Quran files to store every mark with its correct character, and for publishers to check their files with screen readers as well as with their eyes. A simple automated check could find cases like the round zero across a whole file in moments.
The fourth, and the most important, is to involve blind readers at every step. They are the people who know what works, what is missing and what matters most, and no desk check can stand in for their experience.
What we are doing next
This work is part of our research into the gaps in Quran technology, gathered under the Murooah Quran Library. Access for blind and low-vision readers is one of the ten gaps we identified, and as with the others, our aim is to check what exists, to work alongside the people already active in this area, and to help fill the holes that remain.
Our plan is to work with volunteer blind testers, so that the people these findings affect can test them for themselves. We have also shared what we found privately with two organisations working in this area. Nothing has been built or published from this research so far, and the library page will show it when that changes.
If you are a blind reader of the Quran, a teacher of blind students or someone who works on Braille software, and you would be willing to help test or correct these findings, we would be grateful to hear from you through our contact page. Every finding here needs checking by people who read the Quran by touch before anything is built on it, and their judgement will count for more than ours.