W3C home > Mailing lists > Public > www-tag@w3.org > February 2004

Re: Dual opaque and hierarchical URIs

From: Tim Berners-Lee <timbl@w3.org>
Date: Sun, 1 Feb 2004 11:46:11 -0500
Message-Id: <24342D24-54D6-11D8-8473-000A9580D8C0@w3.org>
Cc: www-tag@w3.org, kpako@yahoo.com
To: Elliotte Rusty Harold <elharo@metalab.unc.edu>

I may have said this before.
My feeling is that feed:  is basically harmful in that it confuses the  
identity of the object with the way the user should treat it.  "feed:"  
and "webcal:"  are harmful in the same way, and "http:" should be used.

If one wanted to make something in XLINK to add subscription as a  
possible option, that may be reaonsbale, but by far the best solution  
if for the client to recognize that this is something may way to  
subscribe to
and then ask the user.  Often the user needs  to give other things as  
well, such as frequency.

Tim BL


On Jan 29, 2004, at 9:32, Elliotte Rusty Harold wrote:

>
> I noticed something very unusual about the proposed feed URI scheme  
> for RSS. This is separate from previous concerns that have been raised  
> about the feed URI being a synonym for http.  There may be a deeper  
> architectural principle at stake here too, which is not currently  
> addressed in the URI specs. Is it permissible for one scheme to have  
> both opaque and hierarchical forms?
>
> The draft spec  
> <http://www.25hoursaday.com/draft-obasanjo-feed-URI-scheme-02.html>  
> states :
>
> The following are examples of the "feed" URI scheme
>
> feed:http://example.com/rss.xml  	- Identifies the RSS feed at  
> 'http://example.com/rss.xml' feed:https://example.com/rss.xml   -  
> Identifies the RSS feed at 'https://example.com/rss.xml'  
> feed://example.com/rss.xml  	- Identifies the RSS feed at  
> 'http://example.com/rss.xml'
>
> The first two examples are opaque URIs in an RFC 2396 sense. The third  
> is a hierarchical URI. This would certainly cause problems for some  
> generic URI processing code I have recently written in XOM which I now  
> realize has an implicit assumption that each scheme is either opaque  
> or heirarchical but not both. I'm not sure this is wrong according to  
> RFC 2396. That spec contains no  clear language which would rule this  
> out, and at least one sentence implies that this is POK ("It is not  
> necessary for all URI within a given scheme to be restricted to the  
> <hier_part> syntax, since the hierarchical properties of that syntax  
> are only necessary when relative URI are used within a particular  
> document.") This sentence has been deleted from the draft replacement  
> for RFC 2396.
>
> Regardless, this seems quite confusing and potentially problem causing  
> for existing software. Is it really necessary or advisable to define a  
> URI scheme that can be both opaque and hierarchical? Continuing with  
> this line of thought, the primary purpose of hierarchical URIs is to  
> allow relative references, but does this really apply here? Could  
> there be a relative reference to feed://example.com/rss.xml ? Would  
> this have the expected resolution? Probably. But I expect it would be  
> confusing to users that they could use relative references for  
> feed://example.com/rss.xml but not for the opaque form  
> feed:http://example.com/rss.xml
>
>
>
> --  
>
>   Elliotte Rusty Harold
>   elharo@metalab.unc.edu
>   Effective XML (Addison-Wesley, 2003)
>   http://www.cafeconleche.org/books/effectivexml               
> http://www.amazon.com/exec/obidos/ISBN%3D0321150406/ref%3Dnosim/ 
> cafeaulaitA
Received on Monday, 2 February 2004 09:39:25 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:03 UTC