But the hidden variables that you hypothesize here are precisely the type of
information I would go to the IETF Web site to get rather than use a more
reasonable source like Wikipedia.
In particular finding out what the latest version of a specification is.
Telling people to plug 1724 into the RFC search engine to find the POP3
specification is to put it mildly 1) an arrogant waste of the readers time and
2) gives the wrong result.
I have been thinking quite a bit about usability in the past few weeks, in
particular trying to work out how to codify some of the rules that Nielsen
gives in his book in a form that is verifiable. One approach that does seem to
be useful is to think in terms of task scenarios, similar to use cases but more
Just defining task scenarios is a big start. For the IETF Web site I would
propose that we need to consider the following tasks (amongst others):
1) Find out the status of an IETF proposal or specification
2) Find the latest documents describing an IETF specification
3) Find out how to submit a proposal to the IETF
4) Find out a point of contact for a proposal or specification
We then look at the Web site to determine whether it meets the following rules:
1) Sufficiency of information - is there enough information to complete the
2) Complexity - how many steps does a task require? how much information must
the user remember to complete it?
The IETF Web site is built to the old fallacy that minimizing the information
provided to the user is the same thing as reducing complexity. That is utter
B.S. The user gets confused and considers the problem complex because they have
too little information or irrelevantinformation.
If you build out the state table for what the user is required to do in order
to fulfill these simple tasks it quickly becomes apparent that
1) The user is required to know vast amounts of folklore. That is information
that is not provided to the user.
2) The process is unnecessarily complex
3) Most of the useful information is not even present on the site anyway, the
site you really want to go to is the tools site.
From: Harald Alvestrand [mailto:harald(_at_)alvestrand(_dot_)no]
Sent: Tue 22/01/2008 2:17 AM
To: Henrik Levkowetz
Cc: IETF Discussion; Willie Gillespie
Subject: Re: Finding information
Henrik Levkowetz skrev:
On 2008-01-21 11:24 Stephane Bortzmeyer said the following:
On Sun, Jan 20, 2008 at 03:01:24AM -0800,
Tony Li <tli(_at_)cisco(_dot_)com> wrote a message of 23 lines which said:
Or, you can google IMAP and come up with 3501 straight away...
Bad idea. Not only it makes the RFC process depend on an external
organization, but it often fails for the reasons explained by the
OP. For instance, googling Sieve does not bring back RFC 5228...
Submitting 'sieve' in the document search form on
returns 3028 as the first result, and a link to the htmlized 3028
that it has been obsoleted by 5228. (I'm surprised that the search
doesn't already include 5228, too, but I expect they will fairly soon).
The sieve WG web page also doesn't list 5228 as a product, so I guess
some routine updates are on hold while the transition is taking place.
Ietf mailing list
Ietf mailing list