On 6/8/2014 12:23 AM, Franck Martin wrote:
> I think we need to give advice to MUAs, while letting MUA developers
> some liberty on how to interpret it.
> 
> I'm proposing the following text to be added to the DMARC spec

Directives encased in standards that concern UI/UX design are
appropriate when there are empirical data to support them.

On the average, efforts in IETF-related discussions at making UI/UX
design assertions lack that empirical basis and often even run counter
to existing empirical data.

Often they represent using the speaker as the user exemplar or cognitive
models along the lines of "more information presented to the user is
better".  The former is bad methodology and the latter is known to be
wrong.  Think bell curve...

UI/UX design is a specialty.  Few of us have the necessary training or
experience.

A hallmark of those who do have that training and experience is knowing
that world-class expertise only produces a useful /starting/ point in
design, but that the /finishing/ point is rarely the same.  What is
required to reach that finishing point is testing of the design.

I've been making that above, two-sentence summary for many years, but
finally tested with a group of real UI/YX experts at the SOUPs
conference last year and they nodded yes. (It's the per-eminent 'usable
security' conference. Last year was England; this year is Menlo Park, CA.)


> "MUAs SHOULD display to the end user, in UTF8 (and punycode), in a non
> ambiguous font, the domain used for the assertion of the DMARC policy,
> as well as the result of this assertion. A non ambiguous font is a font
> where the graphical representation of a character is not identical to
> the graphical representation of another character in the same font"

This is an example of a directive for a design choice that has
demonstrated very poor utility.



d/
-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net

_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to