W3C home > Mailing lists > Public > uri@w3.org > April 2010

Re: backronym proposal: Universal Resource Linker

From: Roy T. Fielding <fielding@day.com>
Date: Sun, 18 Apr 2010 23:18:06 -0700
Cc: URI <uri@w3.org>
Message-Id: <FC14CDEB-A9AF-4FB6-8D5C-B457251DD1A8@day.com>
To: Dan Brickley <danbri@danbri.org>
On Apr 18, 2010, at 10:53 PM, Dan Brickley wrote:

> On Mon, Apr 19, 2010 at 2:21 AM, Roy T. Fielding <fielding@day.com> wrote:
>> On Apr 18, 2010, at 4:44 PM, John A. Kunze wrote:
>>> The "it" I had in mind, for those non-mind-readers on the list, was
>>> people's tendency to become irrational when seeing the word Locator.
>>> And the proposal would definitely help reduce that risk.
>> 
>> No, it would just increase the tendency to become irrational for
>> those folks who don't happen to be using a URL for linking.
>> 
>> I think you all need to find a better hobby.  URI is a full
>> Internet Standard and the generic syntax spec isn't likely to
>> change in the next fifty or so years no matter what color you
>> want to paint this shed.  You would need a couple thousand
>> vendors to agree to such a change for it to even come close
>> to overcoming the consensus we spent fifteen years attaining
>> on the current name.
> 
> And how's that working out for you?

Absolutely fine.

> We've all been dutifully being saying "URI" for years, around here.
> Sometimes even URI Reference, and lately IRI when we're feeling
> especially proper. Meanwhile the original name is still very heavily
> used, to understate the situation. In anything but the most
> standards-addled settings, the single basic concept of URI *still*
> needs a special introduction, and solely because it is named URI and
> not URL. 15 years into the great "URI" experiment, 20 years into the
> Web, we can't mention the name of it's fundamental technology without
> having to explain that "it's a bit like an URL". This creates a
> needless barrier between those who are building Web technology, and
> those who use it.

No, it doesn't.  People who need to know the difference already know it.
It is simply IRRELEVANT how many people still use the term URL since
that term encompasses the exact same set as URI.  It always has.
If I am talking to a non-technical person, I will talk about URI
(the standard) and "Web addresses" (browsing) and "references"
(the stuff you type into the address bar or href).

The only problem we've had recently is folks who claim an arbitrary
reference string is a URL, which doesn't make any sense whatsoever
no matter what the acronym spells.  "" is not a URL/URI/IRI.  It might
be a URI reference or an IRI reference.

> What I am suggesting does not ask for any redefinition of URI, it's
> specification, name or syntax. It is just a restoration to greater
> respectability of the colloquial name. We lack, and I believe we need,
> a single simple high level term for these things; something that
> understood by doctors, journalists, artists, engineers, parents,
> grandparents, children and puppies. Those who care will continue to
> check the detail and use the real technical terms, and they will often
> need language that carefully distinguishes IRIs from URIs. But for
> most mainstream use, we lack an umbrella term and have (due to the URN
> vs URI perma-debate) lost the ability to talk clearly about
> UR-Locators. I am suggesting we say "if it's a URI, a URI Reference,
> or an IRI, then ... sure, it's what we now casually call an URL". The
> suggestion to change the acronym expansion could live or die
> separately, but it does emphasise the core feature of the Web.

There is no need to talk clearly about UR-Locators. All URNs are locators.
Being able to use an identifier for locating something has nothing to do
with its syntax.  It has to do with the availability of a
resolution/retrieval mechanism.  Changing the name is not going to
change the desire of some people to keep rehashing that debate --
it will just introduce yet more ambiguity into the term.

> What do you prefer as the most over-arching and inclusive term to use
> in everyday discourse?

URI (for output) and reference (for input).

....Roy
Received on Monday, 19 April 2010 06:18:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:14 UTC