search

erlangen-crm / ecrm

Erlangen CRM - An OWL implementation of the CIDOC Conceptual Reference Model
http://erlangen-crm.org
41 stars 13 forks source link

Added missing rdfs:domain and rdfs:range definitions #2

Open fischmat opened 7 years ago

fischmat commented 7 years ago

Hi all,

when working with the current release of the ECRM I recognized that some RDFS domain and range definitions were missing. I added those following the definition in CIDOC 6.2. Hope I can help to improve the project.

Bests,

Matthias

Knurg commented 7 years ago

Hey Matthias,

welcome to the team - your contribution is appreciated :)

I did not go through all of your changes as the diff for .owl-files is a mess in github. Did you consider that the current version of the ecrm with the missing domains and ranges is CIDOC CRM Version 6.2.2 and not Version 6.2? If you're looking for Version 6.2 you could use version 150929.

The missing domains and ranges in the most recent version of the ecrm are due to points that needed clarification on the CIDOC CRM Version 6.2.2. However the current state of the document looks promising.

Best

Mark

fischmat commented 7 years ago

Hi Mark,

thank you for your quick answer and the clarification. I'm new to CIDOC and the ECRM and want to use it in the context of a project I'm working on. There every property requires a RDFS domain/range definition. Would it be ok if I add the missing statements to version 150929 according to the definition in CIDOC 6.2 and create a new pull request? Thanks in advance.

Bests,

Matthias

VladimirAlexiev commented 7 years ago

Using dated ECRM URLs in your data is not a good idea since that lowers the chance of interoperation to nearly zero. By the same token, IMHO making additions and fixes only makes sense for ecrm_current.owl . The other files have only historic interest. Why is that file (updated 11m ago) older than ecrm_170309.owl (updated 2m ago)?

Version 6.2.2. the current state of the document looks promising

You think so? I think we shoudn't make RDF before CRM SIG makes RDF

That URL change was enough reason for Getty (http://linked.art) to stop using this class.

Knurg commented 7 years ago

Dear all,

thank you for the comments.

@fischmat Could you please check if your additions are also missing on the ecrm_150929.owl? We indeed deleted some domains and ranges because the CIDOC CRM 6.2.1 said to do so. So it might be that in 150929 they were still there. As you know what you've added you're the only one who could check :) If they are missing there too, we can add them there - it is the right place I think.

@VladimirAlexiev typically I would agree with you. However we try to implement the crm-versions as soon as possible to see possible pitfalls in the current development of the cidoc crm. So we could address them on the working group meetings and talk about solutions directly. Of course we could not publish these results as .owl-files here - but I think some others might have interest in them, too.

Currently there are some really rough developments (e.g. deletion of all subclasses of E41) which have major inpact - so using the "current" might make problems there.

Best

Mark

fischmat commented 7 years ago

Hi,

I checked ecrm_150929.owl and found that the following properties have missing definitions:

Without definition of rdfs:domain:

Without definition of rdfs:range:

Without definition of both rdfs:domain and rdfs:range:

Some of the definitions could be inferred by a reasoner through the owl:inverseOf relationship, but for e.g. P28 and P28i this is not the case. It would be great if we could complement these properties. Should I add the definitions according to CIDOC 6.2 definitions and create a new pull request?

Bests,

Matthias

VladimirAlexiev commented 7 years ago

https://github.com/american-art/aaa/issues/16#issuecomment-317450117 is the discussion whether to use E78 at all.