Re: URI Template experience

On May 23, 2009, at 8:38 PM, Joe Gregorio wrote:

> On Sat, May 23, 2009 at 1:46 PM, Jan Algermissen
> <algermissen1971@mac.com> wrote:
>> Hi Joe, hi Roy,
>>
>> sorry if this has been discussed already, i have not been able too  
>> follow
>> the lists closely for a couple of months.
>>
>> AFAIK, there is the idea of using URI templates in the HTTP Link  
>> header,
>> which would allow the server to instruct clients how to construct  
>> URIs based
>> on a set of named parameters (not sure of 'parameter' or 'variable'  
>> should
>> be used here).
>>
>> E.g. (adapted from [1]):
>>
>>        Link: <http://www.example.com/home/{userid}>; rel="home"
>>
>> AFAIKT, this requires shared understanding of the property name  
>> between
>> client and server. Wouldn't it be a good idea to establish a  
>> registry of URI
>> template property names to support this need for shared  
>> understanding?
>
> I'm suspicious of registries to begin with,
> and in this case even if I
> were supportive, I
> believe a registry would be far outside the scope of the URI Template
> specification.

Yes. Another point though is a potential mapping of the property names  
to URIs so client implementations could be coded using these URIs  
instead of the short names. What do you think about baking into the  
draft a proposed way for the client to find the mapping? E.g.

Link: <http://www.example.com/home/{userid}>; rel="home"
Link: <http://www.example.com/property-mapping}>;  
rel="templatePropertyMapping"

Thinking about it, the specification of the link relation would do -  
no 'need' to put it into the draft.

Jan



>
>
>   -joe
>
>>
>> Depending on the question of course, whether a reasonable set of  
>> generally
>> usable properties can be determined or if they are just too  
>> application
>> specific.
>>
>>
>> [1] http://www.mnot.net/blog/2006/06/22/link
>>
>>
>> On May 22, 2009, at 4:24 PM, Joe Gregorio wrote:
>>
>>> On Mon, May 18, 2009 at 5:38 PM, Roy T. Fielding <fielding@gbiv.com>
>>> wrote:
>>>>
>>>> Hi Joe,
>>>>
>>>> Your question implies that the features in the current draft are
>>>> somehow dependent on the extent to which the current draft has
>>>> been implemented in the wild.  I think that is backwards, since
>>>> the draft received many comments and did not change as a result.
>>>> For example,
>>>>
>>>> http://www.w3.org/mid/07109D44-233D-42F3-ACB0-56B4A6562903@gbiv.com
>>>>
>>>> So, the answer to your question is that implementors are patiently
>>>> waiting (perhaps too patiently) for the draft to be updated.
>>>> Would it help if I issued a draft with the alternative syntax?
>>>
>>> I asked the question because there are a bunch of implementations  
>>> and if
>>> there was a great attraction to the current syntax beyond {foo}  
>>> then I
>>> wanted
>>> to know that. From what I can tell from the ensuing conversation  
>>> there is
>>> a need
>>> for more complex capabilities beyond {foo}, but no one is in love  
>>> with
>>> the current
>>> syntax. That's good news to me because I prefer your proposed  
>>> system.
>>>
>>> I can update the current draft to your proposal, or you can generate
>>> a draft yourself if you think that will go faster.
>>>
>>>  Thanks,
>>>  -joe
>>>
>>>>
>>>> ....Roy
>>>>
>>>
>>>
>>>
>>> --
>>> Joe Gregorio        http://bitworking.org
>>>
>>
>>
>
>
>
> -- 
> Joe Gregorio        http://bitworking.org

Received on Saturday, 23 May 2009 19:23:37 UTC