On Tue, Oct 13, 2009 at 11:01:41AM -0600, Clint Savage wrote:
On Tue, Oct 13, 2009 at 10:34 AM, Mike McGrath
<mmcgrath(a)redhat.com> wrote:
>
> I'm not actually sure, here's the app we're using:
>
>
http://code.google.com/p/tgcaptcha/
>
> It's a pretty easy change to start using Vanasco-Dowty but it gets tricky.
> neither one is particularly easy to read (which I guess is the point)
>
> -Mike
Actually the Vanasco-Dowty is *much* easier to read (from what I see
on their google code page). I can read the McDermott CAPTCHA, but it
takes my eyes more effort and I had to squint a little. This is just
my experience, but I think the Vanasco-Dowty would be a good
compromise.
Bodhi switched from Vanasco-Dowty to McDermott because Vanasco-Dowty was too
hard for people to use. I remember playing with it and I think the issue
was that Vanasco-Dowty is much harder to figure out uppercase and lowercase
than McDermott. Note that that's not something you realize by reading the
documentation -- that's only something you realize by having your captcha
rejected because you typed "CIKLOPSUVWXZ" instead of "ciklopsuvwxz".
I'd love to see someone create a new captcha algorithm that was easier to
use. I'd work on it myself but I don't know what a computer can or cannot
read.
-Toshio