Re: reverse mapping for xsd:integer vs xsd:nonNegativeInteger

Then perhaps we should adjust both. We're working in an environment  
where all the specs are interlinked and I think it benefits  
understandability if we use consistent language wherever possible.

In any case, the syntax document is a different context. There, we get  
to define things, as it is our own syntax, some of which we have  
created, and some of which we have borrowed. In the RDF Mapping, we  
are talking about the RDF syntax, (which borrows from XML Schema). So  
in that case, I still maintain, we should use the language that they  
use.

-Alan

On May 29, 2008, at 8:43 AM, Boris Motik wrote:

> Hello,
>
> We use the term "constant" throughout the spec in all other places,  
> cf. http://www.w3.org/2007/OWL/wiki/Syntax#Constants.
>
> Regards,
>
> 	Boris
>
>> -----Original Message-----
>> From: Alan Ruttenberg [mailto:alanruttenberg@gmail.com]
>> Sent: 29 May 2008 13:20
>> To: Michael Schneider
>> Cc: Boris Motik; Peter F. Patel-Schneider; public-owl-wg@w3.org
>> Subject: Re: reverse mapping for xsd:integer vs  
>> xsd:nonNegativeInteger
>>
>>
>> On May 29, 2008, at 6:48 AM, Michael Schneider wrote:
>>
>>> @Alan: Good idea!
>>>
>>> This also leads to a somewhat better alignment between OWL DL and
>>> OWL Full:
>>> In OWL 1 Full, the semantic conditions for cardinality restrictions
>>> included
>>> statements like
>>>
>>>  "y is a non-negative integer"
>>>
>>> I was always a bit unhappy with this, because I thought that this
>>> "sloppy"
>>> formulation does not map to concrete triples. But now it looks
>>> reasonable to
>>> me to just keep it in this form, and understand it in the way
>>> analog to what
>>> Boris just put into the DL spec. That's more flexible, and it even
>>> allows to
>>> use custom datatypes, which provide the ability to express non-
>>> negative
>>> numbers. (I now suspect that it was always meant this way in OWL
>>> Full.)
>>
>> I'm not sure that we want to allow custom datatypes. This is parsing
>> we're talking about, and I worry that might be too high a burden to
>> expect parsers to be aware of the full xml schema datatype system.
>> That's why I think  that the set of datatypes should be explicitly
>> bounded in the parsing document, as Peter suggested. Doing so makes
>> it very clear what a parser writer needs to know about. Boris doesn't
>> like the verbosity of that, I'm assuming. How about referring to
>> http://www.w3.org/TR/xmlschema-2/#built-in-datatypes and saying that
>> any built-in datatype that is integer or a subtype is valid.
>>
>> Also, Boris, you use the term "constant" to describe these.
>> "following notation is used to denote parts of the patterns that are
>> matched to constants with integer value". It would be more precise to
>> use the phrase - these are "datatype literals", or "xml schema
>> datatype literals".
>>
>> -Alan
>>
>>>
>>> Cheers,
>>> Michael
>>>
>>>> -----Original Message-----
>>>> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-
>>>> request@w3.org]
>>>> On Behalf Of Boris Motik
>>>> Sent: Thursday, May 29, 2008 12:22 PM
>>>> To: 'Alan Ruttenberg'
>>>> Cc: 'Peter F. Patel-Schneider'; public-owl-wg@w3.org
>>>> Subject: RE: reverse mapping for xsd:integer vs
>>>> xsd:nonNegativeInteger
>>>>
>>>>
>>>> Hello,
>>>>
>>>> I changed the spec along these lines; here is the diff:
>>>>
>>>> http://www.w3.org/2007/OWL/wiki/index.php?
>>>> title=Mapping_to_RDF_Graphs&di
>>>> ff=8203&oldid=8201
>>>>
>>>> I simplified the notation a bit: we now have POS_INT(n) that
>>>> matches to
>>>> any positive integer, and NN_INT(n) that matches to any
>>>> nonnegative integer. There are no other patterns that need to  
>>>> match a
>>>> particular constant value, so this notation should be
>>>> sufficient (for now at least).
>>>>
>>>> Regards,
>>>>
>>>> 	Boris
>>>>
>>>>> -----Original Message-----
>>>>> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-
>>>> request@w3.org] On Behalf Of Alan Ruttenberg
>>>>> Sent: 29 May 2008 05:51
>>>>> To: Boris Motik
>>>>> Cc: 'Peter F. Patel-Schneider'; public-owl-wg@w3.org
>>>>> Subject: Re: reverse mapping for xsd:integer vs
>>>>> xsd:nonNegativeInteger
>>>>>
>>>>>
>>>>> I wonder if it would be better to not explicitly mention
>>>>> xsd:nonNegativeInteger in the mapping rules and instead do  
>>>>> something
>>>>> like this:
>>>>>
>>>>> Instead of writing
>>>>>
>>>>> _:x owl:maxQualifiedCardinality "n"^^xsd:nonNegativeInteger
>>>>>
>>>>> write:
>>>>>
>>>>> _:x owl:maxQualifiedCardinality  XSDL(n, integer > 1)
>>>>>
>>>>> And then explain that XSDL(n, integer > 1) is any xml schema
>>>>> datatype
>>>>> (from a limited set that Peter enumerated) literal whose value  
>>>>> is an
>>>>> integer > 1.
>>>>>
>>>>> -Alan
>>>>>
>>>>> http://www.w3.org/TR/xmlschema-2/#datatype
>>>>>
>>>>> On May 28, 2008, at 2:53 PM, Boris Motik wrote:
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> I added a slightly less verbose wording; here is the diff:
>>>>>>
>>>>>>
>>>> http://www.w3.org/2007/OWL/wiki/index.php?
>>>> title=Mapping_to_RDF_Graphs&di
>>>> ff=8171&oldid=8131
>>>>>>
>>>>>> It seems to me that we don't need to list all combinations of the
>>>>>> datatypes, as we can just simply refer to the usual equality of
>>>>>> datatypes. Please let me know if you consider this insufficient
>>>>>> and/
>>>>>> or unclear.
>>>>>>
>>>>>> Regards,
>>>>>>
>>>>>> 	Boris
>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-
>>>> request@w3.org
>>>>>>> ] On Behalf Of Peter F. Patel-
>>>>>>> Schneider
>>>>>>> Sent: 28 May 2008 10:57
>>>>>>> To: boris.motik@comlab.ox.ac.uk
>>>>>>> Cc: alanruttenberg@gmail.com; public-owl-wg@w3.org
>>>>>>> Subject: Re: reverse mapping for xsd:integer vs
>>>>>>> xsd:nonNegativeInteger
>>>>>>>
>>>>>>>
>>>>>>> This sounds correct to me.
>>>>>>>
>>>>>>> The wording could be something like
>>>>>>>
>>>>>>> 	When parsing literals in G, literals that use XML Schema
>>>>>>> 	Datatypes derived from xsd:decimal and that are result in  
>>>>>>> values
>>>>>>> 	acceptable for the pattern are parsed as if they used the
>>>>>>> 	particular datatype in the pattern, e.g., "0"^^xsd:integer is
>>>>>>> 	acceptable used when parsing a maximum cardinality restriction
>>>>>>> 	(but not when parsing an n-ary datatype declaration).  The
>>>>>>> 	dataypes allowed here are xsd:decimal, xsd:integer,
>>>>>>> 	xsd:nonPositiveInteger, xsd:long, xsd:int, xsd:short, xsd:byte,
>>>>>>> 	xsd:nonNegativeInteger, xsd:unsignedLong, xsd:unsignedInt,
>>>>>>> 	xsd:unsignedShort, xsd:unsignedByte, and xsd:positiveInteger.
>>>>>>> 	Note that using datatypes in this way is not related to using
>>>>>>> 	these datatypes as OWL dataranges.
>>>>>>>
>>>>>>> This would be placed near the beginning of Section 3.
>>>>>>>
>>>>>>> peter
>>>>>>>
>>>>>>>
>>>>>>> From: "Boris Motik" <boris.motik@comlab.ox.ac.uk>
>>>>>>> Subject: RE: reverse mapping for xsd:integer vs
>>>>>>> xsd:nonNegativeInteger
>>>>>>> Date: Wed, 28 May 2008 09:32:13 +0100
>>>>>>>
>>>>>>>>
>>>>>>>> Hello,
>>>>>>>>
>>>>>>>> This is a problem of equality between datatype constants:
>>>>>>>> "1"^^xsd:integer is in fact equal to
>>>>>>>> "1"^^xsd:nonNegativeInteger. Covering
>>>>>>>> all possible equal lexical forms would be really hard: how  
>>>>>>>> about
>>>>>>>> "1.0"^^xsd:decimal? Or "1"^^xsd:positiveInteger?
>>>>>>>>
>>>>>>>> I believe we just simply need to say that, when matching the
>>>>>>>> mapping rules, we need to match them
>>>>>>> "modulo constant equality".
>>>>>>>>
>>>>>>>> Regards,
>>>>>>>>
>>>>>>>> 	Boris
>>>>>>>>
>>>>>>>>> -----Original Message-----
>>>>>>>>> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-
>>>> request@w3.org
>>>>>>>>> ] On Behalf Of Alan
>>>>>>> Ruttenberg
>>>>>>>>> Sent: 28 May 2008 04:32
>>>>>>>>> To: OWL Working Group WG
>>>>>>>>> Subject: reverse mapping for xsd:integer vs
>>>> xsd:nonNegativeInteger
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> In http://www.w3.org/TR/owl-semantics/mapping.html, it says:
>>>>>>>>>
>>>>>>>>> For the purposes of determining whether an RDF graph is an
>>>>>>>>> OWL DL
>>>>>>>>> ontology in RDF graph form, cardinality restrictions are
>>>> explicitly
>>>>>>>>> allowed to use constructions like "1"^^xsd:integer so long as
>>>>>>>>> the
>>>>>>>>> data value so encoded is a non-negative integer.
>>>>>>>>>
>>>>>>>>> Therefore, for backwards compatibility, should the reverse
>>>> mapping
>>>>>>>>> explicitly have a mapping for the (non qualified) cardinality
>>>> cases
>>>>>>>>> where it currently only says xsd:nonNegativeInteger?
>>>>>>>>>
>>>>>>>>> -Alan
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>
>

Received on Thursday, 29 May 2008 13:21:14 UTC