OK, so I still think that the comparison is not "based on datatype" --
although in the notes we can say that this only works between values of
the same datatype. To me, "based on datatype" would mean that you decide
to do the comparison based on the datatype of all properties of that
type, which is not the case. (That would be #1). The comparison is not
"based on" but may *use* datatypes. What folks on the call felt was
important was to make it clear that we are talking about comparing
values between two statements. That's what I see in your examples below.
As for the #1 case, there's a question of DSP vs. ontology. I could see
a preference for having the DSP be stand-alone so that the requirements
checking does not need to use both the DSP and the ontology. That could
become one of Stefanie's mega-requirements. Then there would be a
mega-requirement that the DSP includes definitions of values -- from
datatypes to value lists. Then again, if some application wishes to also
validate against the ontology, there's no one to stop them from doing that.
kc
On 12/11/14 11:17 AM, Bosch, Thomas wrote:
> I meant your 2. statement.
>
> examples for datatype date:
> ---------------------------------
> dbo:releaseDate > (after) dbo:latestReleaseDate
> birth date < death date
>
> examples for integers:
> --------------------------
> 1 < 2
> -1 < 5
>
> examples for strings:
> ------------------------
> Antoine < Kai
> '-1' > '5' ?
>
>
> But I also think that your 1. statement should also be a requirement.
> We do not have such a requirement?
> When I look over the requirements I do not see that one.
> We should add it.
>
>
> Cheers,
> Thomas
>
>
> Thomas, I'm not clear what you mean by "based on datatype". Maybe you
> can give an example.
>
> Also, I see two possibilities, and we may not have covered both:
>
> 1) make sure that a value is valid for its datatype (e.g. that a date
> really is a date)
>
> 2) compare the values of two different properties to some end (e.g.
> birth date must be earlier than death date)
>
> We were reading this as the latter, which is what "compare" implies to
> me. If it is #1, instead, we could call it "check that value is valid
> for datatype"
>
> kc
>
> On 12/11/14 9:50 AM, Bosch, Thomas wrote:
>> Hi all,
>>
>> I am not convinced to rename R-43
>> (http://lelystad.informatik.uni-mannheim.de/rdf-validation/?q=node/50)
>> to 'Data Value'.
>>
>> When I initially created the requirement
>> R-43-COMPARISONS-BASED-ON-DATATYPE
>> my intention was that comparisons of strings, digits, dates, ... should
>> be possible.
>>
>> So, not only comparisons of digits or only comparisons of dates, ...
>>
>> The comparisons should be done based on the individual datatype.
>>
>> This means, that date values should be compared according to the rules
>> how to compare values / literals of the datatype 'date'.
>> And these rules differ for the individual datatypes.
>>
>> Because of this, I chose the name 'comparison based on datatype'.
>>
>> my recommendation is 'comparison of literals based on datatype'.
>>
>> Thoughts?
>>
>>
>> Best,
>> Thomas
>>
>> --
>>
>> Thomas Bosch, M.Sc. (TUM)
>>
>> PhD Student
>>
>> GESIS - Leibniz Institute for the Social Sciences
>>
>> Social Science Metadata Standards
>>
>> Visitors Address: B2,1, D-68159 Mannheim
>>
>> Postal Address: P.O.Box 12 21 55, D-68072 Mannheim
>>
>> Tel: + 49 (0) 621 / 1246-271
>>
>> Fax: + 49 (0) 621 / 1246-100
>>
>> Web: http://www.gesis.org
>>
>> Website: http://boschthomas.blogspot.com/
>> GitHub: _https://github.com/boschthomas/PhD_
>>
>
> --
> Karen Coyle
> [log in to unmask] http://kcoyle.net
> m: 1-510-435-8234
> skype: kcoylenet/+1-510-984-3600
>
--
Karen Coyle
[log in to unmask] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet/+1-510-984-3600
|