W3C home > Mailing lists > Public > public-owl-wg@w3.org > April 2009

Re: ACTION-333 Quick Review of Quick Reference Guide

From: Rinke Hoekstra <hoekstra@uva.nl>
Date: Thu, 23 Apr 2009 13:46:15 +0200
Cc: Christine Golbreich <cgolbrei@gmail.com>, "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, W3C OWL Working Group <public-owl-wg@w3.org>
Message-Id: <E2876CA9-1BB9-4133-A8A6-CC395778C850@uva.nl>
To: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>
Hi,

Just to be clear on this, it was me who proposed the extra column.  
There were two reasons for this,
1) the occurrence of "(N)" in syntax productions to point to NF&R was  
rather confusing, and
2) I was (still, mistakenly?) under the impression that the web-based  
QRG is to function as an alternative index to multiple documents in  
the OWL 2 specification. I believe we discussed this at F2F1 (may have  
been offline). In this role, having *as many* links as possible is a  
good thing.

On the other hand, I fully agree that a quick ref *card* should  
contain a minimal number of links, and if they do exist, they should  
be inline (as Peter says).

I think the current discussion (at least partly) hinges on a  
different conception of what the role of the QRG/QRC should be. Both  
roles seem perfectly sensible to me. Making the purpose of this  
document explicit would help a lot.

-Rinke



On 23 apr 2009, at 12:11, Ian Horrocks wrote:

> On 23 Apr 2009, at 09:03, Christine Golbreich wrote:
>
>> A first quick comment about links.
>>
>> - NF&R says what a new feature is and not only why a new feature has
>> been added. If it turns out that NF&R has unvoluntary been too much
>> cut and gives a wrong impression, it's urgent to say it. I will  
>> revert
>> and reintroduce some content of previous version.
>
> Please do not do that without discussion -- the vote to publish the  
> document as a WD was made on the basis of the current version.
>
>>
>> - QRG links
>> to simplify, there are 3 profiles of QRG readers : (1) newbies, (2)
>> OWL 1 familiar users, (3) advanced
>>
>> Links useful for
>> (1): to Primer + NF&R for new features
>> (2): to NF&R
>> (3): no "instructional" (Primer or NF&R)  but only "technical"
>> documents i.e. normative specs (syntax and semantics).
>>
>> I suggest to keep all links to NF&R and Primer. Otherwise if you
>> consider that they are "not very useful", remove both links to NF&R
>> and Primer.
>
> I agree with Peter that the extra "links" column is a waste of space  
> and that there are far too many links anyway. IMHO, links to RDF- 
> mapping are clearly inappropriate and *must* be removed; links to  
> NF&R are clearly superfluous (given the purpose of this document)  
> and *should* be removed; links to Primer may be confusing and/or of  
> limited utility and *could* be removed. On the whole, I favour  
> removing as many links as possible.
>
> Ian
>
>
>>
>> 2009/4/23 Peter F. Patel-Schneider <pfps@research.bell-labs.com>:
>>>
>>>       Quick Review of QRG
>>>
>>> I reviewed the version of 22 April 2009.  This version has the link
>>> column up to about Section 2.2.1.
>>>
>>> General comments:
>>>
>>> - Satisfying the following comments will require quite large  
>>> changes to
>>> the document.  A re-review will be needed.
>>>
>>> - Moving the links to a separate column is a bad idea.  It consumes
>>> extra space to no good effect.  Links in the QRG should be from the
>>> relevant information itself, as they used to be.
>>>
>>> - Links to NF&R are not very useful in the QRG.  Users of this  
>>> document
>>> will not be interested in why a new feature had been added to OWL 2.
>>> Instead they will be interested in what the feature is (and thus the
>>> old links from the FS column are useful) and in how to use the  
>>> feature
>>> (and thus the links to the Primer are useful).  I suggest that all
>>> links to NF&R be removed.
>>>
>>> - Links to the RDF Mapping do not provide any useful information and
>>> should also be used.  This document itself provides the information
>>> needed about the mapping, and in a form much more likely to be
>>> comprehensible by the readers of this document.
>>>
>>> - The document is missing named classes, properties, and  
>>> individuals.
>>> Lines like the following should be added:
>>> Named class  U  U
>>> In some places names are required but the document does not
>>> distinguish between names and expressions.  The places that I  
>>> noticed
>>> where names are required are 2.6 Declarations and 2.7 Annotations.
>>>
>>> - This document should use terminology from the normative OWL 2
>>> documents.  This requires at least some changes to the following
>>> terminology:
>>> OWL class -> class expression
>>> object property -> object property expression
>>> I second Christine's comment to use the same short forms as in  
>>> SS&FS,
>>> in general, but it may be that C is better than CE, if only to save
>>> space.
>>>
>>> I suggest the following start of a fix for the notation paragraph:
>>>
>>>   We use the following notation conventions: "C" is a class
>>>   expression, "D" is a data range, "P" is an object property
>>>   expression, "R" is a data property, "A" is an annotation property,
>>>   "a" is an OWL individual, "v" is a literal, and "n" is a
>>>   non-negative integer. All of the previous can have subscripts.
>>>   "_:x" is a blank node.  "(a1 ... an)" is an RDF list.
>>>
>>> Then use subscripts where Q, S, and v were used before.
>>>
>>> - There are some awkward sentences in the document.  I haven't  
>>> changed
>>> them, but a final pass should be made after the content is fixed.
>>>
>>> - Subsections should be used sparingly.  I suggest no sub-sub- 
>>> sections
>>> (2.1.1, etc.) at all.
>>>
>>> - This document does not really need to address n-ary data  
>>> ranges.  I
>>> suggest removing all mention of n-ary data ranges and their support.
>>>
>>> - In some places the document uses explicit iteration (n-ary  
>>> individual
>>> equality) and in some places it uses implicit iteration (equivalent
>>> properties).  Only one should be used.  I suggest implicit.
>>>
>>> - The document is missing some constructs.  I noticed:
>>> - prefixes
>>> - datatype definitions
>>> - names, literals
>>> - individuals - named and anonymous
>>>
>>> - It would probably be better to merge all the datatype and data  
>>> range
>>> stuff into one section.  The built-in datatypes could be more
>>> succinctly presented by just listing the xsd datatypes.
>>>
>>> Specific comments:
>>>
>>> - The abstract needs to be completely rewritten, to be something  
>>> like:
>>>
>>> This document provides a quick reference guide to the OWL 2  
>>> language.
>>>
>>> - Section 1 should be renamed to "Prefixes" and the first sentence
>>> changed to "The standard prefixes in OWL 2 are:" to conform with the
>>> wording in the normative OWL 2 documents.
>>>
>>> - "_:x" is a blank node. All of these can have subscripts.
>>>
>>> - The link for RDF lists should probably point to Turtle or some  
>>> other
>>> place that uses this syntax for RDF lists.
>>>
>>> - "all OWL individuals" -> "universal class" (maybe)
>>>
>>> - The "Every owl:Restriction is an owl:Class." sentence is not  
>>> needed at
>>> all.
>>>
>>> - The presentation of cardinalities should be made easier to  
>>> understand.
>>> I suggest something like
>>>
>>>                   |                              | _:x rdf:type  
>>> owl:Restriction.
>>>                   | ObjectExactCardinalty(n P)   | _:x  
>>> owl:onProperty P.
>>>                   |                              | _:x  
>>> owl:cardinality n.
>>> exact cardinality |  
>>> -----------------------------|------------------------
>>>                   |                              | _:x rdf:type  
>>> owl:Restriction.
>>>                   | ObjectExactCardinalty(n P C) | _:x  
>>> owl:onProperty P.
>>>                   |                              | _:x  
>>> owl:qualifiedCardinality n.
>>>                   |                              | _:x owl:onClass  
>>> C.
>>>
>>> - "<x> Properties are instances of owl:<x>Property" should be  
>>> changed to
>>> just "<x> Property Expressions".  The last bit is not part of the  
>>> FS,
>>> and in any case doesn't add any information.
>>>
>>> - "universal <x> property" and "empty <x> property"
>>>
>>> - The range of a datatype property axiom is a data range, not a  
>>> class.
>>>
>>> - Data range intersection and union are switched up in the table.
>>>
>>> - positive object property assertion needs to indicate that P is an
>>> object property, not an object property expression.
>>>
>>> - A and AP are both used for annotation properties.  Only one  
>>> should be.
>>>
>>> - The section on Annotations needs work.  Object annotations are on
>>> names only, for example.  The full fixing up of this section  
>>> should be
>>> done after some of the above changes are made.
>>>
>>> - The FS for deprecation doesn't need the short form - just write  
>>> out.
>>> Deprecation can be used for individuals as well.
>>>
>>> - The section on ontologies should not say "Annotations of  
>>> Ontologies".
>>>
>>> - This is a quick reference guide and does not need to mention all  
>>> the
>>> arcana.  Therefore the section on deprecation should be deprecated.
>>>
>>> Tyops:
>>>
>>> - I fixed a few typos.  If the document is reverted these may need  
>>> to be
>>> fixed:
>>> Abstract: his -> This
>>> 2: an rdf list.
>>> 2.1: _:x  owl:hasValue a.
>>>      an n-ary data range
>>> 2.2.1: bottom  object property
>>> 2.5: a rdf:type C
>>>
>>>
>>
>>
>>
>> -- 
>> Christine
>>
>



---
Drs Rinke Hoekstra

Leibniz Center for Law      |  AI Department
Faculty of Law              |  Faculty of Sciences
Universiteit van Amsterdam  |  Vrije Universiteit
Kloveniersburgwal 48        |  De Boelelaan 1081a
1012 CX  Amsterdam          |  1081 HV Amsterdam
+31-(0)20-5253499           |  +31-(0)20-5987752
hoekstra@uva.nl             |  hoekstra@few.vu.nl

Homepage: http://www.leibnizcenter.org/users/rinke
Received on Thursday, 23 April 2009 11:46:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 23 April 2009 11:46:58 GMT