Re: LDPR Interaction Model on Create

Hi John,

On 10/10/2014 11:50 AM, John Arwe wrote:
> Moving this onto the right list for drafting a consensus response.
> 
>> I also have a small comment that I've been meaning to send regarding the
>> interaction model.
>>
>> 5.2.3.4 states that "Clients use the same syntax, that is HTTP Link
>> headers, to specify the desired interaction model when creating a
>> resource as servers use to advertise it on responses."
>>
>> I noticed that in the primer, the POST request to an LDP-BC does not
>> contain a link header expressing the type of the resource to be created.
>> That also seems to be the behaviour of test suite. However, the POST
>> request to an LDP-DC *contains* the Link header Link:
>> <http://www.w3.org/ns/ldp#Resource>; rel="type", while *none* of the
>> examples in the LDP spec show a Link header being sent with POST 
> requests.
> 
> THAT should be fixed, for sure.
> The only way a LDP client gets predictable behavior is by specifying the 
> interaction model.
> That's what the examples should show, period... that which is 
> interoperable.
> 
> 
>> Oh, and one more comment -- the LDP spec does not contain any examples
>> covering the creation of containers.
> 
> That is not automatically a problem. 
> Shirley, the non-normative documents can cover it.
>    ^ if you're too young to recognize that, see [1] 
+1 for good quote.

> 
> 5.2.3.4 says it's allowed.
> 
> When we gnawed on this in the past, some people were mildly against even 
> acknowledging it in 5.2.3.4 on the principle that there's probably a 
> bazillion or so consequences that we don't render explicit - we rely on 
> implementers to be able to read specs and reason about consequences to 
> some degree, or every spec would re-state all content from the transitive 
> closure of all its normative references.
> 
> 
>> In conclusion, going through the spec again, I have the feeling that it
>> is heavily oriented/documented towards "read" operations where "write"
>> operations look more like an afterthought.
> 
> I'm sorry, I cannot parse the concrete proposal to improve that situation 
> from your quoted text.
> What is missing wrt write that is not found in dependent specs?
> I'll grant you that doing the sections by method has its down sides; I 
> fought that battle somewhat pre-Submission and lost, so I've learned to 
> live with it mostly.

I think what's missing is actually a couple of examples in 5.2.3 HTTP
POST, showcasing a POST request with the proper headers, especially for
5.2.3.4 (I feel that's the most confusing subsection).

> 
> I did (somewhat recently) take a stab at re-factoring "create" out as a 
> section, since it has always read to me to be POST-centric.  I found in 
> the end (rough figures) 12 clauses that really had to be specific to Post, 
> 2-3 that applied to post/put/patch, and another 1-2 that were put-only, so 
> I decided benefit << effort.

+100 for effort. I know how frustrating spec writing can be at times.
Keep up the good work! :)

-- Andrei

> 
> 
> 
> [1] http://www.youtube.com/watch?v=0A5t5_O8hdA
> 
> 
> Best Regards, John
> 
> Voice US 845-435-9470  BluePages
> Cloud and Smarter Infrastructure OSLC Lead
> 
>