I've ranted about frivolous patents more than once here. Others far closer to the issue do it daily, so my voice is but a drop in the ocean (and yet nothing happens).
This time it's a little different, and yet familiar (read my April post Patents versus Accessibility). Thanks to a tweet from WebAxe, I found out that somebody has patented (granted June 26, 2014) a tool that automatically makes a web site accessible. I find this a preposterous claim (at best), but here's the abstract from the patent:
A method for an accessibility solution provided as a software. The method includes approving or implementing by a website owner of a code into his website and receiving web format by a user device by “scraping” the data from the website pages or by other means of using client side plugin, server side plugin, browser or 3rd party server side or mobile app. The method also includes analyzing the data over the user device or on the server side and clicking a button by the end-user and the original code and the content and code that were collected from the website are rewritten which can also be done automatically as a suggestion to the end user, and the end-user sees or can use a new format according to the updated standard.
Unlike so many frivolous patents that are nothing more than an idea, this patent has some compelling claims. Unfortunately, the likelihood that all those claims can be realized seems pretty small.
Why I Am Wary
Just as anyone who practices accessibility understands that there is no automatic tool that can completely and truly evaluate a site's accessibility, there is also no tool that can automatically make a site accessible. To quote Jared Smith of WebAIM fame:
Accessibility is a continuum. It is based on the user experience. It cannot simply be defined as pass/fail or accessible/inaccessible.
While I don't want to attack what may very well be the well-intentioned best efforts of a dedicated person or team to improve the nature of accessibility on the web, I also have too many conversations with clients who are happy to accept a one-click solution at face value, unaware of the risks. More than anything, this post will serve as a quick reference for the inevitable conversation I'll have with a client or prospect about relying on a tool like this (or perhaps this tool in particular).
I am interested in how this tool can evaluate the value of alt text on images, verify whether an article or section is the correct sectioning element, make sure that placeholder text is not a barrier, understand when a button should be used, and so on. And then fix all these by re-writing HTML, JavaScript and CSS and not introduce more issues. These are just a tiny, miniscule slice of the kinds of decisions that require a person who understands the context in order to be successful.
The Patent Holder's Qualification
The evidence doesn't bear out that this tool can do what it advertises. The company that makes it, which promotes its accessibility services, has many issues on its own site. Which begs the question, did the company use its own tool and get such awful results, or does the company not see enough value in it to use it?
I wasn't the first to tear into the site. Folks on Twitter independently identified issues with color contrast, keyboard navigation (ok, that was me), image alt text, use of a carousel, a blank accessibility statement, incorrect ARIA role use, and misspelled ARIA attributes. All this feedback was unprompted.
To be fair, the tool may not exist yet. The patent was only just approved, though it was filed over two years ago. While I don't want to name and shame, there really is no way to make my point about the proposed tool without doing so.
The Patent Holder's Possible Example
Looking to the site of the registrant of the patent (I also made a Wayback permalink) I see an "Accessibility" box floating on its home page which may provide insight into how this patent might be applied. This "Accessibility" box allows some rudimentary control over the content. For example you can scale the text up or down, something I first built into sites about 15 years ago. You can change the text to white on black or black on white (yes, I also did that about 15 years ago). Oddly, you can convert the page to grayscale, removing all color information (it's not obvious how to turn that off, so it's still gray for me). You can toggle "alternate navigation," which just puts a big box around items that have keyboard focus (the site benefits since it wrongly uses outline:none). There is an option called "Speed control" which may adjust the rate at which its carousel/slider changes. There is an option to "Block blinking," though its affect was not immediately apparent to me.
I'm not sure that the patent would withstand scrutiny from someone who has worked in web accessibility. Some of the ideas outline, while not implemented in the tool on its own site, seem to lean on existing technologies that themselves are subsumed within patents. For example, one of the figures in the patent shows how optical character recognition (OCR) can be used to convert the text in an image into text that can be passed along to the user (figures 7a and 7b).
The Takeaway
As always with what feel like frivolous patents, I fear that they can be used to block truly useful and innovative solutions from coming to market. In this case, I worry that the already poor state of accessibility on the web could be stifled if the patent owners try to exercise any claims against those of us who've been doing accessibility work for years.
