New class: Site Visit

[dr-shorthair]

Propose addition of a class for Site Visits.

A site visit is associated with

0. a project or initiative or investigation
1. a site (geographic location) which might be defined by fiat (e.g. transect, management area), accessibility (e.g. helipad), or as an environmental zone,
2. a time interval
3. an agent (person, team)
4. (optionally) visits to nearby sites associated with the main site
5. one or more observation- or sampling- or specimen collection- or instrument deployment- activities

The sub-visits and activities have a causal dependency on the site visit.

The site-visit serves as a convenience class for data, matching the management of the data collection process.

First specific issue following on from #82

[robgur]

A site visit does not need to generate a collection or positive observation or successful deployment, does it? The visit may yield an absence of all targeted "things", just to be clear. That could still be an observation (e.g. "I didn't see a focal taxon).

…

On Thu, Sep 20, 2018 at 1:31 AM Simon Cox ***@***.***> wrote: Propose addition of a class for Site Visits. A site visit is associated with 0. a project or initiative or investigation 1. a site (geographic location) which might be defined by fiat (e.g. transect, management area), accessibility (e.g. helipad), or as an environmental zone, 2. a time interval 3. an agent (person, team) 4. (optionally) visits to nearby sites associated with the main site 5. one or more observation- or sampling- or specimen collection- or instrument deployment- activities The sub-visits and activities have a causal dependency on the site visit. The site-visit serves as a convenience class for data, matching the management of the data collection process. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#85>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAcc7FmsqmWuMCMrcJE1K6f87G4nOsXnks5ucygegaJpZM4WxaGd> .

[ramonawalls]

@dr-shorthair I am trying to figure out what the specific differentia are that distinguish a site visit from the more general "planned process". If I understand correctly, the key differentia are:

Is part of some investigation
-Is this required or optional? See definition of investigation: http://purl.obolibrary.org/obo/BFO_0000015.

Occurs in some site

• Every planned process happens at some site, but in this case, the site needs to be defined/instantiated.

Has participant some person

• since a team is made of people, this covers teams
• This differentia rules out planned processes that are activated by some machine.
• We are working on specific sub-properties of has participant in RO to specify, for example, who/what initiates a process or is the agent, but for now, using has participant would get part way there.

Has part some sampling process or observation process

• or we could just use planned process as the part
• In this case, the planned process that is part has to be some process other than the site visit itself (so a non-reflexive has part).

[pbuttigieg]

On aligning the usage of site :

It seems that here we don't mean BFO:site (too general) but a BFO:'fiat object part' (like a plot in some ecosystem, with the fiat bit specified in some plan). The latter delimits and overlaps the former.

[dr-shorthair]

• I guess a site visit is usually associated with some investigation, but might be opportunistic

• I would say that a site visit involves a visit by some agent. Need not be a person.

I guess it could be argued that persistently deployed kit is a long-duration site visit? But the key requirement is to be able to characterize a short time-duration activity.

• I looked at the causal temporal relations in RO for some pointers here. Seemed to me that the sampling- and observation-activities that are associated with a site visit are causally related?

[pbuttigieg]

Probably best to keep it simple and have several processes we can chain together freely.

'planned site visit' = A planned process during which an agent moves towards and enters a site'

The axioms wouldn't say BFO:site but probably ENVO:'astronomical object part' (a fiat object part, can't check exact class right now).
Agent is linked to the 'has active participant' relation in RO.

The "planned" part of the label is to differentiate planned visits from accidental visits (we should have a class for that too, perhaps).

This can be upstream of or within observation or sampling processes. I'm not sure it's fair to say that the visit "causes " the downstream processes, I'd stick with just the temporal / ordering relations.

Persistent kit could indeed be seen as involved in a long site visit. We could say that long-term deployment processes are downstream of or within a site visit process.

[pbuttigieg]

PS: SDGIO has "access" semantics which may be useful here (ability, right, permission to approach and enter a site)

[robgur]

Coming from more of the biologists/field monitoring side of this rather than the formal specification side:

The "planned" part of the label is to differentiate planned visits from accidental visits (we should have a class for that too, perhaps).

One point of clarification: Simon, above, says "I guess a site visit is usually associated with some investigation, but might be opportunistic". I will just note that an opportunistic site visit is not the same thing as an "accidental" visit. You can and usually do plan to opportunistically sample somewhere, so it is still a "planned site visit". BCO already does have formal definitions of inventory search types, including "opportunistic". Here is that definition: http://purl.obolibrary.org/obo/BCO_0000053. I think the term I would use is "adventitious" (http://purl.obolibrary.org/obo/BCO_0000055) for what Simon means, above. And that is still part of a planned process, typically.

Persistent kit could indeed be seen as involved in a long site visit. We could say that long-term deployment processes are downstream of or within a site visit process.

For me it helps to think about this in a use case or exemplar context. We have had to model how to represent something like a network of camera traps, mostly in a relational framework and in collaboration with networks like TEAMS and WildlifeInsights who collect these data for NGOs such as WWF and CI. And it is quite complex, with projects, deployments, larger sites for the network, and then each camera defining a smaller site that is being monitored, along with bouts of machine recorded sightings. Not sure any of this helps! But interesting to read the conversation here, in relation to our efforts on this: https://onlinelibrary.wiley.com/doi/full/10.1111/ecog.02942

…

-r

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <#85 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAcc7Asl6n9z0Mb_tWDlMc9bxC2PoU2Oks5ufxLpgaJpZM4WxaGd> .

[dr-shorthair]

Thanks for the clarification @robgur
Indeed, adventitious or incidental is what I meant, not opportunistic ... which might still be planned.