Re: LDP-Server - Issue-58

Arnaud, Henry,


On Thu, Jun 13, 2013 at 5:00 PM, Arnaud Le Hors <lehors@us.ibm.com> wrote:

>  From my point of view, we're not just defining a vocabulary or ontology
> people can use with an existing triple store. We're also defining an
> interaction model in a the form of requirements that requires one to write
> code for in a server to be compliant.
>
> LDP isn't only about RDF. It is also about REST, which clearly involves
> servers and clients. We're not only defining specific states but also how
> to transition from one state to another.
>

I completely agree with Arnaud on that.
However...


> How do you do that without talking about servers?
>

I think Henry's point is that LDP-compliance is not a property of the
server as a whole...
An HTTP server manages a number of resources (roughly, one per URI for
which the server does not return 404);
some of those resources might conform with the spec (either as LDPR or
LDPC), but other may not (e.g. /favicon.ico).

The original text:

> [[
> A conforming LDP Server is an application program that processes HTTP
requests and generates HTTP responses
> that conform to the rules defined in sections onLDPRs and LDPCs
> ]]

may seem to imply that an LDP server manages *only* LDPRs : one might read
it as "that processes any HTTP request", while the intention is, I guess,
"that processes some HTTP requests"

Note that Henry's proposal does talk about the server:

> [[
> A conforming LDP Server is an HTTP Server that contains one or more
> LDPC resources that conform
> to the rules defined in this spec.
> ]]

but aims at making it clearer that LDP-compliance may concern only a subset
of the resources.

Note that two details are bothering me in Henry's definition, though:

* I'm assuming he means LDPR rather than LDPC
* I would prefer "Server that manages" than "Server that contains", but
that's editorial, really

  pa