[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: No new mime types




On 20 Jul 2006, at 17:55, Mark Nottingham wrote:
Big +1 and well-said.

It's easy to forget that the user-agent (or other consuming software) isn't the only thing working with this stuff.

I don't quite understand what you are saying. In the following link

<link rel="categories" href="/cats"/>

which we can understand as pointing to a resource that gives a list of categories available to the poster, there is no need to specify the mime type of the representations served by </cats>. We know that the resource should give us a list of categories. Now if we queried it using application/rdf+xml we could get a list of cats using the atom-owl ontology [1] for example. We could also get a representation back in html perhaps for people who would like to browse the categories, the way I explain in "GET my meaning?" [2].

In RDF that would be expressed by saying the following about the relation iana:categories

@prefix iana: <http://www.iana.org/assignments/relation/> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .

iana:categories a owl:DatatypePropery;
                rdfs:domain :FeedOrEntry;
                rdfs:range :CategoryList .


So there really is no need for an extra mime type to describe what is at the end of the relation. It does help to know that some types of representations are available. But the bulk of the work should really be done by the rel value. As far as possible one should be agnostic about what representations are on the other end. After all, why tie oneself to a format?

Henry

[1] http://bblfish.net/work/atom-owl/2006-06-06/
[2] http://blogs.sun.com/roller/page/bblfish?entry=get_my_meaning


On 2006/07/09, at 1:45 PM, Joe Gregorio wrote:


On 7/9/06, Henry Story <henry.story@xxxxxxxxxxx> wrote:

Here is a suggestion. APP should define no new mime types. XML can
just be served as application/xml  if needed. If there is a need to
identify the type of object at the end of a relation new relation
types should be defined.


Mime-types are used for dispatching on the web, without
looking into the entity body to determine how to
dispatch. It's a fundamental precept of the web.

"I don't like it" is not a reason to change the spec.

You have in now way argued that a new mime-type will hurt
interop or violate the architecture of the web.

  -joe

--
Joe Gregorio        http://bitworking.org



--
Mark Nottingham     http://www.mnot.net/