Replies: 12 comments
-
would it be possible and used by screenreaders when adding mathml directly in the document within a According to PDF/UA accessibility guidelines, “All mathematical expressions shall be enclosed within a Formula tag and shall have an ALT attribute.” |
Beta Was this translation helpful? Give feedback.
-
Well you can add mathml tags. But this is much more difficult to automate (probably even impossible with pdftex) and also much more difficult to adjust if needed. Also even if we manage to add such mathml tags: I have doubts that pdf consumer will use them in the near future.
well you naturally can put a mathml into the alternate text, but if and how a screenreader would make use of them is not something we can decide. Currently they imho do not parse the alternate text to detect mathml and no standard requires from them such a parsing.
That is PDF/UA-1. PDF/UA-2 says
The idea to mark up formulas with alternate text stems from the time when math were mostly given as images. With mathml and unicode it is no longer the right way. See also my issue here pdf-association/pdf-issues#470 |
Beta Was this translation helpful? Give feedback.
-
thanks for getting back that fast and pointing to your issue in pdf-association repository pdf-association/pdf-issues#470 |
Beta Was this translation helpful? Give feedback.
-
If I understand you correctly you think that the Associeted Files concept offer an attack vector by putting malicious conent into such streams that are attached as AF. I would have expected that the the PDF implementations guard against that but if not it is not something we can control or influence from LaTeX other than by putting MathML or tex source into them. Even if we don't use AF ourselves anybody could then write bad code into an AF and put it into a PDF and thus that would need to me taken up with the PDF Association (PDFA) perhaps, but for that one would probably have to prove first that there is a possibility to misuse the concept in the first place. Anyway, I don't think that it is something that affects what we do --- if there is an attack possibility it should, of course, gets fixed but once that happened any of the approaches we consider would remain a viable possibility. |
Beta Was this translation helpful? Give feedback.
-
yep having files within a pdf is a well known security problem. My point is, that because of this,
This is where PDFs with this kind implementation of accessibility features (file attachment) maybe damaged or sorted out by the security measures taken. So the idea/question is: cant we have accessibility for math formulas in a way without opening the possibility to integrate downsides like the chance of being branded as security risk. |
Beta Was this translation helpful? Give feedback.
-
Another field where not "if" there is an embedded file but "how" it is embedded seems to be important:
https://pdfa.org/resource/pdf-2-0-application-note-002-associated-files/ |
Beta Was this translation helpful? Give feedback.
-
We know that. I'm reading PDF-specs for more than six years now ;-). Our files all have a mime-type and a AFrelationship key. BTW: for pdf 2.0 you should check PDF/A-4 not PDF/A-3. |
Beta Was this translation helpful? Give feedback.
-
@hpvd if you look at the examples at #72 you will see that we've annotated each by the standards against which they have been validated and all are valid to PDF/A-4 or PDF/A-4f |
Beta Was this translation helpful? Give feedback.
-
@davidcarlisle many thanks for pointing to the examples! |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
@hpvd you may prefer the PDF 1.7 PDF/UA-1 versions #82 although math accessibility is generally much worse with PDF 1.x. It is well known that at the current time pdf reader support for PDF 2.0 is sadly lacking. An explicit aim of the project is to build up a large corpus of PDF 2.0 files to put some pressure on PDF readers to support the standard. At the current time it's not unreasonable to use PDF 2 for testing or for specific accessibility needs and PDF 1.7 PDF/UA-1 for actual production work. If no one builds PDF 2 files, readers will never support it, so you have to break the cycle somewhere. With development builds of foxit and NVDA/Mathcat you get good screen reading and can navigate directly in to the math expressions from the screen reader, which is a glimpse of how it is supposed to work. |
Beta Was this translation helpful? Give feedback.
-
We are currently adding the files to the embedded files pane, but this is mostly for debugging and to allow people to see the mathml. There is an option to switch that off, and that is what we will probably do later by default as it doesn't make much sense to show hundreds of mathml-files in the pane. That sumatra shows embedded files as bookmarks is clearly a sumatra issue (or feature). |
Beta Was this translation helpful? Give feedback.
-
I have watched your great talk about progress of pdf tagging at TUG 2024
https://www.youtube.com/watch?v=b70CsLh7qU0
You are describing the accessibility tagging for math works via attaching small files including mathml inside.
https://www.youtube.com/watch?v=b70CsLh7qU0&t=1301s
I'm just wondering about the security topic.
Files in / attached to pdf are generally known to be a potential security risk because they can contain malicious content, open other dangerous files, or even launch applications.
Some companies therefore clean all received pdfs before open them...
which damage accessibility and positive findings maybe also lead to blacklisting and block the pdf file...
=> Having this in mind , is there anything one can do in this field?
Are there other ways to make math accessible?
Or can one think of things to tightly couple pdf and mathml files e.g. via signing them? The idea is that it is a least not easily possible to exchange a "good" mathml file with a "bad" actor ...
Of course this is not a latex topic, but we may use the chance to choose the most future proof direction
... security measures taken are constantly rising and pdf-accessibility should not be a blocker for accessing a pdf :-)
Beta Was this translation helpful? Give feedback.
All reactions