Re: [cmis-comment] Comments on CMIS link relations

[Mark Nottingham <mnot@mnot.net>]

Hi Al,

Responses inline.

On 06/05/2009, at 1:09 AM, Al Brown wrote:

> Mark,
>
> Thanks for taking the time for reviewing. You raise some interesting  
> points. I'd like to address them individually:
>
> 1. The spec has a section on IANA registration that defines the link  
> relations and general concept they represent. The spec then uses  
> those link relations and binds them (for the CMIS specification) to  
> specific media types. The links to be registered are not specific to  
> a media type but rather used to model concepts not yet in the link  
> draft you specified.
>
Ah, good.

> 2. Both up and service are not defined in that draft your provided  
> so I am not sure I understand them completely. As long as up is  
> defined to include the up 'document' in 1 hierarchy and if more than  
> 1 hierarchy then a resource describing the collection of up  
> documents then that works. The 1..n aspect is important.  Also, in  
> CMIS, repository maps to a workspace in a service a document, so  
> somehow that association needs to be kept.
>
Sorry about that; both were missed in my last draft, and will be in  
the next one. They're already in the registry;
   http://www.iana.org/assignments/link-relations/link-relations.xhtml


> 3. Agreed. For those that are specific, either URI format or a  
> general purpose link relation that fits will be chosen. On the two  
> you mentioned, we have an issue to remove stream. On allowable  
> actions, it is used to represent a document that describes the  
> actions that can be taken on resource. That concept does not seem to  
> be modeled yet. Would 'actions' be better term for this?
>
I'm not sure. Does the document describe the specific semantics that  
are allowed (potentially in an extensible way), or is it more focused  
on conveying permissions?

Cheers,


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