|
Posted by Alan J. Flavell on 02/28/06 13:56
On Tue, 28 Feb 2006, Jim Higson wrote:
> The URL doesn't have to be the literal name of the file. Why would
> the user care if you store static html files or use shtml? They
> don't, and shouldn't be exposed to this.
Correct and agreed.
> Instead of http://example.com/example.html just use
> http://example.com/example as the address
If you're setting up a new site, then this is certainly an option.
If you're working on existing URLs, then it may be better not to
change them.
> - that way you can change how the site is hosted any number of times
> and the URL stays the same.
Several of my existing *.html files are now SSI or occasionally even
CGI calls (in fact, most of my personal *.html files are now set by
"XBitHack full", to minimise the cacheability penalty for an SSI call,
but that's a local detail).
> Tim Berners-Lee has written on this subject. See:
> http://www.w3.org/Provider/Style/URI
Indeed, "cool URLs don't change". So keep your existing URLs, as
far as possible, by *reconfiguring the server* - to treat them as
SSI, PHP, or whatever you want to introduce, without losing the
existing .html extension. There's nothing sacrosanct about a
particular filename extension.
But whenever introducing server-side processing techniques, *do* give
a thought to cacheability. More and more web sites seem to be getting
intolerably slow to respond, *even for revisiting a page that I was
looking at only moments ago*, and when I look more closely, they
almost always wrecked their cacheability by clumsy use of server-side
processing.
[Back to original message]
|