W3C home > Mailing lists > Public > www-tag@w3.org > March 2008

Re: I-D ACTION:draft-nottingham-http-link-header-01.txt

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Thu, 20 Mar 2008 14:01:44 -0400
To: Phil Archer <parcher@icra.org>
Message-Id: <8C5DCF9C-F6D4-4D76-8E79-5675793EC434@gmail.com>
Cc: Mark Nottingham <mnot@mnot.net>, www-tag@w3.org


On Mar 20, 2008, at 1:18 PM, Phil Archer wrote:

> There's a lot of crossover going on here... the discussion about  
> relationship types has played out extensively on the IETF/HTTP WG  
> list. See, for example, Julian Reschke's comment [1].
>
> Mark's original proposal was to use Link-Profile to make the  
> relationship type extensible which seems sensible at first pass,  
> but...
>
> Link: <file.ext1>; rel="rel-1"
> Link: <file.ext2>; rel="rel-2"
> Link-Profile: <profile.html>
>
> Does the profile define the first or second relationship type, or  
> both?

I'm not sure.

> Link-Prefix would have the same problem but there may be a way  
> through.

It would not - you could have multiple LinkPrefixes. Just as with XML  
or RDF, each rel would have to be resolved relative to one of them.
Each LinkPrefix species two things: the prefix (optionally empty, to  
support the desire to use the current strings) and the URI fragment  
that should be put in its place.

> Could IANA be the default Link-Prefix? So if that header is absent,  
> OK, you know that any relative URI is relative to the IANA namespace.

One could do this, but my feeling is that it might encourage less  
careful behavior. People will need to change their code to generate  
link headers - given that it isn't too much of a burden to have the  
IANA link prefix added too, if that's what they want to use. By doing  
so we clearly link this part of Web architecture with the emerging  
Semantic Web additions to the architecture.

> If supplied, then the prefix URI MUST apply to ALL relationship  
> types, even if using one from the IANA namespace - and whatever the  
> namespace is should use something like owl:sameAs to map from its  
> own terms to the IANA ones - seems a bit wasteful.

I don't think this is a good aspect of the design. It's as if you  
were allowed only one prefix when you write RDF.

> Alternatively, we could, perhaps have an id for each Link,  
> something like
>
> Link: <file.ext1>; rel="rel-1"; id="a";
> Link: <file.ext2>; rel="rel-2"; id="b";
> Link: <my.css>; rel="stylehseet"; id="c";
> Link-Profile: <profile.namespace> for="a b";
>
> (i.e. the Profile header would take a space separated list of  
> identifiers for the link headers to which it applied. Again, the un- 
> qualified relationship type (header c) is from the IANA namespace.

Isn't a "profile" just like any other document that defines a number  
of relationships? The current name for such things are "ontologies",  
and they generally live in RDF or OWL files. I guess I don't see why  
we need to introduce a new name here when we have a mechanism that  
works and is already part of the architecture. i.e. If a rel  is a  
URI that names a property, then we figure out what it means in the  
usual way - by doing a GET of the URI and looking at what is returned.

In the case you give, I would write this as:

Link-Prefix: "a", "http://profile.namespace.com/relation#"
Link-Prefix: "", "http://www.iana.org/assignments/link-relations#"
Link: <file.ext1>; rel="a:rel-1"
Link: <file.ext2>; rel="a:rel-2"
Link: <my.css>; rel="stylesheet"

Assuming that "stylesheet" means the property <http://www.iana.org/ 
assignments/link-relations#stylesheet>
and that <profile.namespace> is actually some URI  such as "http:// 
profile.namespace.com/relation#"

To my mind this maximally borrows from current SW practice

BTW, I don't have any preference for the syntax of the Link-Prefix  
header - could equally be any of
Link-Prefix: "a";"http://profile.namespace.com/relation#"
Link-Prefix: a=http://profile.namespace.com/relation#
Link-Prefix: a;"http://profile.namespace.com/relation#"

Best,

-Alan

>
> Phil.
>
>
>
>
> [1] http://lists.w3.org/Archives/Public/ietf-http-wg/2008JanMar/ 
> 0468.html
>
> Alan Ruttenberg wrote:
>> Hi Mark,
>> Having the link types be URIs will be a great help in solving a  
>> number of issues we've been dealing with around associating  
>> metadata with resources.
>> I have one suggestion for the document - rather than having the  
>> link-extensions, with the default (i.e. not necessarily stated in  
>> the headers) base of
>> http://www.iana.org/assignments/link-relations.html I would  
>> instead either leave that mechanism out, or add something like a  
>> Link-Prefix: header that allows one to set a prefix in the way one  
>> does for namespaces or RDF serializations.
>> Ideally the IANA registry would be served as RDF (perhaps by  
>> conneg). In this way, resolution and discovery of relations could  
>> be uniform. An agent wishing to discover what
>> Link-Prefix: "", "http://www.iana.org/assignments/link-relations#"
>> Link: <http://www.cern.ch/TheBook/chapter2>; rel="Previous"
>> Would do exactly the same thing as working with:
>> Link-Prefix: "dc:", "http://purl.org/dc/terms/"
>> Link: <http://mumble.net/~alanr/ThePersonAlanRuttenberg>;  
>> rel="dc:creator"
>> Regards,
>> Alan
>> On Mar 19, 2008, at 3:11 AM, Mark Nottingham wrote:
>>>
>>> See referenced I-D for a rough idea of what I've been kicking  
>>> around WRT the Link header with a few people.
>>>
>>> Note that while it resolves the relation mess, it still has to  
>>> get some buy-in by both the HTML and Atom communities, as it asks  
>>> some non-trivial things of them.
>>>
>>> Also, this draft is still rough, with some outstanding issues  
>>> already identified. See discussion on the ietf-http-wg list.
>>>
>>> Cheers,
>>>
>>>
>>> [ note: I've deleted the MIME attachment, because the last time I  
>>> forwarded this message, it appeared to crash Mail.app instances  
>>> that received it. Oops. ]
>>>
>>>
>>>
>>> Begin forwarded message:
>>>
>>>> From: Internet-Drafts@ietf.org
>>>> Date: 18 March 2008 5:30:01 AM
>>>> To: i-d-announce@ietf.org
>>>> Subject: I-D ACTION:draft-nottingham-http-link-header-01.txt
>>>> Reply-To: internet-drafts@ietf.org
>>>>
>>>> A New Internet-Draft is available from the on-line Internet-Drafts
>>>> directories.
>>>>
>>>>
>>>>     Title        : HTTP Header Linking
>>>>     Author(s)    : M. Nottingham
>>>>     Filename    : draft-nottingham-http-link-header-01.txt
>>>>     Pages        : 13
>>>>     Date        : 2008-3-17
>>>>     This document clarifies the status of the Link HTTP header and
>>>>  attempts to consolidate link relations in a single registry.
>>>>
>>>> A URL for this Internet-Draft is:
>>>> http://www.ietf.org/internet-drafts/draft-nottingham-http-link- 
>>>> header-01.txt
>>>>
>>>> Internet-Drafts are also available by anonymous FTP at:
>>>> ftp://ftp.ietf.org/internet-drafts/
>>>>
>>>> Below is the data which will enable a MIME compliant mail reader
>>>> implementation to automatically retrieve the ASCII version of the
>>>> Internet-Draft.
>>>> ______________________
>>>> I-D-Announce mailing list
>>>> I-D-Announce@ietf.org
>>>> https://www.ietf.org/mailman/listinfo/i-d-announce
>>>> Internet-Draft directories: http://www.ietf.org/shadow.html
>>>> or ftp://ftp.ietf.org/ietf/1shadow-sites.txt
>>>
>>>
>>> -- 
>>> Mark Nottingham     http://www.mnot.net/
>>>
>>>
>
> -- 
> Phil Archer
> Chief Technical Officer,
> Family Online Safety Institute
> w. http://www.fosi.org/people/philarcher/
>
>
>
Received on Thursday, 20 March 2008 18:02:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:53 GMT