RE: Alertbox: Site map usability (fwd)

> Dan Brickley:
>
> excerpted from http://www.useit.com/alertbox/20020106.html 
> (fwd'd below)
> 
>  "We have often predicted that Internet Explorer version 8 will be the
>  first decent Web browser. One feature request for version 8 is better
>  navigation support by offering a visualization feature that 
> would pull
>  the site map out of the website and make it a standardized 
> browser element."

Hmm. Building site maps into browser functionality? Can't see it.


> So Mozilla kinda does this for the currently viewed page, 
> using HTML LINK
> elements in the HEAD of a document. But it only has the 
> current document's
> point of view, not access to the rest of the sitemap. Having 
> it (and IE,
> Opera...) consume RDF sitemap for navigation support would be 
> pretty cool.
> 
> Has anyone spent any time looking at this? (I'd expect folk 
> working on web
> browsers for mobile phones, PDAs etc would find this of 
> interest too...)

A site map is a table of contents. That would obviously be suited for
device browsing (don't have a style sheet for the user agent on the
server? Then serve up the toc). More useful on the client would be
precanned keywords served as an index page, that you could click on and
have the urls listed. Or indeed downloaded to the client and added to
its own index.

Nonetheless, in many cases what you really want is a good search engine,
and a decent UI. 

One thing that might be useful on the client to people is standardised
help/contact/about/policy pages so they appear in the browser in the
same place (as in windows apps) rather than on the site itself. 

Although...the effort designers go to avoid using browser functionality
for anything is substantial. For app designers, outsourcing key
functionality to the browser is a risky proposition, if not plain
barking; and honestly who can blame them. Hence the baffling popularity
of presentation formats designers and branders can control like flash
and pdf. The lesson learned is that any website that depends on the
browser for anything other than rendering has failed and will be in a
world of usability hurt. You'd have to show it makes the site better and
doesn't hurt the brand and isn't ever going to break across versions and
browsers, as well as undo half a decade's worth of pain. 

But at the end of the day, for site maps, why not just serve a page up?

I've seen two of the better usability lists (sigia and chi-web) slowly
warm to topic maps in recent months (usability people think about site
taxonomy a lot); they're curious. If you gave the usability world a
decent tool (and that's likely to be a plug-in to something like Visio,
or integrated into a workflow system), they'd use topic maps to organise
sites, which could be passed off to the programmers to blow out maps and
site templates, or indeed filter search results. Rdf and ontologies seem
a bit scary for them, compared to topics and taxonomies (it's all in the
perception). There's a genuine lack of tool support for this.

I've seen drill down thematic site maps (think clickable Mondrian
paintings) based on Kohonen maps, can't recall the link, but I know I've
got it bookmarked :( Maybe browser v10 will allow me to do something
useful with my bookmarks, other than have the thoughts "black hole",
"use Google", pop into my mind in quick succession. I use Google more
and more to find something I already know about, and I find the history
pane in IE its most useful feature, which would be even more useful if
there was some meta data stored with the item (the index terms would do
nicely).

For info-centric sites like ora.com, as interesting than site maps is
the facility to feed back server stats in real time to the user and use
them as citation or user flow data, to mimic the way citeseer, amazon
and google work. That could be built into the browser or just appended
to the page to let people know what other people who looked at this page
are looking at, without upsetting the privacy vanguard. It would be nice
to see how others are moving through the site map along with the site
map itself.

regards,
Bill

Received on Tuesday, 8 January 2002 09:40:07 UTC