W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2011

Review of Graph Store Protocol (action 564)

From: Andy Seaborne <andy.seaborne@epimorphics.com>
Date: Fri, 02 Dec 2011 12:09:04 +0000
Message-ID: <4ED8BFE0.2050301@epimorphics.com>
To: SPARQL Working Group <public-rdf-dawg@w3.org>
This completes ACTION 564 [http://www.w3.org/2009/sparql/track/actions/564]

	Andy



== Comment 1

Sandro write:
 > title
 >
 >          I know we went thought a long WG decision process (twice) to
 >          arrive at the current title, but in actually talking about this
 >          document to a few people, I find the only way to have it make
 >          any sense is to use the word "RESTful".    So, I propose we
 >          amend the title to:
 >
 >          SPARQL 1.1 Graph Store (RESTful) HTTP Protocol



I support this suggestion.

Or even

"SPARQL 1.1 RESTful Graph Store Protocol"

Reading the doc, it says "REST" in sentence 1 of the introduction.

Saying "REST" conveys HTTP to me so both aren't needed.

== Comment 2
== Introduction
= Editorial/Not required for LC/required for CR

The list of other SPARQL documents is incomplete (no JSON results, no 
entailment).  I thought we were going to use standard boilerplate in all 
docs to reference the set of documents.

== Comment 3

My preference would be to use

Accept: text/turtle; charset=utf-8

(this is already registered).

for

Accept: application/rdf+xml

== Comment 4
== IRI vs URI

= Section 4.1


The picture does not mention IRI but it's (IRI, RDF) element of graph store.

= Section 4.2

"where the IRI is the result of percent-decoding"
but the percent-encoding is about the ":"

Suggestion: an example with an accented character.

"In the example above, the embedded graph IRI" => confusing

"embedded" => "encoded"

"embedded URI"
=> "query string URI"


== Comment 5
"If the Accept header is not provided... the server SHOULD return RDF XML"

I prefer should "return one of RDF XML, Turtle or N-Triples".  Turtle is 
proving more acceptable (pun) as a serialization of RDF.

== Comment 6
Order of operations.

Could we have HTTP GET first?  It's an operation that applies even if 
the others are not allowed.

== Comment 7
= section 5.2
Bold DROP ==> <tt>DROP</tt>

Bold is used for the HTTP verbs in this section.

== Comment 8
= section 5.2

"If the request body is empty"

If it's content-type: rdf+xml, then an empty request is illegal by 
definition of the content type.  We should not be rewriting HTTP conneg 
rules.

== Comment 9
= section 5.3
"In the event the operation is overridden, ..."
I don't understand who is overriding the operation.

== Comment 10
== section 5.4
= Multipart.

Make this 5.4.1 or place as subsection earlier because it applies to PUT 
as well, presumably.

== Comment 11
== Discussion of service description

Accept: application/rdf+xml

is followed by Content-Type Turtle.

s!Accept: application/rdf+xml!Accept: text/turtle; charset=utf-8!

== Comment 12
== End section 5.4

Include query over POST in POST operations.

== Comment 13

[SPARQL] reference is to SPARQL 1.0.

Some RFC references mention a URL but it's not a link.
Some RFC references don't spell out the URL.
Received on Friday, 2 December 2011 12:09:48 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:47 GMT