Re: User friendly 404s reconsidered

> On a tangent, there is certainly something to be said about URIs that
* are human-readable to a certain extent, as opposed to purely numerical
* ones...

And even more so when all the nodes in the name tree return something
useful.  One of my most common sources of 403s (or 404s if the site
doesn't want to make present but unauthorised resources detectable) is
trying to navigate up the hierarchy by stripping final components.

To me, CMS URLs really defeat the idea of URLs.  Even more so because
they typically make the fact that the real web server isn't the web
server by using ? style URLs, which, because of past abuse, are 
uncachable by default.  (Most sensible CGI implementations allow
the balance of the URL path to be used as a CGI parameter, and one
would have hoped, if not expected, CMS implemetors to be aware of 
that.)

This does introduce a reason why 404s can be more complex than they
seem, as providing the resource has function like characteristics,
there needn't be any pre-existing page file to back a URL.

Received on Tuesday, 25 July 2006 20:26:42 UTC