- From: Leyla Garcia <ljgarcia@ebi.ac.uk>
- Date: Tue, 17 Oct 2017 10:41:59 +0100
- To: BATISTA Dominique <Dominique.BATISTA@france-bioinformatique.fr>, "public-bioschemas@w3.org" <public-bioschemas@w3.org>
- Message-ID: <1ba76026-169d-27c8-c828-b465961d91f6@ebi.ac.uk>
Hi Dominique,
A type in Bioschemas is the same as a type in schema.org. The
BioChemEntity is flexible and extensible thanks to generic properties
such as additionalType, additionalProperty and contains/isContainedBy.
This generic approach was selected because new profiles can be easily
incorporated in Bioschemas without going into schema.org (which requires
some sort of process).
You can see a profile as a customization where additionalType is fixed
(or limited to a predefined set) and specific uses of additionalProperty
(like "transcribed-from-gene") and isContainedIn/contains (like
"in-organism-XYZ"). So, validators and tools will need access to the
profile specifications, including all these customization I just
mentioned. We have not yet fully defined how that customization should
be specified.
I am wondering how this customization will work for validators and
tools. So I was wondering if having the profile name in this new
property preferredLabel would help at all.
Regards,
On 16/10/2017 17:22, BATISTA Dominique wrote:
> Leyla,
> Adding a new property to select a profile is good for me. I wonder how
> it would look like though.
>
> {
> "@type" : "object_type",
> "PrefferedLabel" : "object_profile"
> ...
> }
>
> Would we then have one spec file per profile with common properties in
> the corresponding type spec file? I'm not really sure to have fully
> understand that whole profile thing. What's the difference between a
> profile and inheritance and why did you guys choose one over an other?
>
> Thanks :)
>
> Cheers,
>
> Dominique Batista
> Ingénieur Web IFB Core
> CNRS - UMS3601
> Tel: 01.69.82.46.92
> Site: http://www.france-bioinformatique.fr
> ------------------------------------------------------------------------
> *From:* Leyla Garcia [ljgarcia@ebi.ac.uk]
> *Sent:* Monday, October 16, 2017 6:13 PM
> *To:* public-bioschemas@w3.org; BATISTA Dominique
> *Subject:* Re: New property for BioChemEntity
>
> Hello,
>
> It is in addition to additionalType. additionalType links to ontology
> terms, preferredLabel would point to a plain text. Having just one
> preferred label in English (any other language would be optional)
> would make things easier for validation (I think but Dominique can
> tell better).
>
> Regards,
>
> On 16/10/2017 17:08, Stephen Anyango wrote:
>>
>> Hello Leyla,
>>
>> Is this in addition to additionalType which is also meant to be an
>> enumeration of ontological terms labelling the profile? An example
>> would suffice for clarity.
>>
>> kind regards,
>>
>> Stephen Anyango
>> PDBe
>> EMBL-EBI
>> On 16-Oct-17 4:40 PM, Leyla Garcia wrote:
>>> Hi all,
>>>
>>> I have discussed with Rafael the need of a way to refer to a
>>> particular profile from a type specification. I think adding a new
>>> property could work. Please see below.
>>>
>>> Type: BioChemEntity
>>> Property: PreferredLabel
>>> Expected type: Text
>>> schema.org description: Preferred label in plain text for this entity
>>> Bioschemas description: The preferred label corresponds to the
>>> Bioschemas profile. Please refer to the Bioschemas profile
>>> enumeration to find the available labels <some link will go here>.
>>> Although multiple languages are allowed, a label in English should
>>> always be provided. If the label you are looking for does not exist,
>>> please contact Bioschemas.org <some link will go here>.
>>> Bioschemas restrictions: Minimum, Many (at least one in English),
>>> Enumeration.
>>>
>>> Validators could use this label to "know" better what the minimum,
>>> recommended and so really are as they change from profile to profile.
>>>
>>> I do not think at this point that such property would be
>>> needed/useful for Record. If you think it is, then we could propose
>>> it for Thing rather than BioChemEntity. Please share your thoughts.
>>>
>>> Regards,
>>>
>>>
>>
>
Received on Tuesday, 17 October 2017 09:42:29 UTC