Re: Historical events

Yes, schema.org types might be somehow different from ontology classes.
However, the question still remains as to what schema.org/Event is supposed
to be used for and, more importantly, not used for.


There are 32 properties defined for schema.org/Event.  At least 24 of them
are suitable only for meetings or performances.  So it is quite
understandable that producers and consumers of schema.org information might
think that schema.org/Event should only be used for these kinds of things.

The text description on the page http://schema.org/Event also seems to be
for meetings or performances with its examples, its use of "certain", and
its talk about ticketing.  So again it is quite understandable that
producers and consumers of schema.org information might think that
schema.org/Event should only be used for these kinds of things.


A better setup would be to have one or more different schema.org types for
performances and meetings where the properties for them could be defined.
Then schema.org/Event could be a general type encompassing these types as
well as types for historical events and other kinds of events.  It may be
that this cannot be done because it would be too much of a break with
tradition, in which case a better textual description for schema.org/Event
would certainly help, but a better tradition-preserving way to go might be
to give schema.org/Event a supertype that would encompass events that are
not meetings or performances.


peter


On 06/01/2018 08:06 AM, Richard Wallis wrote:
> Thanks Phil for the name check and pointers to presentations and posts.
>
> Chairing the Schema Bib Extend <https://www.w3.org/community/schemabibex/> &
> Schema Architypes <https://www.w3.org/community/architypes/> groups, and
> participating in several others; this discussion on the suitability, or not,
> of Schema types and properties for a specific domain, or use case, are very
> familiar.
>
> Some of this comes from a subtle difference between a domain specific, often
> constraining /ontology (classes, attributes)/, and a /vocabulary/ of terms
> (types and properties) to describe /things, /such as Schema.org.  As I say
> this difference is subtle and the subject of philosophical discussions
> beyond the scope of the particular concerns in this thread.
>
> One spinoff of this however is the practical difference between the
> definition of a /Class/ and the description of a [Schema.org] Type such as
> Event <http://schema.org/Event>.  The former tends to constrain usage, the
> latter provides guidance to potential usage.
>
> If the current description of Event contained a few extra words, I’m sure
> this discussion thread would have a slightly different focus:
>
>     An event happening at a certain time */past or future/*and
>     location */physical or virtual/*, such as a concert, lecture,
>     */meeting/*, festival, or historical event. Ticketing information may be
>     added via the offers property. Repeated events may be structured as
>     separate Event objects.
>
>
> With such a description, markup such as this would be acceptable:
>
>  1. {
>  2.   "@context":"http://schema.org",
>  3.   "@type":"Event",
>  4.   "location":{
>  5.     "@type":"Place",
>  6.     "name":"Runnymede",
>  7.     "sameAs":"https://en.wikipedia.org/wiki/Runnymede
>     <https://en.wikipedia.org/wiki/Runnymede>"
>  8.   },
>  9.   "name":"Magna Carta",
> 10.   "startDate":"1215-06-15",
>11.
>
> 12.   “sameAs":"https://en.wikipedia.org/wiki/Magna_Carta
>     <https://en.wikipedia.org/wiki/Magna_Carta>",
>13.
>
>14.
>
>
> 15. }
>
> Proposals to adjust descriptions of terms in Schema.org are not uncommon.
> This brings me to the question of proposing a new type of /HistoricalEvent/.
> My question being, what is special about a historical event - other than it
> being [potentially a long time] in the past.
> Also discussed here is how, with the current definition of ISO 8601, you
> define open ended and other less common time periods. I think the previously
> referenced movements around that area should solve those problems.
> Finally, I believe there is an issue with /Event/ and its subtypes where it
> is difficult to describe time periods too vague to usefully use the
> combination of startDate, endDate & duration. Periods like “The Dark Ages”,
> “The Bronze Age”, etc.  Maybe a new HistoricalPeriod type would be
> appropriate, not sure, maybe one or two extra properties would do it - I am
> open to suggestions.
> A final general comment, to those who are uncomfortable with the way
> properties are lumped into Schema types to cover many generic use cases, All
> properties are optional. If your Volcano does not have a fax number, there
> is no need to define it. If Ticketmaster did not offer tickets to the
> signing of the Magna Carta, there is no need to describe the offer. Using
> the /Event/Type in markup only infers that the entity you are describing has
> taken place, or will do, at a known, or unknown location. It is the
> selection of a subset of available properties that then starts to refine
> that meaning.
> ~Richard
>
>
> Richard Wallis
> Founder, Data Liberate
> http://dataliberate.com
> Linkedin: http://www.linkedin.com/in/richardwallis
> <http://www.linkedin.com/in/richardwallis>
> Twitter: @rjw
>
> On 1 June 2018 at 12:42, Phil Barker <phil.barker@pjjk.co.uk
> <mailto:phil.barker@pjjk.co.uk>> wrote:
>
>     Hello all, I've been following this discussion with interest, and can't
>     resist jumping in now.
>
>     Allison, I wouldn't read too much in to the semantics of a schema.org
>     <http://schema.org> Type based on the properties of that type.
>     Specifically, I wouldn't infer that schema.org/Event
>     <http://schema.org/Event> is a commercial event just because the Event
>     type has properties that are relevant only to commercial events.
>     Schema.org is generally 'lumping' in its approach to classifying,
>     lumping stuff together rather than creating fine distinctions; put this
>     together with the inheritance hierarchy of schema.org
>     <http://schema.org> Types and you get such oddities as fax numbers for
>     volcanoes.
>
>     My own 2p-worth is that the defining feature of a schema.org/Event
>     <http://schema.org/Event> is that has a time and a location (and that
>     the place can be abstract, like the Internet for webinars).
>
>     I think a big change like moving the commercial properties out of Event
>     into a subtype that sits alongside Historical Event would break too much
>     of the existing data. Consequently, I think a type for Historical Events
>     (i.e. events of historical significance, used as reference points in for
>     other events, not just any old past event) would probably sit in
>     schema.org <http://schema.org> as a subtype of Event, and inherit all
>     those 'commercial' properties from Event. FWIW, I think WW2 would be a
>     fine as an example of such an event.
>
>     Also, I think there may be a case for something like a Named Period
>     type. The Victorian Age, Paleocene Epoch would be examples. And yes,
>     "WW2 era" could be a named period (if post-war and pre-war are, you kind
>     of need something for the bit in between).
>
>     I am not sure whether you would need both Historical Event and Named
>     Period, but I think they might facilitate slightly different types of
>     statement (something and Historical Event X were contemporary; something
>     was a consequence of Historical Event Y, something happened during the
>     Named Period Z [but wasn't part of it in an event/sub-event way]).
>
>
>     As for extending schema.org <http://schema.org>, I've run a couple of
>     W3C community groups. I would suggest first make sure that schema.org
>     <http://schema.org> is the best vocabulary for this type of information,
>     e.g. by thinking about use cases that fall within the scope of its
>     mission. This thread could be the start of this. Then raise an issue on
>     the schema.org <http://schema.org> github issue tracker if you think
>     there is anything that is unaddressed. Setting up a W3C community group
>     is straightforward and worth doing if you have complex set of use cases
>     and a need to build some sort of community consensus around how to
>     address them. Then you just need to recruit participants and steer the
>     discussions around a coherent resolution of meeting your use cases--that
>     can be more of a challenge.
>
>     Richard Wallis has written & presented a lot about extending schema,
>     perhaps most usefully in this context would be his presentation about
>     extend schema.org <http://schema.org> for bibliographic use:
>
>       * Schema.org: Part 1 - Fit For a Bibliographic Purpose
>         https://www.youtube.com/watch?v=ZiijJj1v1bQ
>         <https://www.youtube.com/watch?v=ZiijJj1v1bQ>
>       * Schema.org: Part 2 - Extending Potential and Possibilities
>         https://www.youtube.com/watch?v=sKHw3UsD3r8
>         <https://www.youtube.com/watch?v=sKHw3UsD3r8>
>
>     (He has also written a 3-part series of blog posts on the technical side
>     https://dataliberate.com/2016/02/10/evolving-schema-org-in-practice-pt1-the-bits-and-pieces/
>     <https://dataliberate.com/2016/02/10/evolving-schema-org-in-practice-pt1-the-bits-and-pieces/>
>     )
>
>     Hope this helps, Phil
>
>     On 31/05/18 21:07, Muri, Allison wrote:
>>     I think, respectfully, Thad, that you might not be understanding what
>>     is needed. I would not really have a strong desire to use sameAs,
>>     myself. This isn’t about “too much work,” really. What I am trying to
>>     do, and I might be inferring correctly that Roger is, too, is to have a
>>     simple schema.org <http://schema.org> markup that addresses events
>>     other than commercial events. That would be either to have Event as a
>>     category under which commercial events and historical events are
>>     subcategories, or have Occurence or some such category.
>>
>>     I am curious about Peter Patel-Schneider’s concern “about
>>     stretching schema.org/Event <http://schema.org/Event> to cover … things
>>     like WWII whose location is certainly not ‘certain’, or even
>>     anything that is not the kind of event that has performers and
>>     attendees.” Why is that? 
>>
>>     I see, when I read about the Community Group and Steering Group
>>     (http://schema.org/docs/about.html#cgsg
>>     <http://schema.org/docs/about.html#cgsg>), that “other W3C Community
>>     Groups exist that are focussed partially or entirely on schema.org
>>     <http://schema.org> improvements, e.g. health and medicine
>>     <https://www.w3.org/community/schemed/>, sports
>>     <https://www.w3.org/community/sport-schema/>,archives
>>     <https://www.w3.org/community/architypes/>, libraries and bibliography
>>     <https://www.w3.org/community/schemabibex/>, autos
>>     <https://www.w3.org/community/gao/>…” 
>>
>>     I wonder if there is someone on this list who is knowledgeable about
>>     these other groups and could provide some information about what one
>>     might do if one wanted to improve on markup for, say, history and culture? 
>>
>>     - Allison
>>
>>
>>
>>     On May 31, 2018, at 1:33 PM, Thad Guidry <thadguidry@gmail.com>
>>     <mailto:thadguidry@gmail.com> wrote:
>>
>>>     Roger,
>>>
>>>     Depends on what you are trying to do.
>>>
>>>     I taking a stance that most of the time there's no need to tell a
>>>     machine that something is historical if you can reconcile the entity
>>>     before hand and provide data about that.
>>>
>>>     Most machines will know and understand (if they are decently built and
>>>     programmed) to know that "Battle of Gettysburg" is a historical event.
>>>     There's no need to tell most machines that...
>>>
>>>     HistoricalEvent: Battle of Gettysburg
>>>
>>>     if you could possibly reconcile your entities (your welcome to use my
>>>     community's latest OpenRefine against Wikidata for that)
>>>     and instead of providing Strings...provide Things...
>>>
>>>     HistoricalEvent: "Battle of Gettysburg
>>>     sameAs: https://www.wikidata.org/wiki/Q33132
>>>     <https://www.wikidata.org/wiki/Q33132>
>>>
>>>     Machines can easily find out lots of information by parsing Wikipedia
>>>     itself like Google does, if you provide a sameAs url, or perhaps even
>>>     better a Wikidata or DBPedia url.
>>>
>>>     Giving us a better example of what you are trying to do would be most
>>>     appreciated by all.
>>>
>>>     My hunch is that you don't so much care about "Battle of Gettysburg"
>>>     but relations around it ?  What are those relations that you trying to
>>>     establish ?  That it was partOf: American Civil War ?
>>>     What else ?
>>>
>>>     Help us and we can help you,
>>>     -Thad
>>>
>>
>>     ....................................................
>>     Allison Muri
>>     Department of English
>>
>>     Arts 418
>>     University of Saskatchewan
>>     Saskatoon, SK, Canada
>>     ph: 306.966.5503
>>
>
>     -- 
>
>     Phil Barker <http://people.pjjk.net/phil>. http://people.pjjk.net/phil
>     PJJK Limited <https://www.pjjk.co.uk>: technology to enhance learning;
>     information systems for education.
>     CETIS LLP <https://www.cetis.org.uk>: a cooperative consultancy for
>     innovation in education technology.
>
>     PJJK Limited is registered in Scotland as a private limited company,
>     number SC569282.
>     CETIS is a co-operative limited liability partnership, registered in
>     England number OC399090
>
>

Received on Friday, 1 June 2018 15:46:56 UTC