Re: Proposed HTTP SEARCH method update

> On 17 Apr 2015, at 11:09, Julian Reschke <julian.reschke@gmx.de> wrote:
> 
> On 2015-04-17 10:37, henry.story@bblfish.net wrote:
>> 
>>> On 10 Apr 2015, at 18:00, James M Snell <jasnell@gmail.com> wrote:
>>> 
>>> Please see: http://datatracker.ietf.org/doc/draft-snell-search-method/
>>> 
>>> Comments welcome.
>> 
>> [[
>>   The payload returned in response to a SEARCH cannot be
>>   assumed to be a representation of the resource identified by the
>>   effective request URI.
>> ]]
>> 
>> I think that should rather be
>> 
>> [[
>>   The payload returned in response to a SEARCH is a partial
>>   representation of the resource identified by the effective
>>   request URI.
>> ]]
> 
> I don't think that's the intention, nor would it be consistent with the current use of SEARCH.
> 
>> The way I think one has to understand SEARCH is in the context of the
>> other the two other HTTP verbs:
>> 
>> - GET: returns a representation of the effective request URI
>> - PUT: changes the representation of the effective request URI to the new representation
>> 
>> then SEARCH is like PATCH just a way of efficiently doing GET and PUT
>> 
>> - SEARCH: allows on to get a part of the full representation of the request URI ( using whatever is the
>>  most appropriate query language )
> 
> Depending on the query language, the result might just be a list of URIs. I don't see how to interpret this as partial representation...

Ah yes, I had forgoten that case. 

In that case can one not think of the query combined with the answer as forming a partial 
representation of the resource?

In your hypotehtical SPAQRL example you request parts of the remote table. If the client 
made a number of such requests, say asking for each column individually, it could in the 
end re-constitute the whole table. ( assuming the etag had not changed )

It would be interesting if the client could make such deductions from a number of 
queries.



> 
>> - PATCH: allows one to replace part of the representation of the request URI
>> 
>> 
>> It may be that by making this clear in the document it becomes easier for reviewers
>> to understand that of course this does not require a representation to be specified, the
>> same way that the GET verb does not require one.
>> 
>> Also I am not sure what in 4.2 you mean by a ""SPAQRL query.
>> 
>> 
>> Henry
>> 
>> Social Web Architect
>> http://bblfish.net/
> 
> Best regards, Julian

Social Web Architect
http://bblfish.net/

Received on Friday, 17 April 2015 10:14:13 UTC