Purdue Mathematics IT

There are at least two web accessibility guidelines: Section 508 from the federal government and WCAG from the W3C .

What’s an quick, simple, cheap method for testing or validating web accessibility?

Their are a number of tools available that attempt to provide automated accessible testing:

• Firefox Plugins
  • Fangs – a screen reader emulator
  • Firefox Accessibility Extension – from uiuc.edu
  • TAW3 – uses tawdis.net and the WCAG 1.0
• Validation services:
  • webxact
  • Cynthia Says – can select section 508, or WCAG 1 through 3
  • FAE – Functional Accessibility Evaluator
• Testing
  • VisCheck – simulation of web page experience for the color blind

These tools are a great place to start and by passing these tests your website will be usable by screen-readers, braille printers and other accessibility tools. That said, making an excellent site really cannot be automated. Improving accessibility includes things such as descriptive links, content that doesn’t require a user to visually scan a page, using the <abbr> tag, useful description of images, labels, allowing text to be resized (use em not px), etc. The W3C provides an excellent article entitled: How People with Disabilities use the Web.

Accessibility software:

• Kurzweil 3000 – “the comprehensive reading, writing and learning software solution for any struggling reader, including individuals with learning difficulties, such as dyslexia, attention deficit disorder or those who are English Language Learners”
• JAWS – Windows screen reader
• Dolphin Hal – Windows screen reader
• Orca – gnome’s visual assistive technology
• Sun’s Solaris 10 includes much built-in accessibility functionality, including a screen reader for gnome
• Apple’s Mac OS X contains tightly integrated accessibility features

What about PDF’s?

Isaac Vetter – Webmaster

Every so often I find a need for a small little web service type widget.

I leave many of them open to the world, others are restricted to just .purdue.edu, or maybe just .math.purdue.edu.

Here’s a few of them:

• “What’s my PUID?” – enter your Purdue Career Account username and password and get back your PUID and name (from i2a2).
  • www.math.purdue.edu/resources/whatsmypuid
• ip2country – Checks your ip against an incomplete database of ip ranges and the country listed (with cool country flags.)
  • www.math.purdue.edu/resources/ip2country
• PDFtoHTML – Converts and hosts a pdf file (that’s currently on a purdue.edu url) to an html file. PDF images are converted to .png’s.
  • www.math.purdue.edu/resources/pdftohtml

If you want to use something or need something changed, let me know.

Isaac Vetter

Webmaster

Do sites using Google Analytics (GA) allow Google to correlate GA traffic with user’s search/gmail/maps/doubleclick/etc traffic?

Google Analytics is a web traffic analysis service offered for free by Google to website owners. The owner of a site merely needs to register with Google and place a snippet of javascript on each web page. Here’s an example of that snippet:

<script src="http://www.google-analytics.com/urchin.js" type="text/javascript">
</script>
<script type="text/javascript">
	_uacct = "UA-555555-1";
	urchinTracker();
</script>

As one can see from the above code, a javascript file on google.com is included, the site owner’s identifier (_uacct) is defined and then the urchinTracker() function is called.

This function places four cookies on the end user’s browser. It’s important to note that these are first party cookies, meaning that the domain associated with the GA cookies is the current website and therefore, these cookies can only be read by the current website. This means that if x.com uses GA, and a user visits x.com, and then visits google.com, google can not connect the user’s analytics traffic from x.com with the user’s google.com visit using the unique identifier in the cookie.

That said, Google could connect those two visits with an ip address or ip and user-agent string. This is less reliable that the cookie, but still possible.

These four cookies are:

• __utma
  • This cookie expires in year 2038 and contains a long string that looks like this:
    • 2252076.984943658.1189608517.1189608517.1189610067.2
  • The first number (2252076) is a hashed value of the current website’s domain. The second number (984943658) is just a random number. The third, fourth and fifth are timestamps of different events. The third records the date/time that the cookie was first written. The fifth contains the most recent date/time that this cookie was edited (this cookie is written to every time __utmb or __utmc is written, i.e. every 30 minutes or anytime the browser has been closed and reopened). The fourth contains the previous value of the fifth. The sixth value is an integer that records the number of times this cookie has been written.
• __utmb
  • This cookie expires thirty minutes after last visiting a tracked page for the current website. (This is how GA defines a session.)
• __utmc
  • This cookie expires when the browser is exited.
• __utmz
  • This cookie expires in five months (?) and contains another long string that looks something like this:
    • 2252076.1189608517.1.1.utmccn=(direct)|utmcrs=(direct)|utmcmd=(none)
  • The first number is a hashed value of the current website’s domain.
  • The second number is the timestamp that this cookie was written.
  • The utm (Urchin Traffic Monitor) fields, utmccn (click content), utmcsr (click source), utmcmd (click medium), and the optional utmctr (click term), and utmcca (campaign). Each of these fields can be set by adding HTTP GET parameters, like so:
    • http://www.domain.com/
      • ?utm_source=sitethelinkison
      • &utm_medium=organicORcpc
      • &utm_term=successfulsearchtime
      • &utm_campaign=AdCampaign
      • &utm_content=

If you wish to specifically stop website owners from aggregating details of your use of their website, a good firefox extension is CustomizeGoogle.

Here are some interesting pages about Google Analytics:

• Want to Hack Google Analytics? Simple!
• Here’s a rarely used fix to the above method.
• Enhancing Web Analytics Tracking
• Testing Keyword Landing Pages

Isaac Vetter

Webmaster

Department of Mathematics

Purdue University

www.math.purdue.edu/help

Email clients, like Mail.app and Thunderbird, like to know that they’re the only program accessing email stored on the local machine. This avoids potential email file corruption.

Sometimes, however, these email clients get confused and insist that another program is writing to their files. If this happens to you, you can assure Mail.app that its files are okay by removing a lock file.

From the command line, run this command:

rm ~/Library/Mail/Envelope\ Index.lock

From Finder, navigate from your home directory, to Library, to Mail and drag the Envelope Index.lock file to the Trash.

Of course, sometimes Mail.app files really do get corrupted. If you continue to experience problems, you can reset Mail.app and wipe out all of its settings and start fresh*.

From the command line, run this command:

rm -rf ~/Library/Mail ; rm ~/Library/Preferences/com.apple.mail.plist

Now open Mail.app and (re)create your mail account.

* Sometimes a fresh start is a bad thing. If you use POP to grab your email from the server, you should not do this. If you have spent lots of time writing client side email filter/rules, you should not do this.