W3C home > Mailing lists > Public > public-awwsw@w3.org > December 2010

Re: Looking for examples of non-REST information resources

From: Jonathan Rees <jar@creativecommons.org>
Date: Mon, 6 Dec 2010 12:11:57 -0500
Message-ID: <AANLkTimt7xiMUymB3VXyoYQGnQaQtVRkqVj9FWwJ7RHc@mail.gmail.com>
To: David Booth <david@dbooth.org>
Cc: AWWSW TF <public-awwsw@w3.org>
On Sun, Dec 5, 2010 at 2:24 PM, David Booth <david@dbooth.org> wrote:
> On Fri, 2010-12-03 at 17:31 -0500, Jonathan Rees wrote:
>> The subject is sort of a teaser, let me be more precise.
>>
>> I'm looking for examples URIs with the following properties
>>   - GET requests result in 200 responses
>>   - The URI, when used to refer (e.g. as a noun phrase in a sentence
>> or as a subject or object in an RDF statement), seems to refer to
>> something that does not naturally fit the REST mould.
>>
>> By 'the REST mould' I mean what Roy F means by REST resource; my
>> summary would be
>>   - has a current state consisting of "abstract information"
>>   - the state may or may not change through time
>>   - the state can be "represented" by "representations" (bit strings +
>> metadata) synthesized by a server
>>   - the state might be updated by PUT and POST, or by some other process
>>
>> My examples are
>>
>> http://en.wikipedia.org/wiki/Special:Random
>> http://www.random.org/integers/?num=10&min=1&max=6&col=1&base=10&format=plain&rnd=new
>
>
> In what way do you see these as not naturally fitting the REST mold?
> They seem to me to fit it.

What are their states?

See my reply to Henry.

> BTW, the other input that you didn't mention in your REST summary above
> is the client's request, since that can also affect the representation
> that is returned.

I'm not talking about HTTP. I'm talking about REST. Roy has gone to
great pains to separate the two.

>> Can you name more? Are there lots of these resources?  Or only a few?
>> Do we need to be able to use their URIs to refer to them, or should we
>> refer to them in some other way, or not at all? Are 200 responses
>> still "representations" of these resources, or do they bear a
>> different relationship to the resources? Are they "information
>> resources"?
>
> To my mind if a 200 response is returned then the thing *is* an
> "information resource" -- there is no better qualifying criterion.  In
> my view, the reason that the role of "information resource" exists in
> the web architecture is to enable us to talk about the things that have
> "representations" that may be returned in 200 responses.

Please reread my original post. I'm not asking about information
resources, I'm asking about REST resources.

If REST is a term that distinguishes some things from others, one
ought to be able to name something that isn't a REST resource (but
still has associated GET/200 exchanges).  If on the other hand it
vacuous, I guess that's good since we needn't be concerned with it any
longer, but then I haven't a clue what Roy was writing about, and what
people mean when they complain that an interface is non-RESTful.

Jonathan

> David
>
>>
>> If it's necessary in order to formulate answers, pretend you're the
>> URI owner, and are entitled to make decisions about what these
>> resources are.
>>
>> Jonathan
>>
>>
>>
>
> --
> David Booth, Ph.D.
> Cleveland Clinic (contractor)
> http://dbooth.org/
>
> Opinions expressed herein are those of the author and do not necessarily
> reflect those of Cleveland Clinic.
>
>
Received on Monday, 6 December 2010 17:12:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 6 December 2010 17:12:26 GMT