Re: [apps-discuss] [OAUTH-WG] draft-jones-appsawg-webfinger-04

["Martin J. Dürst" <[e-mail]>]

On 2012/05/16 16:16, Graham Klyne wrote:

> There's an interesting maybe-comparable example to consider: Digital
> Object Identifiers. Several years ago, these were defined to provide a
> way for allocating unique identifiers to research publications, etc.
> (http://www.doi.org/,
> http://en.wikipedia.org/wiki/Digital_object_identifier).
>
> A doi: URI scheme was proposed but has not subsequently been approved
> for registration
> (http://human.freescience.org/htmx/digital_object_identifier.php);
> rather the doi was registered as a sub-namespace in another new URI
> scheme, info: (http://www.ietf.org/rfc/rfc4452.txt) which was defined as
> an umbrella for several of these "non-web" identifiers.
>
> But it's interesting that a DOI is most commonly cited in a web page and
> similar contexts, and often in print, as http://dx.doi.org/<handle>
> (http://www.doi.org/doi_handbook/2_Numbering.html#2.6.2) ... so the
> other URI scheme registrations don't seem to have that much traction.

Things like these happen mostly because new URI(/IRI) schemes are hard 
to deploy. That initially led to a lot of pressure against new schemes, 
but those who liked to invent new schemes didn't get convinced, so those 
whot didn't think new schemes were such a good idea just switched from a 
"we know, and we don't want you to find out the hard way" to a "let them 
find out themselves, we told you so" policy.

Of course, what I wrote didn't happen explicitly, but I hope it 
describes the general developments well. One reason for the change may 
be that when URIs were not that well known yet, those who invented them 
felt that they had to protect users from bad experiences. These days, 
URIs as such are extremely well established, and there isn't much of a 
danger anymore that a failure with a specific URI gets blamed on URIs in 
general.

Regards,   Martin.