-----Original Message-----
From: DCMI Accessibility Community [mailto:[log in to unmask]]
On Behalf Of Patrick H. Lauke
> [1] http://searchthresher.com
Looking at the source code for the extension shows that, for the time
being, it only actually does anything when doing a Google search (based
on the function below)
[PW] Yes, proof of concept which is used to demo to the likes of Microsoft
to encourage them to pay for the build of an IE plugin. Support for other
search engines will be included shortly. We haven't put it into Mozilla labs
for all the reasons you put forward.
isValidSearchSite : function (aHref) {
var name = aHref.match(/^http:\/\/www\.google\.[a-z\.]+\/search/i);
if(name){
//name[0].toLowerCase()
return true;
}
The output that the extension overlays on top of the Google search
results is, ironically, not keyboard accessible (the icons can be tabbed
to, but not activated other than with a mouse/pointer).
[PW] Thanks. We haven't had time to update the site so this is something
we'll need to look at. We should document known issues so we don't waste
your time.
At this stage, due to the not very widespread use of the trustmarks, it
usually happens that one does a Google search for verified or
self-labelled sites only to be presented with an empty results page
(empty because the extension has removed the non-verified or
non-labelled search results from the page after it's been sent by
Google), but with the results page counter/links at the bottom still
there (as the extension only transforms the results one page at a time).
[PW] Yes we know. This is easy to fix. Also, the extension reads other
labels such as the partners on the Quatro programme -
http://www.quatro-programme.org There are more compelling use cases for
Content Labels in the pipeline.
Lastly, in order to look for the potential presence of trustmark
metadata, the extension executes a full page fetch for each result
(potentially messing up stats collection on those pages).
[PW] The extension checks to see if there's a link tag in the page. If it
has the appropriate name space etc. it'll return the relevant icon. It's
only when you click on the icon, (or in future, check preferences) does it
go looking for the Content Label. You can pretend that you've been
independently verified but the extension recognises that there's no
legitimate Label and as such, warns the users before they enter the site.
It's a good idea, but I'd say it's something that would be far better
handled at the server/search-engine end of things. As a proof of
concept, though, it's interesting. And the concept of trustmark as
RDF/metadata is far more appropriate than the original "Segala approved"
trustmark (which was effectively just an image with a link to the
certificate on the Segala site) I remember seeing back in 2006 on five
sites (two Segala and three O2 sites). Once/if site owners actually
start using this new trustmark metadata, the idea could prove useful.
[PW] There are other sites. However, we've been spending 98% of our time in
the lab re methodology/processes and other stuff such as w3c MWI - with 0
time dedicated to sales. The partner programme includes agencies and
freelance developers across 6 countries and we haven't tried to build it
yet. The idea is to enable them to certify and label sites. Same business
model as VeriSign. Around 2 years has gone into this, so we don't intend on
diluting the brand before it has had time to get started. So you can rest
assured that our processes are pretty tight (although not perfect I'm sure).
As you say, visual badges and certificates have very limited benefit. So do
SSL Certificates. I've demonstrated this to Philip Hallam-Baker amongst
other equally qualified people and they think its compelling - Content
Labels that is. The extension is an extremely small part of the puzzle to
help search engines and browsers better understand how they can enable more
trust using meaningful data. It's in no way a commercial product. It will be
subjected to the usual open source community review just like any other
extension.
As you point out Patrick, it's a proof of concept, but once we start
development again it'll move forward pretty quickly. What's there was only
built in a couple of days (about 5 months ago) to 'show' how it 'will' work.
I just thought it was a good idea to show the group so you'd know what's out
there. I'm a member of the Semantic Web Education and Outreach programme and
intend to use Content Labels as a good use case. There's no reason this
group's metadata shouldn't be in there.
Thanks for the feedback!
Paul
