I think you can use either in other contexts. I see the tradeoff being that the inclusion syntax and templates for referring to properties of another item might be slightly longer. For the example of children of Charles Dickens you definitely want to have each child be their own item, as it currently is: http://www.wikidata.org/wiki/Q5686. The question for construction phases of Roman forts is whether or not each phase has enough information to justify being a complete item. Although if you are ready to go right now then it isn't really a question because qualifiers aren't implemented yet. I suggest creating the extra properties you need and creating separate items for the construction phases, even if they only have a few properties each, and then later on if downgrading them to just qualified values of properties for the main fort item simplifies some of your queries, inclusion syntax usage, and template boxes, then you can certainly do that.


Date: Wed, 27 Mar 2013 20:44:39 +0200
From: saturnian@gmx.com
To: wikidata-l@lists.wikimedia.org
Subject: Re: [Wikidata-l] A data model for Roman forts (castra)

Thanks Denny and Michael, it really helps. It is indeed a major difference between string enumerations of XSD and the values that should link to other items resulting in a knowledge network.

Choosing between lists of values and own item as value, I would prefer usually an own item because it could be used in other context, e.g. a sentence. A good example could be "The <children of Charles Dickens - Qxxxxxxxx> were younger than ...".



On 03/27/2013 04:53 PM, Michael Hale wrote:
Regarding the construction phases complex type that you want you have a couple of options. Properties support lists of values, so you could split the type into multiple properties and give a list of values for each. Then when qualifiers are added you could add a date range qualifier to each value to specify the phase or just use string qualifiers that say "phase 1", "phase 2", etc. depending on how detailed the information is. You can also group the properties in their own item. So you could create an item called "Potaissa phase 1 construction details" and then have construction phase just be a list of those specific items, which themselves contain the information.


Date: Wed, 27 Mar 2013 12:55:18 +0100
From: denny.vrandecic@wikimedia.de
To: wikidata-l@lists.wikimedia.org
Subject: Re: [Wikidata-l] A data model for Roman forts (castra)

I would say this is a good starting point. The Wikidata data model is described in full detail here [1], and a introductory primer is given here [2]. Qualifiers are not implemented yet, but will be there soon, and followed by more datatypes (like time, geo, etc.).

The major difference is that values like "stone" for material or "opus-quadratum" for technique should not be strings - this does not translate well. They should be pointing to items, e.g. Q8063 instead of "stone" and Q2631941 instead of "opus-quadratum".

The other thing is that Wikidata does not really intend to enable constraints in that very strong sense that your schema chooses. So if someone wants to add a value for material that you did not preconceive, like Q40861 (Marble), Wikidata-as-a-software will not stop them from doing so (just as Wikipedia-as-a-software does not stop you from entering that, either) (see also [3]).

I hope this helps,
Denny





2013/3/27 Flaviu <flaviu@gmx.com>
I fully agree with you that the XSD model cannot by precisely integrated into Wikidata and also I know Wikidata development is in progress. I think I could deal with simple properties like "material" but I'm not sure how to deal with complex properties like "construction phases" I'm not sure. Even if it is no implementation yet, how these complex properties could be defined?


On 03/26/2013 11:53 PM, Michael Hale wrote:
You can't integrate the XSD model precisely as it is defined because Wikidata doesn't allow all of the constraints that XSD allows. Specifically, you'll notice that you can't force an item to have a specific property (like the document or epigraphic reference in your model) and enumerations aren't currently supported. Wikidata has a global collection of properties and any item can use any arbitrary subset of them. The list is here: http://www.wikidata.org/wiki/Wikidata:List_of_properties. Some of the ones you want already exist, like material used for a building. http://www.wikidata.org/wiki/Property:P186 You'll want to propose the other ones you want here: http://www.wikidata.org/wiki/Wikidata:Property_proposal. I just added two statements (when you use a property, called a claim without references) to the Potaissa item that you use as an example. http://www.wikidata.org/wiki/Q7234612

> Date: Tue, 26 Mar 2013 22:50:52 +0200
> From: saturnian@gmx.com
> To: wikidata-l@lists.wikimedia.org
> Subject: [Wikidata-l] A data model for Roman forts (castra)
>
> Hi all,
>
> I recently attended "Wikidata meets archeology" symposium and I came
> across the question how a well structured and comprehensive data model
> of Roman forts (castra) could be integrated into Wikidata and then used
> on Wikipedia. For this I developed a raw XSD model based on the data
> model of Wikipedia's Infobox castrum. I put this model on the talk page
> of Q88205 (https://www.wikidata.org/wiki/Talk:Q88205) adding also a
> simple example. I would like to know this data model could be integrated
> into Wikidata? It is feasible? Is it too complex?
>
> Thank you,
> Saturnian
>
> _______________________________________________
> Wikidata-l mailing list
> Wikidata-l@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l


_______________________________________________
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l


_______________________________________________
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l




--
Project director Wikidata
Wikimedia Deutschland e.V. | Obentrautstr. 72 | 10963 Berlin
Tel. +49-30-219 158 26-0 | http://wikimedia.de

Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e.V. Eingetragen im Vereinsregister des Amtsgerichts Berlin-Charlottenburg unter der Nummer 23855 B. Als gemeinnützig anerkannt durch das Finanzamt für Körperschaften I Berlin, Steuernummer 27/681/51985.

_______________________________________________ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l


_______________________________________________
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l


_______________________________________________ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l