← Prev in month ← Prev in thread
Next in thread → Next in month →

Some observations on obix v1.1 PR doc (WD38)

From
Gemmill, Craig <>
Date
2014-12-11T18:34:34+00:00
ID
Thread
Some observations on obix v1.1 PR doc (WD38)
In working on the white paper for example OBIX implementations, I came across a few issues with the current PR doc.  Please don’t reply to this email as it will be better to address them with separate email threads.  I will shortly create
 JIRA issues for each of the items, and then we can launch separate threads if need be, to discuss each of them.

 

BTW, Toby, can you create a PRD02 component for the various docs, so I can mark these items against that?  In the interim, I can mark it against WD38.

 

·        
Section 5 is somewhat inconsistent – the Lobby contract defines bindings as a list of str, but in the examples later in Section 5.5.3 it describes the bindings as a list of uri.  There are a few other inconsistencies, of the same
 nature, in Section 5, which ought to be cleaned up.

·        
I noted that the requirement to put the version and/or retrieval time in as a child element sort of violates the rule about Extent in 9.3, which is that you only use non-refs for primitive children with no further extent; complex
 children should generally be represented with refs.  Conceptually I think it’s important not to have to ask again for these additional details, but it is the first place a client looks, and it violates the rule.  The ‘rule’ is a should, not a must, so it’s
 not a definite problem, but still worth noting and considering.

·        
The namespace location seems to be different for the core, than what is referenced in the encodings doc – not sure if that’s okay, or if one is (or both are) wrong:

o  
core: 
http://docs.oasis-open.org/ns/obix/201410

o  
encodings: 
http://docs.oasis-open.org/obix/ns/201312/schema (for the core spec), and 
http://docs.oasis-open.org/obix/ns/{short-identifier and version}  (for the encodings spec)

o  
at the least, I think a specific version needs to be used in the encodings spec, and this should be the same for the core as for the companion standards.

·        
What should the xsi:schemaLocation be in the xml root object’s attributes?  This seems unclear to me from what we have in the specification so far.

 

Thanks

Craig
← Prev in month ← Prev in thread
Next in thread → Next in month →