W3C home > Mailing lists > Public > semantic-web@w3.org > June 2006

Re: iana mime types as rdfs:Datatype ?

From: Henry Story <henry.story@bblfish.net>
Date: Tue, 27 Jun 2006 00:28:03 +0200
Message-Id: <78F38DBF-4F82-42AF-AF30-6FB0C0C2210A@bblfish.net>
Cc: Semantic Web <semantic-web@w3.org>, atom-owl@googlegroups.com
To: David Powell <djpowell@djpowell.net>

Yep. I agree with your point below. The fact that one can have  
";format=flowed" makes the whole thing
a lot more complicated. It may be that this still means that one can  
use the url as a datatype. I am not sure what all the parameters  
after that mime type can mean. If they are mostly encoding  
specifications then this would be ok since when one uses it as a  
datatype then the encoding issues would already have been dealt  
with... "<b>hello</b>"^^ianatxt:html no longer requires the encoding  
information. The issue is interesting but a little too complicated  
for us to look at now. We'll leave that as a puzzle for someone else  
to look at.

On the latest atomOwl ontology [1] we currently do 2 things:

1. All text constructs get translated to one of the following literals
    -simple text construct to literals such as [] :title "my first  
entry".
    -html constructs to literals such as [] :title "My first  
<i>important</i> entry."^^:html
    -xhtml constructs to literals such as [] :summary "<div  
xmlns...xhtml>hello</div>"^^:xhtml

2. <content> and <link...> objects are turned into relations  
to :Content objects, that
keep all of the information from atom xml.

[ a :Entry;
  iana:alterate [ a :Content;
                  :type "text/html";
                  :length 976;
                  :src <http://eg.com/entry1.html>
                ];
  :content [ a :Content;
             :type "text/xhtml";
             :body """<div xmlns="http://www.w3.org/1999/xhtml"  
xml:lang="en" xml:base="http://diveintomark.org/">
         <p><i>[Update: The Atom draft is finished.]</i></p>
       </div>""" ]
].

Essentially a :Content is what one can GET at a url. It has therefore  
very similar properties
that an HTTP header + body has.

Henry

On 26 Jun 2006, at 22:27, David Powell wrote:

>
> Friday, June 23, 2006, 11:46:33 AM, Henry Story wrote:
>
>> Every iana mime type has an associated url. So for example text/html
>> is  http://www.iana.org/assignments/media-types/text/html
>
> What is a mime (media) type? Does it allow parameters?
>
> The RFCs don't define the term "media type", and it is used in
> contradictory ways, including:
>
>  + A top level type, eg: "text"
>  + A two-part type, eg: "text/plain"
>  + A full type with parameters, eg: "text/plain; format=flowed"
>
> Does Atom allow parameters in its @type attribute?

I don't think so: see paragraph 4.1.3.1

"but MUST NOT be a composite type"

>
> How should RDF vocabularies refer to mime types that include
> parameters?
>
> It is all a bit vague, I find it easiest just to model the mime type
> as a Literal property.

agreed for non TEXT relations. Text relations can clearly be  
immediately converted to Literals,

>
> -- 
> Dave

[1] https://sommer.dev.java.net/atom/
Received on Monday, 26 June 2006 22:28:15 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:41:52 UTC