Hey blind people, what is the recommended accessible format for distributing text-only content?
-
@skye partially sighted here, hope reply was okay, our eyes just flat out do not work a lot of the time nowadays, like rn
@chaos thank you!
-
@skye PDF loses absolutely all semantics, it essentially turns all your formatting into blocks of letters with a specific font, size and position, it only cares about the visual layout of the document.
Using a screen reader on one, you'll at best get it to read line-by-line, but it will treat a line as a paragraph and make weird pauses and generally be a bad experience.
You can make a standalone HTML file that people can open in their browser locally, as long as you don't put in any external resource references in it, HTML is great at conveying semantics if you use the correct tags.@skye I've prepared a comparison with PDFs generated with pandoc vs the same content as HTML, need to edit the recordings a little though, but here are the files.
I did notice that you can get somewhat more accessible PDFs with different Pandoc backends (like typst) though even then it only fixes headings. -
@skye PDF loses absolutely all semantics, it essentially turns all your formatting into blocks of letters with a specific font, size and position, it only cares about the visual layout of the document.
Using a screen reader on one, you'll at best get it to read line-by-line, but it will treat a line as a paragraph and make weird pauses and generally be a bad experience.
You can make a standalone HTML file that people can open in their browser locally, as long as you don't put in any external resource references in it, HTML is great at conveying semantics if you use the correct tags.@lunareclipse @skye Possibly relevant: PDF is often absolute garbage at non-ASCII. Sometimes ç or ã render properly but the actual text generated is incorrect when one copies it, and for some reason "ti" becomes a theta on some documents. Everything about PDFs sucks. EPUBs are better for very long, book-styled documents, HTML is a better option regardless.
-
Hey blind people, what is the recommended accessible format for distributing text-only content? So that a short text of a couple hundred words is browsable, searchable and the structure is understandable?
Am I correct is assuming that a well-structured PDF with all the titles set correctly would do the trick? Or is there some secret file format that is much better suited?
The document is intended to be distributed as downloadable file, not as a website.
(Sighted people feel free to boost but for the love of everything please do not try to answer)
Edit: So what I have gathered
- PDF actually a terrible option
- HTML good, but not easily downloadable
- Epub probably good, and easily downloadable
- Markdown might be preferred for people that know how to work with them
- Why the fuck is this still so difficult in the year 2026
@skye Markdown. Never do PDF ever. Saving it as HTML is best too since they can open local HTML files
-
@skye Markdown. Never do PDF ever. Saving it as HTML is best too since they can open local HTML files
@WeirdWriter Thank you, this is helpful.
-
@skye I've prepared a comparison with PDFs generated with pandoc vs the same content as HTML, need to edit the recordings a little though, but here are the files.
I did notice that you can get somewhat more accessible PDFs with different Pandoc backends (like typst) though even then it only fixes headings.@skye I made a video showing the difference with Orca, not that it's a great screen reader but it's good enough for a demonstration. Apparently my phone cannot record TalkBack's sound because google decided it's magically excluded so I only have what I got on desktop.
-
@skye I made a video showing the difference with Orca, not that it's a great screen reader but it's good enough for a demonstration. Apparently my phone cannot record TalkBack's sound because google decided it's magically excluded so I only have what I got on desktop.
@lunareclipse Yeah, it's too bad that html is usually not accepted as a document type, I will need to see what I can do
-
@lunareclipse Yeah, it's too bad that html is usually not accepted as a document type, I will need to see what I can do
@lunareclipse It surprises me though, I would have thought that at some point in the last 25 years or so that pdf has been the default document type we would have managed to make them accessible…? Clearly I continue to be too optimistic about these things
-
Hey blind people, what is the recommended accessible format for distributing text-only content? So that a short text of a couple hundred words is browsable, searchable and the structure is understandable?
Am I correct is assuming that a well-structured PDF with all the titles set correctly would do the trick? Or is there some secret file format that is much better suited?
The document is intended to be distributed as downloadable file, not as a website.
(Sighted people feel free to boost but for the love of everything please do not try to answer)
Edit: So what I have gathered
- PDF actually a terrible option
- HTML good, but not easily downloadable
- Epub probably good, and easily downloadable
- Markdown might be preferred for people that know how to work with them
- Why the fuck is this still so difficult in the year 2026
@skye honestly html file would work really well tbh
-
Hey blind people, what is the recommended accessible format for distributing text-only content? So that a short text of a couple hundred words is browsable, searchable and the structure is understandable?
Am I correct is assuming that a well-structured PDF with all the titles set correctly would do the trick? Or is there some secret file format that is much better suited?
The document is intended to be distributed as downloadable file, not as a website.
(Sighted people feel free to boost but for the love of everything please do not try to answer)
Edit: So what I have gathered
- PDF actually a terrible option
- HTML good, but not easily downloadable
- Epub probably good, and easily downloadable
- Markdown might be preferred for people that know how to work with them
- Why the fuck is this still so difficult in the year 2026
@skye as in a downlodable one in a zip file
-
Hey blind people, what is the recommended accessible format for distributing text-only content? So that a short text of a couple hundred words is browsable, searchable and the structure is understandable?
Am I correct is assuming that a well-structured PDF with all the titles set correctly would do the trick? Or is there some secret file format that is much better suited?
The document is intended to be distributed as downloadable file, not as a website.
(Sighted people feel free to boost but for the love of everything please do not try to answer)
Edit: So what I have gathered
- PDF actually a terrible option
- HTML good, but not easily downloadable
- Epub probably good, and easily downloadable
- Markdown might be preferred for people that know how to work with them
- Why the fuck is this still so difficult in the year 2026
@skye I am sighted but I used to work as an assistant for a blind student and specifically it was my job to convert text into something screen reader compatible/comfortable, so I think I can add a little bit:
Not all screen reader software is similarly capable. Also braille has no capital letters and some people use 8 (syllables) and 6 (letters) point braille, while some only 6 (letters) and others only text to speech.
PDF is a real surprise box of a format. You never know what it is really like. It is all but guaranteed that your PDF will handle semantics properly and that the different screen readers will do the same. Avoid it whenever you can.
Simpler formats are much better. Plaintext, RTF, Markdown, HTML (only text and lists, no tables, with semantic formatting).
Remember that tables work barely or not at all and less but consistent semantic formatting is more. (Sub)chapter titles are good. Many screen readers allow jumping between them, useful. Braille is slower to read than black print -
Hey blind people, what is the recommended accessible format for distributing text-only content? So that a short text of a couple hundred words is browsable, searchable and the structure is understandable?
Am I correct is assuming that a well-structured PDF with all the titles set correctly would do the trick? Or is there some secret file format that is much better suited?
The document is intended to be distributed as downloadable file, not as a website.
(Sighted people feel free to boost but for the love of everything please do not try to answer)
Edit: So what I have gathered
- PDF actually a terrible option
- HTML good, but not easily downloadable
- Epub probably good, and easily downloadable
- Markdown might be preferred for people that know how to work with them
- Why the fuck is this still so difficult in the year 2026
@skye Properly tagged PDFs can work, but depending on how the PDF is generated, getting the tagging done correctly can be anything from slightly tricky to impossible. But if you go with epub, it's pretty much the same as generating accessible HTML. A lot easier. -
@skye I am sighted but I used to work as an assistant for a blind student and specifically it was my job to convert text into something screen reader compatible/comfortable, so I think I can add a little bit:
Not all screen reader software is similarly capable. Also braille has no capital letters and some people use 8 (syllables) and 6 (letters) point braille, while some only 6 (letters) and others only text to speech.
PDF is a real surprise box of a format. You never know what it is really like. It is all but guaranteed that your PDF will handle semantics properly and that the different screen readers will do the same. Avoid it whenever you can.
Simpler formats are much better. Plaintext, RTF, Markdown, HTML (only text and lists, no tables, with semantic formatting).
Remember that tables work barely or not at all and less but consistent semantic formatting is more. (Sub)chapter titles are good. Many screen readers allow jumping between them, useful. Braille is slower to read than black print@skye in practice using markdown to get proper semantic HTML is a good workflow. There are nice free markdown editors that offer you some comfort as an author, even if you never wrote a line of HTML in your life. Also those you can send to anyone because due to the ubiquitous web there is almost no device that does not render HTML somehow
An HTML website that is just a single document is easy to download and people who navigate the world with a screenreader usually know how to do that (it is „file>save as“ or the share button on your phone if you forgot)
-
@skye Markdown. Never do PDF ever. Saving it as HTML is best too since they can open local HTML files
@WeirdWriter @skye The issue with HTML is that a lot of browsers these days will refuse to download HTML, and will just open it. For less techy people, actually saving an HTML file they got from a link can be challenging. And markdown is fine on desktop, but I'm not aware of anything on mobile that allows navigating by headings in a markdown file. Personally I'd say epub. There are free programs on all the platforms that can open epub, browsers just download it by default, making accessible epubs is as easy as making accessible html, and it will work for blind and sighted alike so you don't need to create two different documents. -
Hey blind people, what is the recommended accessible format for distributing text-only content? So that a short text of a couple hundred words is browsable, searchable and the structure is understandable?
Am I correct is assuming that a well-structured PDF with all the titles set correctly would do the trick? Or is there some secret file format that is much better suited?
The document is intended to be distributed as downloadable file, not as a website.
(Sighted people feel free to boost but for the love of everything please do not try to answer)
Edit: So what I have gathered
- PDF actually a terrible option
- HTML good, but not easily downloadable
- Epub probably good, and easily downloadable
- Markdown might be preferred for people that know how to work with them
- Why the fuck is this still so difficult in the year 2026
@skye just chiming in to say: an .epub is basically a self contained website that behaves like a book when opened in a reader. you get the semantics of HTML without the logistical hurdles of having to inline all resources (although in this case it doesn't look like there'd be many to begin with)
-
@skye in practice using markdown to get proper semantic HTML is a good workflow. There are nice free markdown editors that offer you some comfort as an author, even if you never wrote a line of HTML in your life. Also those you can send to anyone because due to the ubiquitous web there is almost no device that does not render HTML somehow
An HTML website that is just a single document is easy to download and people who navigate the world with a screenreader usually know how to do that (it is „file>save as“ or the share button on your phone if you forgot)
@skye also be aware that the overwhelming majority of blind people are those who lose their sight at very old age (think 80+). They are often unwilling or incapable of learning braille and screenreaders properly. So some advice written by non-blind and especially institutions of care or the state centres their experience. But realistically the people who work with screenreaders are mostly young(ish) people who have been blind their whole life or lost sight early in life and of course move very differently through life. Lots of things focus on the elderly who are blind for a few years before dying. If in doubt, and if you aren’t specifically offering something consumed mostly by very old people, it is better to focus on the young and their needs
-
@skye also be aware that the overwhelming majority of blind people are those who lose their sight at very old age (think 80+). They are often unwilling or incapable of learning braille and screenreaders properly. So some advice written by non-blind and especially institutions of care or the state centres their experience. But realistically the people who work with screenreaders are mostly young(ish) people who have been blind their whole life or lost sight early in life and of course move very differently through life. Lots of things focus on the elderly who are blind for a few years before dying. If in doubt, and if you aren’t specifically offering something consumed mostly by very old people, it is better to focus on the young and their needs
@skye also just to be clear in my time as an assistant the blind student was my boss and I did whatever she wanted the way she wanted it as best as I could. I sometimes stumbled upon new features/ideas and proposed/demonstrated them, but in the end the important part in that kind of work is that you very strongly centre your boss and what they want and not the ton of sighted people trying to influence them. Be they doctors, relatives, bureaucrats, professors or welfare institutions.
-
@lunareclipse It surprises me though, I would have thought that at some point in the last 25 years or so that pdf has been the default document type we would have managed to make them accessible…? Clearly I continue to be too optimistic about these things
@skye @lunareclipse i recently noticed an option to make an "accessible pdf" when you export a libreoffice writer document to a pdf. no idea what that does though and considering i only had one headline and a couple paragraphs of text with no images or anything else and it asked me to fix several things it might not be very stable / easy to use...
-
@lunareclipse It surprises me though, I would have thought that at some point in the last 25 years or so that pdf has been the default document type we would have managed to make them accessible…? Clearly I continue to be too optimistic about these things
@skye @lunareclipse PDF files can be made accessible by ensuring that they're tagged properly with semantic structure. There are multiple tools that can do this. Word, Google Docs, Acrobat Pro and others can all provide the tools you need. If you can provide some details on the types of files, I can give you further guidance.
-
@WeirdWriter @skye The issue with HTML is that a lot of browsers these days will refuse to download HTML, and will just open it. For less techy people, actually saving an HTML file they got from a link can be challenging. And markdown is fine on desktop, but I'm not aware of anything on mobile that allows navigating by headings in a markdown file. Personally I'd say epub. There are free programs on all the platforms that can open epub, browsers just download it by default, making accessible epubs is as easy as making accessible html, and it will work for blind and sighted alike so you don't need to create two different documents.
@fastfinge @skye Yeah, Epub is fantastic too far far far better than PDF
