cost of new/private URI schemes well known? [was: section 1, intro, for review]

On Mon, 2002-03-18 at 20:10, Paul Prescod wrote:
> Chris Lilley wrote:
> > 
> > ...
> > 
> > Okay, but with a flourish I can produce the URI
> > zip://atm.example.org/06902 or, worse, zip:06902 but just because
> > these now have URIs does not lessen in any way their proprietary
> > nature, and the second one is probably not dereferencable either.
> 
> Okay, but
> 
>  1. Why would anyone do such a thing? People know that non-HTTP URI
> schemes have a massive barrier to acceptance.

Ha!

er.. ahem...

Would that it were so.

As maintainer of an index of URI schemes,
  http://www.w3.org/Addressing/schemes
I can tell you, they don't. I can't even
keep up with all the crazy new URI
schemes I hear about.


I went looking into why mozilla parsed
	irc://irc.openprojects.net/rdfig
as a relative URI, and found that they
treat new URI schemes potatoe chips: "crunch
all you want. We'll make more."

KDE likewise.

Mozilla has since cleaned up their act a little;
I know they fixed
	http://bugzilla.mozilla.org/show_bug.cgi?id=2110

I haven't looked into KDE in a while.

Nor have I checked whether this persists:

  To the question "how many private URL schemes are there?," the
  answer was given that there were perhaps 20-40 in use at
  Microsoft, with 2-3 being added a day; WebTV has 24, with
  6/year added. Maybe others have similar number of schemes.

  -- minutes of the
  Uniform Resource Locator Registration Procedures (urlreg) WG meeting
  at the 39th IETF Meeting in Munich, Bavaria, Germany, Aug 1997
  http://www.ietf.org/proceedings/97aug/transit97aug-30.htm


You might think that these schemes are just private and
they don't matter. But then try using one in an HTML
document from the web, say:

	<a href="about:plugins">...</a>

and you'll quickly realize that it is an interoperability
issue: nobody else can use about: without bumping
into mozilla/netscape's use of it.

Then there's stuff like opaquelocktoken: and DAV:
that went all the way thru the IETF standards
process without much consideration of the cost
of real-estate in the URI scheme list.
(which, I suppose, is not really all that high,
in this case...)

Norm/Stuart/Ian, please let's add something to
section 2 on naming about the costs of
new URI schemes. Maybe it doesn't fit
in the one-page version, but before
we'er done elaborating, please let's.

By way of suggested text, see:

  The danger of too many accesses schemes
	in
  http://www.w3.org/DesignIssues/Model.html

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Tuesday, 19 March 2002 17:14:15 UTC