Vote requested on two proposed changes to Linked Places format (LPF)

8 views
Skip to first unread message

Karl Grossner

unread,
Aug 15, 2022, 10:44:02 AM8/15/22
to Linked Places
Hello all,

At the last meeting of the Linked Places format (LPF) working group, it was agreed to proceed stepwise on revisions to LPF by voting on two that have been proposed. Please indicate yes or no (plus any comment you might want to add) for both of these in reply to this message. Following the 16 Aug zoom meeting (https://pitt.zoom.us/j/93496830115) we will circulate this poll more widely.
  1. adding "wd" as a namespace prefix for Wikidata in the LP format context like so:    "wd": "http://www.wikidata.org/entity/"
  2. adding an optional "certainty" attribute to any timespan or named period within any "when" element, wherever it is placed (i.e at feature level or within any name, type, geometry, or relation).
           For example
        "when": {
          "timespans": [
            { "start": { "in": "yyyy-mm" },
              "end": {
                  "earliest": "-yyyy",
                  "latest": "yyyy-mm-dd" },
              "certainty": "less-certain"
            }
          ],
          "periods": [
            { "name": "Anglo-Saxon Period, 449-1066",
              "uri": "http://n2t.net/ark:/99152/p06c6g3whtg",
              "certainty": "certain"
            }
          ],
          "label": "for a century during the Anglo-Saxon period",
          "duration": "P100Y"
        }


Thanks,
Karl

--
Karl Grossner, PhD
Technical Director, World Historical Gazetteer
ka...@kgeographer.org
@kgeographer



Shaw, Ryan

unread,
Aug 15, 2022, 8:45:30 PM8/15/22
to Karl Grossner, Linked Places

> • adding "wd" as a namespace prefix for Wikidata in the LP format context like so:
> "wd": "http://www.wikidata.org/entity/"

👍

> • adding an optional "certainty" attribute to any timespan or named period within any "when" element, wherever it is placed (i.e at feature level or within any name, type, geometry, or relation).

Does this correspond to `lpo:has_certainty`?

https://github.com/LinkedPasts/linked-pasts-ontology/blob/main/lpo_latest.ttl#L121-L123

The documentation there says

> Preferred values 'certain' and ‘uncertain'

What are the allowed values? What does it mean that ‘certain’ and ‘uncertain’ are preferred? Does that mean other values can be used but may not be interoperable? Are ‘certain’ and ‘uncertain’ guaranteed to be acceptable here (for software adhering to the LPF spec)?

Karl Grossner

unread,
Aug 16, 2022, 12:44:39 AM8/16/22
to Shaw, Ryan, Linked Places
Thanks for noting this, Ryan

Yes it does correspond to that draft lpo: ontology, a Linked *Pasts*
Ontology, thought at the time to be ultimately broader than for Linked
Places format

The rdfs:comment for that is incorrect. As stated, any string is okay,
making 'preferred value' meaningless.

Not sure how one would restrict allowed values for an xsd:string, or
whether that is even possible. Maybe needs owl:withRestrictions?

But in the LPF spec we can say that only values ‘certain’,
'less-certain', and ‘uncertain’ are allowed for software supporting LPF.
I should have included that.

The lpo: ontology is an early draft and does need revisiting!

kg

Graham Klyne

unread,
Aug 16, 2022, 2:07:27 AM8/16/22
to Karl Grossner, Shaw, Ryan, Linked Places
Hi Karl, all,

(I’ve been lurking and watching with interest, but as I’m now actually retired I haven’t previously got actively involved with your discussions… as a non-participant, I offer the following as a suggestion, but don’t feel I should have any “voting” rights in the matter.)

 If you wanted to restrict possible values for certainty, I’d suggest introducing a new class (@type) and predefined instance values for those values.  An advantage of such an splash would be that certainty values could have separately defined labels and descriptions for display purposes.


(This is Turtle rather than JSON, but the ideas should be easily transferable.)

#g


Sent from my iPhone

On 16 Aug 2022, at 05:44, Karl Grossner <karl...@gmail.com> wrote:

Thanks for noting this, Ryan
--
You received this message because you are subscribed to the Google Groups "Linked Places" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linked-place...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/linked-places/c8e35923-d34b-acb6-51a5-7808f3ae7d8c%40gmail.com.
Reply all
Reply to author
Forward
0 new messages