W3C home > Mailing lists > Public > public-xg-ssn@w3.org > December 2010

Section 5 of final report: reviewing

From: <Kerry.Taylor@csiro.au>
Date: Fri, 17 Dec 2010 16:04:51 +1100
To: <coryhenson@gmail.com>, <Laurent.Lefort@csiro.au>
CC: <public-xg-ssn@w3.org>
Message-ID: <98CC33188056EF4996107897D51E15AA03072F5BA1@EXNSW-MBX05.nexus.csiro.au>
SSN-XG:  Cory and Laurent in particular,

I am reviewing the final report.  I've fixed some minor things in section 5 - but these below  are a bit harder.
Perhaps Cory and/or Laurent could have a look at these.  Laurent might not see this rapidly - but
if you (Laurent) can respond broadly (most especially point 8 below) that would help. Cory - I think you can see that some of these
things would best be handled by you.

I'm working through section by section - I'll move on to  Section 4 (Ontology)  next and other sections before returning to this one.

Kerry
---------------------------



SECTION 5: Work on Semantic Markup
http://www.w3.org/2005/Incubator/ssn/wiki/Report_Work_on_Semantic_Markup


I've fixed some minor things in section 5 - but these are a bit harder:


0.       "Work on ..." is a poor title.  Just "Semantic Markup" is good.



1.       The whole section 5 should be more careful about use of the word "concept". Where something that is formally an owl:Class
Is meant, then the word "class"  should be used instead. e.g below:


Describing XLink annotation as a semantic annotation, xlink:href would be used to point to an instance of a concept in an ontology and xlink:role would be used to point to a concept in an ontology.

2.      Similarly, in this case  above, for example, "instance of a concept" should surely be "individual".  An "instance of a concept"
would probably be interpreted as  meaning a "triple" (where the property/predicate  is rdf:type), whereas I think "owl:Individual" is probably meant. Can the word "individual" be used?
.


3.      Section 5.3 is missing



4.      There is a problem with the table in 5.4.  "xlink:arcrole" is used for two different purposes.  All the examples show only the model reference purpose.  The "ontology object property"  purpose seems like it needs to exist - but shouldn't it be another xlink attribute? And doesn't it then need two  individuals - so how would that work?  And also what about datatype properties too?   Perhaps one way out of this at this stage is just to explicitly state that there *is no way* to annotate with ontology properties (which seems like a reasonable thing anyway to me).



5.      I recommend dropping 5.5.4  and 5.2.1  and 5.2.3 as unwritten and also interesting but unnecessary .



6.      Section 5.5.3 inference example - seems to be missing the xlink:arcrole attribute. Actually  seems to be missing everything that is recommended just above - including using a different (knoesis) sensor ontology. The latter is ok but needs to be specially noted to avoid confusion. This is an interesting  example, but not a necessary one, so one way to handle this could be to  just  remove it.


7.       In sec  5.5.2: the table "Attribute, Definition"  and footnotes from sec 5.3 is repeated  (this time, correct, I think). But it shouldn't be here.



8.       I've only looked at the "report" so far, not the "deliverable". I am really confused - why is the deliverable a separate document that repeats great parts of the report?  They should be the *same thing* and any summary of "deliverables" just refers to this section of the report. The trouble is, they both seem to have been edited in parallel.  I propose to REMOVE THE DELIVERABLE http://www.w3.org/2005/Incubator/ssn/wiki/Semantic_Markup_Draft_Deliverable and move extra  content I can find there (which seems to be the untitled introduction) to the Report.  I would also NOT preserve the  sections of the "draft deliverable"  called   "status of this document",  "term Definitions" and "motivating use cases" as these seem inappropriate here.  Use cases are given in section 3 of the report.  I realize this might remove some good work that Laurent and/or others  have done, but  it seems necessary for rapid completion at this stage.



9.       There are some slight strangenesses with thing s that appear to be headings e.g. "Sensor Discovery Use Case:" and "SOS GetCapabilities -" and "Sensor Availability". Not sure what to do here - but it is readable as it is.











----------------
Received on Friday, 17 December 2010 05:05:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 17 December 2010 05:05:45 GMT