[Auth]Re: Auth digest, Vol 1 #29 - 7 msgs

[Ron Burk]

> Why restrict the information to a set of standard information fields?

I see what you're saying in general. I think one answer is
convenience for the user. I only want to have to type my
credit card number in once, not once for every site that
requests it. Likewise with shipping address and so on.
I don't know how to do that without having both client
and server agree on the names of things.

I also see this as sort of the crux of the "let's automate
the Web movement". All the stuff about agents and
the semantic web and the like has largely this one sticking
point -- people have to agree on the names for and meaning
of particular bits of data. If it's not possible to get widespread
adoption of such agreement for the relatively small set
of personal data that most users provide to web sites,
then I think all the grander schemes have little hope.

OTOH, there's nothing to prevent existing form filler
software from continuing to do exactly what you're
saying, in addition to supporting a dotGNU method.

Also, now that I've seen how many vendors launch
the browser to the desired site rather than integrate
with the browser, I'm feeling it's important to get the
possible mechanisms a bit more nailed down for
how the web site gets its information request to the
client software. Maybe it's possible to make a relatively
simple Windows DLL that acts as an intermediary
between IE and any third-party password tool. E.g.,
the DLL gets loaded by IE, and the third-party tool
(they often are running in the system tray anyway)
registers with the DLL to get a callback if the browser
encounters a request for dotGNU-type information.
Anybody on this list got experience coding software
that hooks into IE?

Ron Burk
HighTechInfo.com, www.hightechinfo.com