Re: [weirds] feedback on draft-designteam-weirds-using-http

[Andy Newton <[e-mail]>]

On May 17, 2012, at 3:34 PM, Julian Reschke wrote:

> On 2012-05-17 20:57, Andy Newton wrote:
>> 
>> On May 17, 2012, at 1:39 PM, Julian Reschke wrote:
>> 
>>> On 2012-05-17 19:32, John Levine wrote:
>>>>> Using URI templates sound promising.
>>>> 
>>>> I sure hope not.  URI templates are a reasonable way to try to put a
>>>> consistent layer on top of a bunch of existing inconsistent http
>>>> services.  But in this situation, where we're able to spec the URIs
>>>> from scratch, they're just an extra level of unneeded complexity and
>>>> cruft.
>>>> 
>>>> R's,
>>>> John
>>> 
>>> I have no problem with the protocol hardwiring things into URIs, if that can be done. But please stay away from the term "restful" then.
>> 
>> Can you elaborate on this? Are you making a REST purity argument? I'm confused.
>> Sorry to be thick headed here.
> 
> If you hard-wire the URI format, and there's no discovery whatsoever, you don't take advantage of <http://en.wikipedia.org/wiki/HATEOAS>.
> 
> Now my experience is that most of the time, discussions about whether something is "restful" aren't productive. It's not essential that everything is "restful". What's essential is that it works.
> 
> That's why I initially said that you should stay away from the term unless you really plan to fully embrace it :-).

We are using the term RESTful and not REST, which is generally interpreted as borrowing from REST but not being totally compliant to the principles.

-andy