Re: Best practises for using multiple schemes to describe resources

RDF is property-centric, so I would say using the right properties is more
important that the right classes, because classes can be inferred from
properties.

There is also nothing wrong with using multiple types for a resource,
coming from different vocabularies. That does not constitute a conflict.

You could start with properties, and when patterns emerge of what your
domain model really is, you can create your own vocabulary with classes
that extend (possibly multiple) of those 3rd party classes, and define
additional properties for them that you couldn't find somewhere else.

On Thu, Feb 15, 2018 at 10:38 AM, Mikael Pesonen <mikael.pesonen@lingsoft.fi
> wrote:

>
> Thank you everyone for feedback.
>
> About the consumption: this is fully internal system, but we need to
> publish/rebuild data from it for other systems too.
>
>
> So there are several ways to handle this:
>
> 1) Easy / lazy way:
> Just pick properties that are suitable enough from various schemes without
> worrying about classes (ignoring class definitions and possible conflicts).
> Classes can be derived from the properties when needed. Validator tools may
> not work correctly with this data?
>
> 2) By the book a:
> As in 1), but taking class definitions into account and avoid conflicts
> (use a scheme/class only when it fits the resource type fully). This data
> should be fully functional with data validators. There might not exist
> fully compliant schemes for all purposes
>
> 3) By the book b:
> Define proprietary scheme from scratch and link it's properties to widely
> known scheme properties, such as Dublin Core. This is most work to do.
>
>
>
> For me, 2) seems the best way to go, especially if we can find the correct
> schemes. Maybe EDM, FRBR could work here.
>
>
> Mikael
>
>
>
> On 14.2.2018 18:01, Simon Spero wrote:
>
> On Feb 14, 2018 8:13 AM, "Mikael Pesonen" <mikael.pesonen@lingsoft.fi>
> wrote:
>
> We are describing resources by (almost randomly) picking suitable
> properties from various schemes, for example FOAF, Dublin Core, Nepomuk and
> Organization ontology.
>
> If a scheme defines class for the property, should the resource which is
> being described be always defined as such?
>
>
> If an ontology *defines* a domain for a property, then anything to which
> that property is applied is by *definition* a member of the specified
> class.  So the decision is made for you ☺️.
>
> The bigger question is whether you are applying multiple properties to the
> same individual in such a way as to require it to be an instance of
> incompatible classes. For example, a graphical  calendar app might create
> an individual that is a green Wednesday, but the source ontologies might
> require colored things to  be visible. (Blue Monday is a special case :)
>
> Proper document description is not a simple problem; you may be better off
> using a pre-baked solution as much as possible.
>
> The development of the Europeana Data Model may be  a good example to look
> at-
> See e.g. https://pro.europeana.eu/resources/standardization-
> tools/edm-documentation
>
> Simon
>
>
> --
> Lingsoft - 30 years of Leading Language Management
> www.lingsoft.fi
>
> Speech Applications - Language Management - Translation - Reader's and Writer's Tools - Text Tools - E-books and M-books
>
> Mikael Pesonen
> System Engineer
>
> e-mail: mikael.pesonen@lingsoft.fi
> Tel. +358 2 279 3300 <+358%202%202793300>
>
> Time zone: GMT+2
>
> Helsinki OfficeEteläranta 10 <https://maps.google.com/?q=Etel%C3%A4ranta+10&entry=gmail&source=g>
> FI-00130 Helsinki
> FINLAND
>
> Turku OfficeKauppiaskatu 5 A <https://maps.google.com/?q=Kauppiaskatu+5+A&entry=gmail&source=g>
> FI-20100 Turku
> FINLAND
>
>