Last year I was involved in the organisation of the Western Australia Web Accessibility BarCamp, a great event with many excellent speakers, a great vibe with around a hundred people in attendance and lots of great information on accessibility.
As part of the event I organised a mock debate titled ‘PDFs are AWESOME!!’ in which the affirmative side struggled semi-intentionally to make a good case for the use of PDFs, while the negative side had plenty of ammunition about why PDFs should be avoided at all costs. Unsurprisingly the entire room voted the negative team the winner and the affirmative team weren’t particularly disappointed.
While the mock debate was a lot of fun, genuine debates over the merits of PDFs and their accessibility often result in heated discussion in ways which don’t seem to be the case for general accessibility discussions.
I’ve attended conference presentations, including one by Adobe that claimed that PDFs can be made very accessible, which resulted in fierce heckling from the audience and accusations of lies being used to sell product.
Yet, several presentations which followed on how we can make web content accessible were met with no scepticism, no shouting, and the general feeling in the room was one of appreciation and an understanding of the needs of people with disabilities.
So, why is it that when we talk about making PDFs accessible there’s criticism and anger, yet when we talk about making the web accessible we’re enthusiastic and supportive?
A good place to start is the conversation that took place in the WAI Interest Group (WAI-IG) in January around PDF accessibility. For regular readers of the column, you may recall that the last time I talked about something in the WAI-IG, The ‘100% accessible website’ joke—do web accessibility professionals have a sense of humour?, it resulted in some passionate commentary of its own, but I’m going to risk it again as I think it helps to highlight the point, and all WAI-IG emails are publicly available [link is external].
PDFs and signatures – it all started so innocently
In January 2015, a post was made to the WAI-IG by Sarah Kane asking about how WCAG viewed putting a signature in a PDF document. The question was framed like this:
Though it seems to be not recommended to add signatures to PDFs on a public website, I have not been able to find if it’s within WCAG 2.0 guidelines to include a signature either by adding an e-signature or inserting an image of a signature with alt text. Any suggestions or resources?
Personally, I thought at the time this was a really good question. While there are techniques relating to PDF accessibility provided by W3C [link is external], some of the nuances of PDF files can make it tricky to know where they fit. I was really interested to see how the conversation would play out.
The first reply to the post was as follows:
In a strict sense, WCAG 2.0 doesn’t apply to PDFs, but rather to web content in the sense of HTML (and dynamic coding, etc.) pages. Having a signature on any form of web content is outside of WCAG 2.0’s purview, as a signature wouldn’t in itself be an accessibility matter- the means by which a signature is presented, is.
My first reaction to this was, “But what about the WCAG PDF techniques?” The next post seemed to agree:
I have to disagree slightly with Andy’s claim – PDFs *are* covered-by and included in WCAG, and in fact there is a whole section of Success Techniques provided by the W3C. Please see: http://www.w3.org/WAI/GL/WCAG20-TECHS/pdf.html
Post number four highlighted why this was starting to become an issue:
To a degree, there are many facets of PDF that WCAG does not cover. For a complete understanding of PDF accessibility it is necessary to look to PDF/UA, the ISO standard for accessible PDF.
All these emails also had a bit more discussion on the actual signature issue, and if it were an issue relating to HTML rather than PDF, I suspect it would have stopped there. Instead, the subject line was changed—“PDF accessibility guidelines. WAS: Re: PDF’s and Signatures”—by the next post, which tried to share some direct experiences relating to the original question, but the thread quickly exploded into 50 posts over the course of a week.
Rather than focus on the original issue, the conversation migrated into a space where one person said that their recommendation was that PDFs are not accessible at all, another stated that WCAG doesn’t really address PDF accessibility issues, and others argued that PDFs can be accessible if the right tool is used, particularly if designed with PDF/UA in mind.
The issue then became about having to pay for the PDF/UA document, because ISO (International Organization for Standardization) charges for its standards. Universities were then blamed for not putting accessible PDFs online, moving on to this fiery statement:
No matter how much PDF/UA or ISO 32000-1 are prettified with access capability, PDFs remain as Jakob Nielsen described on https://web.archive.org/web/20030716054032/http://www.useit.com/alertbox/20030714.html more than a decade ago – designed for printing, not reading on a PC screen. As most are presented, they are a web accessibility plague exceeded only by Flash, if even that.
The conversation then progressed to IBM moving away from PDF to ePub, the need for office suites to improve PDF tagging functionality, and someone’s church newsletter. While there was the odd post that tried to steer the conversation back to the original question about signatures, it was too late, with criticism of other portal document formats, Adobe conspiracy theories, the impossibility of getting anyone to take PDF accessibility seriously and things getting a little too personal. Eventually W3C’s Shawn Henry called it a day with the following message:
Thank you to those who kept the discussion professional and constructive.
However, some recent comments are not appropriate.
I think most further exchanges would be best privately, rather than on this list.
As a reminder, parameters for the WAI IG mailing list include:
* E-mail messages should contribute to a comfortable, constructive exchange of ideas to advance web accessibility. Before you write an e-mail, think about how it will come across to the hundreds of people around the world who are on the list.
** Be respectful. Be positive. Be open to different perspectives.
** Provide constructive information. Focus on developing solutions, rather than just complaining.
** Avoid personal attacks, naming calling, and such…
* When your comments are not useful for the entire list, reply off-list — that is, reply directly to the sender and do not send the message to the list.
It’s a rarity for a conversation to get to a point where it’s necessary—and I agree it was necessary—for Shawn to pull the plug on a WAI-IG conversation. And while it may be rare to break up a thread in this space, it’s not unusual to see this conversation unfold when PDF accessibility comes up in conversation, wherever that might be.
Argument for the negative
To return to the mock debate I mentioned at the start, there are some good reasons as to why PDF documents are challenging. As a legally blind person relying on assistive technologies, I can’t help but cringe when I discover that information is only available in a PDF. Not because it’s necessarily going to be hard, but because it’s inconsistent.
I’ve had circumstances when I’ve received a PDF which can only be viewed in a specific version of Adobe Reader, and my other PDF viewing software just gives a ‘please wait’ message.
Sometimes it might be scanned and I’m going to have to play with the colours in my screen magnifier to view it. Other times my high contrast colour scheme will create a black background and black text scenario which makes the document impossible, and then there are times when it works perfectly. Knowing that, in all likelihood, the PDF sitting in my inbox is going to take a while to read makes it difficult to plan my day, especially at work.
The second thing is the lack of native support. Yes, there are many, many PDF creation and viewing tools, and even Word allows you to save to PDF these days, but to use the file you’re always at the mercy of the viewer. For example, when using a screen magnifier I find it difficult to scroll through the PDF and scroll across the screen with the magnifier. With Word documents you can format the windows, resize the text and do many other things to make the two work together. It’s not impossible to do this with a PDF, but it’s much harder.
Thirdly, moving outside my personal assistive technologies, is the point made in the WAI-IG thread that the WCAG 2.0 guidance is helpful, but not definitive. There are always circumstances where PDFs are created using an unexpected source. One example shared with me recently is the use of PDFs for archiving rare text and images. PDF’s archive extension is absolutely great in ensuring that the document can still be used as time goes on, but what’s the chance of that document ever being accessible? It’s a tricky issue.
Finally, there are scanned PDFs which aren’t done for any specific archiving purpose, but because it is convenient. An example I was given at a disability conference recently was around job advertisements. For a job description form to go online, it’s not uncommon for it to be written in a Word document, printed out, signed for approval, scanned with the signature and then published online as the completely inaccessible document it has become. This is by far my biggest frustration with PDFs because without optical character recognition (OCR) software a number of people with disabilities will have absolutely no chance of using it—not because of any deliberate attempt to create problem documents, but because the process was so easy.
Argument for the affirmative
Before I descend any further into PDF-bashing, it’s important to acknowledge that the format actually has many benefits and this can include accessibility— a brave statement, I know. The first important thing to consider with PDFs is that they are with us and are here to stay. No matter what people may think of them in terms of accessibility, they are used across the world all the time and serve many useful purposes, most notably for getting a lot of information into a small file that can be viewed and printed consistently. While it’s not the only file format with such properties it’s one of the most popular, so we really need to find ways to make it work.
Secondly, as indicated in the WAI-IG conversation boards at beginning of this article, we do have help when it comes to PDF accessibility. While the WCAG techniques aren’t perfect and may not cater for all situations, they provide an excellent guide, and for most PDF creators they will help turn an inaccessible PDF disaster into a usable document. This resource is getting better all the time and so we can look forward to ongoing PDF improvements, which has to be a good thing.
Thirdly, there’s PDF/UA. This standard is likely to result in new PDF authoring tools continuing to improve accessibility, and perhaps we’ll reach a point in the near future where most PDFs are created accessible by default, not because the user necessarily tried to, but because the tool made it happen. While we’re not there yet, it’s very possible.
Another improvement is the accessibility of other file formats which spill into accessible PDFs. Having accessibility checks in products such as Office 2010 and 2013 provide additional opportunities for source documents to be made accessible, meaning that if these documents are in turn made into a PDF there’s more chance that they will be accessible. It’s not a perfect process but it helps a lot.
Finally, there’s a lot of innovation out there in the accessibility community— if we can stay civil to each other long enough—that can greatly improve the accessibility of PDFs.
To return to the archive example in the negative, I’m aware of at least one museum that has a lot of scanned PDFs, but it endeavours to address the accessibility issue by creating an accessible abstract for each item or a sample of the text, and then works to provide a full transcription of the scanned item on request.
While it doesn’t make the document accessible, and there is a delay in getting the information, it ultimately results in the user being able to gain access to the information. And the requests come in at a speed manageable for their staff rather than trying to convert hundreds of thousands of PDFs, some of which involve rock painting images in ancient dialects—a virtually impossible task on a large scale.
PDFs are whatever you make them
So balancing out the different arguments for PDFs, I come back to the first statement of the affirmative case: PDFs are here and they’re here to stay. The time for arguing over whether or not we should be using them has passed as the world has embraced them, and ultimately the accessibility community needs to as well.
As we all come together to improve the techniques we use to make web content accessible, we need to also continue finding new ways to make PDF accessibility not just better for the end user but easier for the creator, to ensure that accessibility is built in.
PDF/UA is a good guide but few would argue that we can’t do better, so I’m hoping that the next accessibility conference I attend will lead to a much calmer and more collaborative approach when the inevitable question about PDFs comes up.