I think all five criteria below are correct. I also believe we will meet
all of them in our next release (in testing) of TrustBar, and meet
almost all even in our current release (which has many downloads, happy
users). Here are details:
Heikki Toivonen wrote:
Ka-Ping Yee wrote:
1. We want an antiphishing tool that does not transmit a record
of the user's browsing activity.
Absolutely. And yes, several commercial tools do, at large cost to
privacy and even performance hit.
2. We want an antiphishing tool that occupies modest or minimal
screen space.
Agreed, and also admit that this is not so true for TrustBar 0.3.1.
Fixed in version 0.4.
3. We want an antiphishing tool that is deployable without
requiring major changes to server security infrastructure.
Any short term solution will have a requirement that says: no server
changes required. Long term everything is possible, but the less changes
the better, of course.
Agreed and of course TrustBar requires no such change in server.
I think a fourth point is required as well:
4. No (or minimal) input from user.
Agreed; and in fact, I believe `provide useful function even with no
input` is actually a good goal, and we meet (even) that.
Current SSL system generally requires no input from user (exceptions are
when some problem with the certificate the server presents). petname is
an example where input is required for every SSL-enabled site the user
visits more than once.
Not with TrustBar, which displays the name from the certificate (and the
logo of the CA) by default. We do allow user to change the name/logo
for the site (i.e. use `petname` model) but this is optional.
And perhaps another point should be explicitly mentioned:
5. Easy to use.
You could elaborate 5th a lot: trivially easy to use, idiot-proof, fail
safely, ...
Our usability experiments show TrustBar meets this as well.
Best, Amir Herzberg
_______________________________________________
Mozilla-security mailing list
[email protected]
http://mail.mozilla.org/listinfo/mozilla-security
