OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

xri message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Comparison Rules


So first, I'm sorry it's taken me so long to get this last set of  
changes done... it's taken things a little while to settle down.  I  
have a couple of questions regarding these comparison rules, that  
weren't obvious from the minutes...

On Jan 28, 2010, at 6:26 PM, Drummond Reed wrote:

> a) Comparison of Subject and Alias values MUST based on the  
> comparison rules
> defined in RFC 3986, specifically section 6.2.3, Scheme-Based  
> Normalization.
>
> b) Comparison of Link/rel values MUST be as defined in Web Linking.
>
> c) Comparison of Link/type values MUST be after normalization  
> according to
> the rules for media types in section 3.7 of RFC2616.
>
> d) Comparison of Property and Link/Property values MUST follow the  
> same
> rules as for Link/rel values (above).

Is this for the 'type' attribute of the Property element (a URI value)  
or for the actual Property value (a string value) ?  Or both?  I would  
think it only actually makes sense for the 'type' attribute, and that  
the comparison rules for the Property value would be inferred based on  
the type value?


> e) In section 1, we should add text that says that comparison rules  
> for
> strings or URI values defined by extensions MUST be defined by the
> extension.

Judging from the above, it doesn't look like we have general  
comparison rules that can be included in section 1, is that right?   
Comparison seems to be pretty specific to each individual element.  As  
such, it seems strange to state in section 1 that extensions must  
define comparison rules for string and URIs, when we aren't doing so  
ourselves (at least not at this point in the spec).  It seems that  
this would fit better in section 3.2 "Schema Extension"?  And I'm not  
sure that a MUST is necessary... comparison rules are only necessary  
for those values that are expected to be compared.  How about relaxing  
this to a SHOULD?  Or state that comparison rules MUST be specified  
for those string or URI values that are expected to be compared?

-will


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]