Re: updating RFC 2718 (Guidelines for new URL schemes)

Hello Harald,

At 08:36 04/08/30 +0200, Harald Tveit Alvestrand wrote:

>Thanks for looking at this again, Larry!
>
>Yes, chunks of it comes from my "things to think about" memo.... that was 
>my only real claim to authorship on this one!
>
>I feel that there is still value in keeping an Informational resource 
>about URI registration - there is much that is common (or uncommon?) 
>knowledge about URIs and registration, that shouldn't be part of the 
>ruleset (because there are sometimes good reasons to override them), but 
>should definitely be thought about before designing/registering an URI scheme.

I think the direction at the BOF in San Diego was to have only
one document, basically a combination of the current two documents,
because this makes it much easier for scheme
designers (one-stop shopping).

Also, if we make the registration easier, as also discussed at the BOF,
this may mean that there is more material that is only advisory/informational,
as opposed to fixed rules.

But I think it should be possible to put this all in one document,
there are things such as 'MUST' and 'SHOULD', as well as the possibility
to label sections as informative,...



>It also allows keeping more of the dry humor that I sometimes appreciate 
>in the RFC series, such as the "UR* = flamewar" comment... if you have 
>compelling arguments for making it a BCP resource, I am listening....

'One stop shopping' is a very strong argument.


>Note - the original document was deliberately limited to URL schemes - 
>this is not an accident; stuff like the HTTP proxy and "can all objects be 
>referenced" mostly makes sense when it's an URL; the comment from 
>draft-ietf-url-process-00 that got deleted from the "uninteresting 
>questions" question:
>
>>2.5.1 Is it an URL, an URN or something else?
>>
>>   This classification has proved interesting in theory, but not
>>   terribly useful when evaluating schemes.
>
>is a prime candidate for reinstatement if we change it to be a document 
>about URI registration.....
>
>(<http://lists.w3.org/Archives/Public/uri/1997Jan/0000.html> is one URL 
>for this I-D...)

I think we need some considerations both for schemes that for similar
to http:/file:/..., for schemes that work similar to things like telnet:,
and for schemes that do things similar to tag:,..., as well as probably
others. Some rules and advice will be common, but in other cases, we
will have to just say: make clear which kind of use this is for.


Regards,    Martin.

Received on Tuesday, 31 August 2004 11:53:38 UTC