seeing the affect of checkpoint criteria and tests on real sites

We were discussing yesterday testing site, how many sites to test before we
know how reliable a test or affect is. (I mentioned testing for computer
components but that was probably not the point.)

When UB was  making the test plan for SWAP on  whole sites what we tested
separately on different  verticals, in other words types of sites, because
what typical code and components on one type of site is vastly different
content to a different type of site. For examples, you could take the
vertical of middle sized law firms, and you would find very different
behavior and use of mark up to, say, middle sized law firms, or large sized
banks  or universities etc....
We found the magic number to be a bout five sites per vertical to get to
about 90% of the issues.
In a way a commercial company has it easer because we did not have to
segment all the types of sites out there, only the ones that were relevant
for our marketing. So , for example we did not test five blogging sites (
Middle sized, small or large...) but i would assume typical blogging HTML
and accessibility issues to be very different form bank law firms or
universities.....

So while the conversion of test affects are very important  it will be a lot
of work by and getting them stable is going to be a lot moving target. And
the only way to get anywhere close will be by using a lot of testers -in
which case we should be asking the interest group to help go over the tests
and tell us what

I would still like to see on the test pages, even the drafts,  the
disclaimer that we discused. Some thing along the lines of, "these test have
been developed as a useful tool to help people find and understand
accessibility violations. They do not guarantee accessibility, and user
testing is always recommended. "


Keep well and all the best,

Lisa

Lisa Seeman
UB Access
Tel: +972-2-648-3782 (please note our new number)
Website:   www.ubaccess.com

THIS E-MAIL CONTAINS CONFIDENTIAL INFORMATION AND IS INTENDED FOR THE
RECIPIENT OF THIS E-MAIL ONLY.