Re: Proposed improvements to the mailing lists archives

At 05:04 AM 2002-03-29 , Dominique Hazaël-Massieux wrote:
>le ven 29-03-2002 à 05:12, Al Gilman a écrit :
>> 1. I would like to see David's hierarchical-theme navbar concept.  That probably stands between us and the next step. 
>
>David has sent a new mail with public references. Let us know what you
>think.
>

OK I looked at them.  In a quasi decision-tree development:

? At the top level of the decision tree, to include a navbar like this or not:

.. definite yes, but

? At the head of the page, in order from outer scope to inner scope?

.. can live with this.  I believe that there is a genuine cross-current between what is fluent and usable in vision and in speech in this regard.  For speech the optimum thing would be to have all these links present in the page but after the main content and in ascending order of breadth of scope, from inner to outer.  *But,* on one key condition:

? What broke?  The skip link.  The skip-nav link should be the first or second link encountered as one processes the page in tabbing order on opening the page from a URI-reference with no #fragment qualifier.  This navbar is the kind of stuff the skip-nav link was invented to skip.  David put the navbar ahead of the skip link by visually-optimized design thinking, not knowing about the function of this link.  The speech-using visitor should not be subjected to this litany on every page without access to a link to the main business of _this page_ appearing first.

So we have to come up with a way to get a skip-to-main business (which we verbalize as 'jump' still) link integrated with this hierarchical path summary and we have a candidate design change fit for prototyping.  The actual verbosity vs. clarity tradeoff for the link text, the link TITLEs, and the whole page if read in that alternate reading mode in the screen reader (which the reference user will only do under duress as a repair strategy) needs to be checked out by real qualifying individuals in the context of a working model with at least a whole month cast in this model.  Because it is the pattern of repetitions across pages as you navigate life-scale trajectories (and I don't mean three clicks) through the fabric that is needed to tell us if we are done.

Some further details:

.. The first verbal link to "W3C Home" is sufficiently redundant with the icon link that there should not be two links separately in the tabbing sequence of the page.  I think the logo wins.  Suggest we kill the first entry in the hierarchical descent.  The logo for the site in the upper left hand corner is such a strong cliche that it's not necessary to provide connectives to convey that that is the root of the "you are here" trace once it is clear that that is the dance that is being done.

.. @@ unresolved diction problem.  The sense of the link sequence in this bar is very dependent on the punctuation (flow arrows at the moment) which is not present when the link text is read as the links are tabbed through.  We have not yet provided title attributes for these links.  I would try for a happy medium at two word noun link texts, somewhat longer TITLEs, and an IMG for the 'containing' relationship with colloquial English TBD for the ALT that states the containment relationship.


An example full narrative rendition in descending order of scope would for reference be 

At W3C you find _mailing list archives_ including _public lists_ including _wai-xtech@w3.org_ including what is here, =March 2002 by thread=.

I am not saying we have to go all the way there, but just as an asymptote.

I would venture it is OK to have the user have to slow down and switch modes to get the explicit connective verbiage that talks about the relationships among the nominative link text phrases.  So long as the link texts themselves and TITLEs are clear enough.  The objective is to support a user-controlled range of verbosity vs. clarity balance points so first time visitors can get well oriented and inveterate visitors can go to the sub-blocks quickly. 

>>  Which I think of as having a whole list or at least a whole period
>> working with multiple index pages and multiple messages cross-linked.
>
>Hmm... I'm not sure how easy that would be. I'll need to check with
>Jose, but I agree that it would be a good way to proceed.
>
>> 3. The skip-nav link in the 'period view' index page mockup needs a title, perhaps
>> 
>> from 
>> 
>> <a href="#first" accesskey="j">58 messages</a>
>> 
>> to
>> 
>> <a href="#first" accesskey="j" title="jump to">58 messages</a>
>
>Implemented.
> 
>> 5. Use of the label "contemporary messages":
>> 
>> I agree it is awkward.  I was trying to save the implementer putting in [this month | this quarter] as appropriate.  
>> 
>> Furthermore @@ I don't know what to do about the following:
>> 
>> The actual function if your browser implements #fragment references right is to take you to "this message in date context"  and respectively " in thread context,"  "...in subject context,"  "...in author context."
>> 
>> It takes you not just to these indices, but to them shifted to put the current message at the origin of navigation.  The different staring point for navigation is important to understand because in speech it won't be obvious that it is the same page as arriving at the top of the page.  These links are not synonymous with the links emanating from other places, all of which start one at the top of the page.
>
>I see your point but don't have a good solution either.
> 
>> 6. The page title element for the message page should be front-loaded better [start with information specific to this message, not list-wide info]
>> 
>> from
>> 
>> <title>wai-xtech@w3.org from August 2001: Re: archive: top-level index</title>
>> 
>> to
>> 
>> <title>Re: archive: top-level index page with adjusted link text <?punctuation?/> wai-xtech@w3.org</title>
>> 
>> The <?punctuation?/> could be "-on-" or a dash or whatnot.
>> 
>> I would even consider dropping the list address from this page title.
>
>Implemented.
>
>Thanks for the feedback!
>
>Dom
>-- 
>Dominique Hazaël-Massieux - http://www.w3.org/People/Dom/
>W3C's Webmaster
>mailto:dom@w3.org
>

Received on Tuesday, 2 April 2002 08:55:56 UTC