Semantics doc review.

Basically a thumbs-up.

Overall, the style is somewhat more conversational than most technical
documents but I rather like that.


Fixes follow: [section 4 has a serious error, everything else is
trivial or typo. By "serious" I mean, "requires fixing", not
"show-stopping"]


TOC and section headings:

- indentation of TOC needs fixing.
- heading for section 2 (not in the TOC) ends with a "."; none of the
	others do.
- capitalisation (throughout): you've got "RDF graphs", but
	"Rdf-entailment" (at a sentence start) and "rdf closures".

Throughout:

- rdfs:XMLLiteral appears. I've not done a global search on the
	document for all bad namespace assignments but there may be
	others.

Section 0.3

- "A subgraph of an RDF graph is simply a subset of the triples in the
	graph. Each triple in a graph is considered to be a subgraph."
	...there's a type clash here (!). Pedants might expect,
	"... Each set consisting of a single triple from the graph is
	considered..." although it's clear from context what's going on,
	and this kind of shorthand seems commonplace elsewhere.

Section 1.1, para 3.

- I note in section 3.3 that subClassOf and subPropertyOf are given
	extensional definitions (iff, not if), not intensional ones,
	despite the nod to intension here and elsewhere.

Section 1.3

- I've put a "!" by the first paragraph about LV, but I can't now
	remember why I had a problem with this - can someone else
	scrutinise this?

Section 1.4

- last but one paragraph: "all literals containing a plain literal"
	should be "all triples containing a plain literal"

Section 3.3

- iff conditions on subClassOf, subPropertyOf are extensional, not
	intensional. Just a note, no fix required.

- given the above, the weak rules for range and domain with
	superproperties seem slightly incongruous. More of a "comment"
	than a problem.

Section 4

- "A graph rdf-entails another just when its rdf-closure simply entails
	it."

	This isn't true as this currently stands, and the RDF entailment
	lemma doesn't hold, because (as I read the document) simple
	entailment doesn't know about rdf:XMLLiteral.

	Therefore, rdf2a and rdf2b don't help if the consequent graph
	contains an XML literal in noncanonical form that differs from
	the form in the original premise graph.

	I think rdf2a and rdf2b need to be replaced with an
	XML canonicalisation operation (replacement, not addition)
	which is applied to both premise and conclusion graphs for
	this to work. Either that, or the interpretation of XMLLiteral
	typed nodes needs to be restricted in the simple interpretation/
	entailment case.

Section 4.2

- mislabelling "rdfs1" as "rdf1"

Appendix A.

- question (I can't find an up-to-date Lbase doc at the time of the
	review): does L_base use "/" for escaping?

Appendix B.

- RDF closure lemma proof requires definition of HP. (It's obvious
	what's going on though and looks ok to me.)

- entailment lemmas ignore XML literal issues I think.


That's everything I could find, nice one Pat (and happy new year).

jan

PS. Apologies if I'm missing something in section 4.


-- 
jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287088 Fax +44 (0)117 9287112 http://ioctl.org/jan/
"...perl has been dead for more than 4 years." - Abigail in the Monastery

Received on Monday, 6 January 2003 10:24:43 UTC