W3C home > Mailing lists > Public > uri@w3.org > March 2006

RE: [dix] on the dix: URI scheme for DIX/SXIP

From: <noah_mendelsohn@us.ibm.com>
Date: Mon, 20 Mar 2006 21:49:34 -0500
To: "McDonald, Ira" <imcdonald@sharplabs.com>
Cc: "'Ben Laurie'" <ben@algroup.co.uk>, Digital Identity Exchange <dix@ietf.org>, uri@w3.org
Message-ID: <OFB1AD2C6D.14349C30-ON85257138.000E6159-85257138.000F86A4@lotus.com>

Ira McDonald writes:

> Although I have mixed feelings about the logic, the brand new 
> RFC 4395 "Guidelines and Registration Procedures for 
> New URI Schemes" argues that the bar should be very high for a 
> new 'Permanent' URI scheme, because so many browsers
> and other bits of client software will have to updated for the 
> URI scheme to become widely deployed and used.

FWIW, I think that's sometimes among the reasons for viewing scheme names 
as scarce, but not the only reason or even the most widely applicable.  As 
noted later in this thread, there are many potential schemes for which the 
required scheme-specific software would likely by quite isolated.

As far as I'm concerned, the scheme names that are ultimately scarce are 
the ones that are simple to type and easy to remember, and especially 
those likely to be suggestive of particular real world concepts.  If 
someone wanted to argue for schemes like:

        a1324987123904873209487:
        a8976872648574368573469848765:
        a3487687683687463874683:

I would say they have a variety of important downsides, but scarcity isn't 
among them.  (I'm showing scheme names with the trailing ":" even though 
that's not technically part of the name).  On the other hand, if someone 
wants to have a scheme:

        phone:

then that's it for all time.  We can't come back in 20 years and decide 
we'd like to use the word phone for some more robust abstraction of 
phones, or for something entirely different. 

The situation with scheme names is aggravated by the fact that, like top 
level domain names, they are not fundamentally hierarchical per RFC 3986. 
One could build a hierarchical convention within scheme names (Roy 
occasionally suggests xxx.yyy), but then the xxx becomes the limited 
resource. 

Of all the potential uses of the three letters "dix" for a scheme name, is 
the proposed use likely to be among the most important over a period of, 
say, 50 years?  If not, I think assignment of the scheme is suspect.  It's 
in that sense that I see scheme names as scarce.

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------
Received on Tuesday, 21 March 2006 02:51:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:36 GMT