RE: Fusebox (was: Web Accessibility for Cognitive Users)

Good point. Most of this stuff has been terribly abused - partly out of
laziness, but also partly because no one anticipated how big the Web would
be and the sorts of commercial uses it would be put to.

The two cookies were probably CFID and CFTOKEN, which are used for session
management. I have no idea why Cold Fusion needs two when everyone else
seems to get by with one. Me, I'm moving my sites away from Cold Fusion and
into Java servlets, mostly because CF is proprietary and requires the use of
a CF server.

My sites are designed so that the owner can maintain the site
(add/edit/delete text and images, rearrange page hierarchy, add/edit/delete
pages, etc.) without knowing any code. This has the advantage of allowing me
to lock in accessibility features and maintain standards-compliant code.
Yes, they can still make a mess of the site, but not as badly as otherwise,
and they can update the pages without wasting my time.

It's easier to do this if the site is stored in a database and the pages are
generated dynamically (which means I can regenerate the site metadata at
will). So if the owner moves a page in the hierarchy, all links and the site
map are automatically updated to reflect this.

I could write the software to write pages to the server (so they would be
static), but these sites are intended to be hosted on rented space. While I
can easily get use of a CF server and an Access database (or SQL Server),
getting hosting companies to let me write to the server is almost
impossible. So the database route it is, for now...

Charles F. Munat
Seattle, Washington

-----Original Message-----
From: w3c-wai-ig-request@w3.org [mailto:w3c-wai-ig-request@w3.org]On
Behalf Of David Woolley
Sent: Friday, May 18, 2001 3:33 PM
To: w3c-wai-ig@w3.org
Subject: Fusebox (was: Web Accessibility for Cognitive Users)


> Yes, that's true and one reason why I rejected the Fusebox method as it's
> currently applied. The trick is to avoid the ?. On my sites I substituted

It's a pity that few people are aware of PATH_INFO, which allows, servers
to construct virtual directories behind CGI applications.  Orignally
GETs to ? URLs were supposed to be cacheable, but abuse of this means
that HTTP 1.1 requires a positive indication of cachability and some
proxies don't even bother looking at the headers on request for them.

This particular page also tried to dump two useless cookies on me,
which is another thing that needs careful handling for cachability.

Received on Friday, 18 May 2001 22:11:16 UTC