RE: Excess URI schemes considered harmful

I've read through the discussion of this and other points on the
URI mailing list, but the discussion seems to have wandered.

I find most of the arguments for "Excess URI schemes harmful"
to be circular. They're harmful because they're harmful.
I don't understand the assertion "The Web depends on a very
high shared knowledge of the properties of URI schemes".

I don't see anything more harmful in using an obscure and
unpopular URI scheme than there is in using an obscure and
unpopular content-type or XML namespace. If you communicate
and start talking about flxxernsties, and no one knows what
a flxxernsty is, then you won't be understood. Similarly,
applications that use novel URI schemes will find that those
schemes won't travel well.

The harm in *registering* new URI schemes is that the registration
process is cumbersome, and that short scheme names might,
at some point in the future, see the same kind of vanity
use as top level domains. (NO NO, *I* want to define "coffee:").

The problems you allude to are problems with indiscriminate *use*
of novel URI schemes in any context where it might escape,
but the problem is much less than that experienced in using
a context-sensitive identifier. At least with URIs, the receiver
knows that the receiver doesn't understand what the sender
has sent.

(FWIW, webdav defined "DAV:" and not "webdav:").


> -----Original Message-----
> From: Tim Berners-Lee []
> Sent: Monday, September 24, 2001 11:15 AM
> To: Larry Masinter;; Dan Zigmond; Rich
> Petke
> Cc:; Dan Connolly
> Subject: Excess URI schemes considered harmful
> I note, for includion in a future version of
> that there is a very high cost
> associated with the deployment of new URI schemes. There a number of
> flexibility points in web architecture, and
> those such as  Content-Type, and XML Namespace creation have a much lower
> cost.  The Web depends on a very high shared knowledge of the properties
> URI schemes. New ones should only be introduced is absolutely necessary.
> Content-Types should be defined by URIs, as are XML Namespaces. These then
> leverage the existing URI schemes to anchor thier meanings in the web.
> allows anyone to make a local private Content-Type or namespace for their
> own use. This does NOT apply to URI schemes.  The process has to be rooted
> somewhere, and that root is the URI spec and the *small* set of  URI
> schemes.
> I would recommend that this be emphasized.  Specific examples of the
> creation of  arbitrary trivial schemes such as the "webdav:" have violated
> this rule in the past.
> Tim Berners-Lee

Received on Friday, 28 September 2001 03:25:31 UTC