Search Mailing List Archives
[protege-owl] Asserted Condition
b.x.jackson at world-net.co.nz
Wed Aug 6 18:30:40 PDT 2008
Thank you for your suggestions. Have reworked the material and feel I
have resolved most of the issues.
On Aug 6, 2008, at 8:30 AM, Kaarel Kaljurand wrote:
> On Mon, Aug 4, 2008 at 11:36 AM, Barry Jackson
> <b.x.jackson at world-net.co.nz> wrote:
>> I am in need of help with applying some Asserted Conditions. This
>> is part
>> of an e-government ontology.
>> I have divided completely a country into components using three
>> methods. I have called these classes in my ontology
>> Districts and Cities. (A city is a special type of district based on
>> population) I created two sub-classes under the class DandC.
> Some naming suggestions:
> * don't mix singular and plural (e.g. "Electorates" vs "Country"),
> use only
> singular for classes.
> * don't create "weird" names like "DandC", because (1) it's hard
> to understand the intended meaning of this class;
> (2) it seems to hide semantics into the name, rather user a complex
> class (district and city), or (district or city)
> * use lower-case for class names, i.e. "district" instead of
> "District", unless
> your class is called e.g. "EU-country" where uppercase is more
> (I know that the Protege manual would disagree with this point. :()
>> I have created a Class called Country and placed the above as sub-
> If "city" is a sub class of "country" then "every city is a
> country.". This
> is not correct, consider changing this to
> "Every city that is not Singapore and is not ... is not a country."
>> In each of the classes I have placed individuals.
>> In addition I have created a symmetric object property "includes"
>> Country as the Domain and Range.
> Why? I'm not sure I understand the modeling goal here.
> Can you express your intention in natural language?
>> Except for District and City I have no disjoint classes.
>> It is possible for the following to occur and it is these
>> assertions I wish
>> to construct:
>> A territory "includes" zero or more electorates and one or more
>> and/or city).
>> A district "includes" at least one or more territories and zero or
>> electorates but not a city
>> A city "includes" at least one or more territories and zero or more
>> electorates but not a district
>> An electorate "includes" one or more territories and one or more
>> and/or cities).
> I'm writing this is ACE because I don't want to torture the internet
> and the mail clients/servers
> with RDF/XML. ;)
> /* begin */
> No district is a city.
> If something X includes something Y then Y includes X.
> Everything that includes a country is a country.
> Everything that a country includes is a country.
> Every territory includes something that is a district or that is a
> Every district includes at least one territory.
> No district includes a city.
> Every city includes at least one territory.
> No city includes a district.
> Every electorate includes a territory and includes something that is a
> district or that is a city.
> /* end */
> To convert this into RDF/XML do the following:
> 1. Go to http://attempto.ifi.uzh.ch/ape/
> 2. Copy-paste the ACE text into the textarea
> 3. Check "OWL RDF" in the menu
> 4. Press "Analyse"
> 5. Copy-paste the OWL RDF/XML output into a file.
> (6. Open it with Protege (I tried with Protege 4 r11075))
>> Of course, no city can overlap with another city, no territory
>> with another
>> territory, no district with another district and no district with
>> a city.
> By "overlap" you mean "include"?
> In any case, I'm not sure how to express "no city can overlap with
> another city" in OWL,
> given that you allow a city to overlap with itself. (Do you?) If it's
> possible in OWL at all
> then you would need OWL 2 for that. I have to think about it.
> Hope this helps,
> protege-owl mailing list
> protege-owl at lists.stanford.edu
> Instructions for unsubscribing: http://protege.stanford.edu/doc/
More information about the protege-owl