Re: Registration of acct: as a URI scheme has been requested

On Thu, Jun 21, 2012 at 3:39 AM, Larry Masinter <masinter@adobe.com> wrote:
> looks like the registration form is missing pointers to the actual semantics
> of the scheme as well as to the only actual use case.

In defense of the webfinger draft, the form of the registration is no
different from the http: scheme registration in this respect.  In both
cases, the syntax of the URIs is specified, but the registration
section of the RFC says almost nothing about semantics. Yes, there is
strong guilt by association because the scheme description happens to
occur inside the documentation of a protocol, but the HTTP protocol
(like webfinger) can be used with any kind of URI, and there is no
explicit statement in the RFC linking the scheme to the protocol.

There is the one semantic-sounding statement 'The "http" scheme is
used to locate network resources via the HTTP protocol' but this is
completely uninformative since other schemes, such as ftp:, can *also*
be used to locate network resources via the HTTP protocol.

I don't think HTTPbis is substantially different from 2616 in this
respect, either.

So, if the criticism applies to acct:, it also applies to http:.

Jonathan

> -----Original message-----
>
> From: Kingsley Idehen <kidehen@openlinksw.com>
> To: "www-tag@w3.org" <www-tag@w3.org>
> Sent: Wed, Jun 20, 2012 21:57:13 GMT+00:00
> Subject: Re: Registration of acct: as a URI scheme has been requested
>
> On 6/20/12 5:30 PM, Noah Mendelsohn wrote:
>>
>>
>> On 6/20/2012 1:42 PM, Kingsley Idehen wrote:
>>
>>> If the architecture of the world wide web can't accommodate new URI
>>> schemes
>>> then its broken. The great news is that it isn't broken.
>>
>> The way this is phrased, you leave out a crucial subtlety. The
>> Architecure of the World Wide Web is quite clear on the fact that
>> definition of new schemes is allowed, but costly [1]:
>
> True!
>
>>
>> ===========
>> While Web architecture allows the definition of new schemes,
>> introducing a new scheme is costly. Many aspects of URI processing are
>> scheme-dependent, and a large amount of deployed software already
>> processes URIs of well-known schemes. Introducing a new URI scheme
>> requires the development and deployment not only of client software to
>> handle the scheme, but also of ancillary agents such as gateways,
>> proxies, and caches. See [RFC2718] for other considerations and costs
>> related to URI scheme design.
>>
>> Because of these costs, if a URI scheme exists that meets the needs of
>> an application, designers should use it rather than invent one.
>>
>> Good practice: Reuse URI schemes
>
> Yes.
>
>>
>> A specification SHOULD reuse an existing URI scheme (rather than
>> create a new one) when it provides the desired properties of
>> identifiers and their relation to resources.
>> ===========
>>
>> Many TAG members seem to be concerned that the advice above is too
>> often being ignored. In addition to the points made above, each token
>> in the scheme space is valuable, precisely because it is, by
>> tradition, one in which the names are short, central registration is
>> required, and no two facilities can use the same scheme name for
>> different purposes. Even if no acct URI's "leak" out from the intended
>> private use within Web finger, that relatively suggestive short name
>> is now not available for any other purpose as a scheme.
>>
>> For all these reasons, the bar should be set very high on the
>> registration of new schemes. That does not mean, as you suggest, that
>> there is a question of not "accommodating" new schemes. In the rare
>> cases where a new scheme is appropriate, the architecture can
>> accommodate it.
>
> As you know, there's no bar higher that knocking HttpRange-14
> distractions at every turn. WebID solve a major problem. Its reliance of
> Linked Data ultimately brings the denotation (name) and web resource
> identification duality issues into scope. Thus, we have an example of a
> situation where a new URI scheme is warranted albeit solely as an
> additional option for entity (web or real-world) names that use
> indirection to resolve to web based descriptor documents/resources.
>
> Kingsley
>>
>> Noah
>>
>>
>>
>> [1] http://www.w3.org/TR/webarch/#URI-scheme
>>
>>
>>
>
>
> --
>
> Regards,
>
> Kingsley Idehen
> Founder & CEO
> OpenLink Software
> Company Web: http://www.openlinksw.com
> Personal Weblog: http://www.openlinksw.com/blog/~kidehen
> Twitter/Identi.ca handle: @kidehen
> Google+ Profile: https://plus.google.com/112399767740508618350/about
> LinkedIn Profile: http://www.linkedin.com/in/kidehen
>
>
>
>
>

Received on Thursday, 21 June 2012 15:02:07 UTC