Eliot Lear wrote:
I strongly disagree. Many IETF protocol specifications have the user
interface in mind from the get go. This is particularly true in mail.
Take the simple construct of the Subject: line. It is there because
Very, very little IETF work has had user interface in mind. Some work has
user "information" in mind, but that is quite different.
The example of email is restricted to very, very well-established models, such
as the office memo. Hence it was sure to plug into human communications quite
That's not true for anything like what is being discussed here.
Netnews is the same way with the Newsgroups:
huh? What is the basis for believing it is useful to the general population of
users? (I don't mean what was intended, I mean knowing that it is broadly
construct. That is there because users are interested in specific bits
One of the challenges in having human factors discussions in geek circles is
the need to be very careful to distinguish average, mass market users from
geeks. This is an issue of defining the population properly and sampling it
For quite a few critical factors, we ain't nuthin' like them.
No human factors experts were involved in the creation
of Email or Netnews protocols. I know. I was there for the latter. SSP
There is a difference between what the engineers designing something think
will be useful, versus what actually will be useful.
Designing user interface details is a sufficiently heuristic process that it
is essentially impossible to predict exactly what will work. Expertise can
produce an initial design that has lower risk, but not one that is guaranteed
to be acceptable. Only testing of the construct can do that.
(I recently had the fun of being a subject for a design at a well-known apps
company. They have a massively successful track record. So imagine my
surprise at how unusable much of their prototype was...)
I'll refrain from launching into a lecture about cognitive load,
attention/distraction, and the rest of that domain. It's not stuff we deal
with in the IETF but it is fundamental to usability design.
does NOT tell applications what to display or how to display
information, but rather makes basic observations and conclusions about
behavior of users and spammers that we see today. That is: users look
at From lines and spammers and phishers try to fake them. Anyone
DISAGREE with that assertion?
Users mostly look at the display string of the From line. Users are generally
lousy at distinguishing particular domain names and matching correct ones to a
brand, versus cousin domains. DKIM and SSP provide no protection against these.
NOTE WELL: This list operates according to