|
|
Posted by Alan J. Flavell on 09/27/06 07:01
On Tue, 26 Sep 2006, Jonathan N. Little wrote:
> Alan J. Flavell wrote:
>
> > [...] The distinction between differently-cased URLs disappears
> > when the typical win32 server maps URLs into the *file system*,
> > it's true, but that's not because the URLs have magically become
> > case-insensitive, but because the *file system* is, for these
> > purposes, case-insensitive. The web doesn't know that, nor care
> > about such internal server details - that's an important
> > principle: URLs define a hierarchy of their own, which was
> > designed to be server-agnostic.
>
> It may be semantics here, and that the HTTP transaction will be
> case-sensitive and it is the mapping file system case-insensitive
> occurs.
As several other contributors have also pointed out: URLs are, by
definition, case-sensitive. You can't argue that away (if that's what
you were trying to do). It's an important principle, as I said, that
the URL scheme was defined to be server-agnostic: the web does not
know, and does not need to know, these details about a server, any
more than it would need to have the "/" replaced by ":" or "\"
depending on details of the server's file system.
> The file system does come to play here though. Yes your can
> configure the web server to compensate, but I have really come
> across any.
I think you missed a negative there, but this is an important
principle, and applies irrespective of whether you or I have seen
examples of it actually being done.
This is rather a fundamental principle of the web architecture - it's
not just an argument about "counting angels on pinheads", which is why
I'm still dragging this argument out as long as you seem to be
obscuring the principle. Sorry.
Navigation:
[Reply to this message]
|