On Jan 26, 2009, at 1:44 PM, Gili wrote:
>
> Originally I thought a 300 response is issued if the client did not
> issue the
> right kinds of headers and say multiple content-types are available.
> I no
> longer think this is the case because the specification talks about
> the
> server returning multiple URLs so this isn't a simple matter of
> headers
> against the original URL.
>
My understanding is a 300 response is about *related representations*
not about related resources i.e. the former is more specific than the
latter, and i my assumption is your use-case fits into the latter
category. Is my assumption correct?
> Can you please give a concrete example of when a 300 response would be
> appropriate?
>
A request to "/content" returns distinct URLs:
/content.xml
/content.json
Or returning the languages supported when the accept language header
is not supplied.
In such circumstances the server could also return a Location header
with it's preference for automatic redirection by clients.
Paul.
> Gili
>
>
> Paul Sandoz wrote:
>>
>> Hi Gili,
>>
>> My understanding is a 300 response is returned when the client may
>> not
>> send enough information to indicate what it wants from the server.
>> From section 10.3 [1]
>>
>> This class of status code indicates that further action needs to be
>> taken by the user agent in order to fulfill the request. The
>> action required MAY be carried out by the user agent without
>> interaction with the user if and only if the method used in the
>> second request is GET or HEAD. A client SHOULD detect infinite
>> redirection loops, since such loops generate network t
>> traffic for each redirection.
>>
>> In that respect i think it does fit in with the current model. For
>> example, the developer is using the client API and has written code
>> to
>> expects a 200 response, so a 300 would be the exception in this
>> respect. We want to avoid the client having to switch on a status
>> code
>> for the common cases.
>>
>> But it is a bit of a grey area. A 202 Accepted could return a
>> different representation to that of a 200 OK.
>>
>> However from your use-case it seems to me that "similar tags" are
>> actually different but *related* resources rather than the requested
>> resource "corresponding to any one of a set of representations" for
>> the purposes of agent driven content negotiation. Thus, IIUC your
>> use-
>> case, i think you may be using a 300 for something different than it
>> is intended.
>>
>> In that respect a 200 response with hypermedia defining the
>> relationship seems fine. Namely, given a tag you want to get tag-
>> based
>> resources that are related by "similarity", for example:
>>
>> http://.../<tag>/similarTo
>>
>> or
>>
>> http://.../tagSimilarTo?tag=<tag>
>>
>> or
>>
>> http://.../tags/<tag>
>>
>> where the representation returns also contains some related tags as
>> well as additional information on the tag.
>>
>> and "it is business as usual" from the perspective of the client
>> GETing a representation and following links in that representation
>> that are related in some manner.
>>
>> Paul.
>>
>> [1] http://greenbytes.de/tech/webdav/rfc2616.html#status.3xx
>>
>> On Jan 24, 2009, at 5:31 PM, Gili wrote:
>>
>>>
>>> Upon further reflection, perhaps 200 OK is meant to be used for
>>> resources
>>> that are expected to return multiple representations, whereas 300
>>> MULTIPLE
>>> CHOICES is meant to be used for resources that aren't expected to do
>>> so.
>>>
>>> For example:
>>> http://example.com/myDirectory/ should return 200 OK with a list of
>>> nested
>>> files
>>>
>>> http://example.com/tags/NewYorc should return 300 MULTIPLE CHOICES
>>> and
>>> suggest "similar" tags the client may wish to investigate.
>>>
>>> This would fall in line with Jersey's treatment of 300 as an
>>> exception. What
>>> do you think?
>>>
>>> Gili
>>>
>>>
>>> Gili wrote:
>>>>
>>>> Hi,
>>>>
>>>> I wanted to discuss the fact that Jersey throws
>>>> UniformInterfaceException
>>>> if the return-code is equal to or greater than 300.
>>>>
>>>> I started a discussion at
>>>> http://stackoverflow.com/questions/401191/how-to-return-random-items-restfully
>>>> and someone suggested that return-code 300 is appropriate for a
>>>> resource
>>>> whose state consists of links to sub-resources. In such a case,
>>>> HTTP 300
>>>> would not constitute an error at all. What are your thoughts on the
>>>> matter?
>>>>
>>>> Thanks,
>>>> Gili
>>>>
>>>
>>> --
>>> View this message in context:
>>> http://n2.nabble.com/300-MULTIPLE-CHOICES-tp2209176p2209401.html
>>> Sent from the Jersey mailing list archive at Nabble.com.
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe_at_jersey.dev.java.net
>>> For additional commands, e-mail: users-help_at_jersey.dev.java.net
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_jersey.dev.java.net
>> For additional commands, e-mail: users-help_at_jersey.dev.java.net
>>
>>
>>
>
> --
> View this message in context: http://n2.nabble.com/300-MULTIPLE-CHOICES-tp2209176p2217683.html
> Sent from the Jersey mailing list archive at Nabble.com.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_jersey.dev.java.net
> For additional commands, e-mail: users-help_at_jersey.dev.java.net
>