Re: [WNET] New Draft

Hi Ralph,

Thanks for the detailed reading. I've addressed your comments and 
updated the document [1] accordingly, see below.

[1]http://www.w3.org/2001/sw/BestPractices/WNET/wn-conversion-20062304

> I guess "first" is as given in the Princeton source and is not intended
> to be represented anywhere else (e.g. in the RDF), correct?  There
> are later references to the order of the word senses not being significant.

Correct on both accounts. Added:
	[[The %lexform% is the lexical form of the first WordSense of
	the Synset (the first
	WordSense in the Princeton source as signified by its 	
	"wordnumber", see <a href="#distribution">Overview of the
	WordNet Prolog distribution</a>). ]]


> Figure 2: missing hypernymOf and the subproperties for classifiedBy
> and meronymOf.

All the subproperties are present. The hyponymOf property is also 
present. You may be alluding to the fact that the inverse properties are 
not in the list. I leave those out at that point in the document to not 
give the false impression to RDFS-only users that they are indeed in the 
source.

> Typo in first query; the last pattern should be

Thanks for spotting, fixed.

> Typo: "There is one more specific kind of /-adverb-//+adjective+/ called
> an adjective satellite."

Fixed.

> 3. Selecting and Querying ...
> 
> Typo after the list of triples:  "... same as the result /-to-//+of the+/
> following SPARQL query"

Fixed.

> Appendix D
> 
> It's hard to understand which way hypernymOf and hyponymOf go.
> I suggest including some triples here for an example.
> The same is true of the classifiedByTopic, Usage, and Region
> properties as well as the meronymOf subproperties.

Some of them did not have an example from the original source, so added 
those. Concerning the direction: the text states quite unambiguously the 
direction by three statements:
- original source format, e.g.: mm(Synset_ID_A, Synset_ID_B).
- example, e.g. mm(100006026,107463651).  [Person,People].
- Maps to: memberMeronymOf(Synset_ID_A, Synset_ID_B)

I don't think adding triples will make things more clear as these
would look very similar to the "Maps to" part, only then in RDF syntax.

> Appendix F
> 
> At the end of the paragraph comparing this conversion to the
> Neuchatel conversion, there's a typo; this conversion uses
> slash URIs, not hash URIs.

Nasty one. Thanks, fixed.

> Appendix G
> 
> Is '/' the only character that appears in Princeton's dataset that
> needed to be changed?  Are there any other NCname reserved
> characters used?  I gather than "_" is used consistently in place
> of (a) space character(s).

No, there are a few more. The text now reads
[[
Some words contain characters that are not allowed in NCNames. In order 
to generate a correct URI we changed the following characters into 
underscores: '/', '\','(', ')' and ' ' (space).
For example, the URI for the word "read/write_memory" becomes:

http://www.w3.org/2006/03/wn/wn20/instances/word-read_write_memory/
]]

> Appendix H
> 
> Second to final sentence; I think you mean "uniqueness" rather
> than "unicity".

Thanks, fixed.

> Appendix I
> 
> Typo in second paragraph: s/OLW/OWL/

changed to 'an RDF/OWL'

Thanks for the careful reading,
Mark.

Received on Monday, 24 April 2006 09:43:52 UTC