W3C home > Mailing lists > Public > wai-xtech@w3.org > August 2001

Re: archive: top-level index page with adjusted link text

From: Al Gilman <asgilman@iamdigex.net>
Date: Mon, 27 Aug 2001 10:54:51 -0400
Message-Id: <200108271434.KAA7328778@smtp2.mail.iamworld.net>
To: jasonw@ariel.ucs.unimelb.edu.au
Cc: wai-xtech@w3.org
At 11:31 PM 2001-08-26 , Jason White wrote:
>
>I would be interested to know what the supposed problems are because I
>can't find them.
> 

If one were to arrive at the top level index page and tab through the page one
would have no clue as to which of the many links whose link text is "date," or
is "subject" etc. is the one they wish to visit.  This violates checkpoint
13.1, or the reason why we included 13.1, as I understand it.

   13.1 Clearly identify the target of each link. [Priority 2]
          Link text should be meaningful enough to make sense when
          read out of context -- either on its own or as part of a
          sequence of links. [snip]

The links to the sub-indices from the top level index page clearly flunk this
test because the time period to which they relate is never mentioned if one
just takes the sequence of link text phrases.

In the message page, the issue is more subtle.  Take for the example the link
to the next message in time order.  Here the destination is clear enough, but
the motivation is in the relationship which is stated immediately previous to
the link.  If one tabs to this link, and is not sure if they wish to visit it,
then shift-TAB, TAB does not solve the problem, but up-arrow, down-arrow [I am
guessing the keystrokes, but the latter is for reading the normal or buffered
mode chunks that Kelly depicted as lines in his rendition of the interface]
does solve the problem.  This to me puts the message page issue in the
definite
maybe category.  But on the top level index page to repair the confusion when
you stop on an 'author' link you have to up-arrow back three times to get to
the time-period information, regain orientation, and then move on ahead.
This,
to my mind, is demanding too much understanding of the macrostructure of the
information from the user.

If you already know that there are a sequence of time periods and four
sub-indices for each time period, of course you won't have trouble reading the
tea leaves if you arrow through [not TAB] and the date and then the sorting
flavor step by in your ear.  But if you don't have that pattern in mind or if
you TAB through the links, it is quite meaningless.

That's the general idea.

The rest of the message is a more discursive ramble which may be skipped on
first reading.

There is an old joke about mathematics, how a given notion is perceived either
as intuitively obvious or totally incomprehensible.  The story goes on that it
is possible for a particular notion to pass from the incomprehensible to the
obvious state as you learn, but that it doesn't seem to pass through any
intermediate states, it just snaps in place.  And after that, we might add, it
is soon impossible to recostruct the 'incomprehensible' perception.

Browsing these archives is a little too much like that story about
mathematics.  If it all makes sense, it all makes sense.  On the other
hand, we
should make the archives so that they are operable if you only half understand
what you are doing.

And furthermore, there is a general principle that I think we will need to
understand and honor on the way to a device-independent,
accessible-by-construction plan for building the Web.  This is that one wants
to fix problems, when they occur, as locally as possible.  I was reminded of
this by the way the AT industry has largely ignored our flying leap to the
headers with the HEADERS attribute, and even worse the possibility of
inverting
a SCOPE attribute.  Things that work locally with more kinds of elements are
going to succeed, both with the consumer and with the AT developer, before
special purpose things that involve more of the context before they work
right.

Al 
Received on Monday, 27 August 2001 10:34:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:15:36 GMT