I've done some searching around and I can't seem to find any command line utilities out there that will allow me to evaluate accessibility on web pages.
Essentially I want to automate the process of wget'ing a large number of websites and evaluating their accessibility.
So I would have a cron job that would get all of the necessary pages and then run the evaluation software on them. The output would then be parsed into a website ranking accessibility.
Does anyone know of anything that may work for this purpose?
Thanks!
+1 to @BrendanMcK answer ... and for the part that can (and should (*)) be automated, I know of Tanaguru and Opquast.
Tanaguru is both a SaaS and free software. Based on checklist Accessiweb 2.1 (that follows closely WCAG 2.0), it can audit pages or thousands of them. You can try it for free here: http://my.tanaguru.com/home/contract.html?cr=17 > Pages audit
I never installed it on a server, there's a mailing list if you've difficulties installing this huge Java thing
Opquast is a service that you can try for free for a one page audit but otherwise isn't free.
It'll let you audit your site with quality checklist (the Opquast one), a brand new "Accessibility first step" (aimed to problems so obvious that they should be corrected before contacting an accessibility expert) and also accessibility checklist Accessiweb and RGAA (both are based on WCAG 2.0 but I don't think that RGAA has been translated from french to english).
EDIT 2014: Tenon.io is a fairly new API by K. Groves that is very promising
(*) because it's tedious work like checking if images, area and input[type="image"] lack alt attribute ... That is work better done by computers than by people. What computers can't do is evaluating if alt, when present, are poorly written or are OK.
If only accessibility evaluation were that simple... Unfortunately, what you're looking for isn't reasonably possible.
The main issue is that it's not possible to evaluate for accessibility by programatic/automated means alone. There's certainly some things you can check for and flag, but it's rare that you can say that they are either in error or correct with 100% accuracy.
As an example, take the issue of determining whether an IMG has suitable ALT text. It's impossible for a tool to determine whether the ALT text is actually meaningful in the overall context of the page: you need someone to look at the page to make that determination. But a tool can help somewhat: it can flag IMGs that don't have ALT attributes; or perhaps even flag those that have ALT attributes that look like filenames instead of descriptive text (a common error). But if there is already ALT text, it can't say for sure whether the ALT is correct and meaningful or not.
Similar with determining whether a page is using semantic markup correctly. If a tool sees that a page is not using any H1 or similar headers and only using styles for formatting, that's a potential red flag. But if there are H1's and other present, it can't determine whether they are in the right meaningful order.
And those are just two of the many issues that pertain to web page accessibility!
The issue gets even more complicated with pages that use AJAX and Javascript: it may be impossible to determine via code whether a keyboard user can accesses everything they need to on a page, or whether a screenreader user will understand the controls that are used. At the end of the day, while automated tools can help somewhat, the only way to really verify accessibility in these cases is by actual testing: by attempting to use the site with a keyboard and also with a screenreader.
You could perhaps use some of the existing accessibility tools to generate a list of potential issues on a page, but this would make for a very poor rating system: these lists of potential issues can be full of false positives and false negatives, and are really only useful as a starting point for manual investigation - using them as a rating system would likely be a bad idea.
--
For what it's worth, there are some tools out there that may be useful starting points. There is an Accessibility Evaluation Toolbar Add-On for Firefox, but it doesn't actually do any evaluation itself; rather it pulls information from the web page to make it easier for a human to evaluate it.
There's also the Web Accessibility Evaluation Tool (WAVE); again, it focuses on making accessibility-relevant information in the page more visible so that the tool user can more easily perform the evaluation.
Also worth checking out is Cynthia Says, which does more of what could be called 'evaluation' in that it generates a report from a web page - but again its only useful as a starting point for manual investigation. For example, if an IMG tag has empty ALT text - which is recommended practice if the image is purely cosmetic or is a spacer - then the generated report states "contains the 'alt' attribute with an empty value. Please verify that this image is only used for spacing or design and has no meaning." - so it's flagging potential errors, but could flag things that are not errors, or possibly miss some things that are errors.
For other information on Web Accessibility in general, I can recommend the WebAIM (Accessibility In Mind) site as a good starting point for everything Web Accessibility related.
