Open GoogleCodeExporter opened 9 years ago
Do we need to make it a principle, or should it just be part of the IAO
metadata set (like for other annotation properties)?
Original comment by mcour...@gmail.com
on 16 Oct 2013 at 11:11
See also 2007 "Towards naming conventions for use in controlled vocabulary and
ontology engineering"
http://msi-ontology.sourceforge.net/namingconventions/ cited in OBO Foundry
Coordinated evolution...
http://en.wikipedia.org/wiki/OBO_Foundry
Survey-based naming conventions for use in OBO Foundry ontology development
http://www.biomedcentral.com/1471-2105/10/125 and table 1 "The initial set of
OBO Foundry naming conventions" section 1 "Be clear and unambiguous". "Be
univocous" http://www.biomedcentral.com/1471-2105/10/125/table/T1
Authors: Daniel Schober, Barry Smith, Suzanna E Lewis, Waclaw Kusnierczyk, Jane
Lomax, Chris Mungall, Chris F Taylor, Philippe Rocca-Serra and Susanna-Assunta
Sansone
- Note distinction between editor preferred label (follows these guidelines),
other label properties.
Responding to Melanie's question: The issue first is whether it a principle in
spirit, is there contention about the need. Second how it is reflected in our
documentation
Original comment by alanruttenberg@gmail.com
on 6 Mar 2014 at 4:01
At the Editorial WG meeting on Mar 6, we discussed whether or not Foundry
principle 12 already included the notion that there should be unique Foundry
labels. I think there was general agreement that there should be unique labels,
and that the current wording that
"Naming conventions will ease the application of computerized lexical analysis
and processing and will further:
reduce lexical variance
increase human term recognition velocity
increase precision in the interpreted meaning conveyed by name
increase precision and recall in term query results
help automatic string matching"
should be interpreted to mean that there should be Foundry unique labels.
However, the group was divided as to whether it should be part of principle 12
or a separate principle.
We discussed using SKOS properties for unique labels. Some members of the
community have suggested this. I looked into this and it looks like
skos:prefLabel would work because:
1. skos:prefLabels are instances of owl:AnnotationProperty,
2. skos:prefLabel is a sub-properties of rdfs:label,
3. A resource has no more than one value of skos:prefLabel per language tag.
[See http://www.w3.org/TR/skos-reference/#labels]
As long as the prefLabel is specified in English, it would have to be unique
for each ontology. There was some concern that prefLabel was a data property,
but it is an annotation property.
Original comment by rlwalls2...@gmail.com
on 12 Mar 2014 at 12:35
Original issue reported on code.google.com by
haen...@ohsu.edu
on 14 Oct 2013 at 6:18