Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Question on qualities of spatial regions #83

Open
ajnelson-nist opened this issue Mar 8, 2024 · 24 comments
Open

Question on qualities of spatial regions #83

ajnelson-nist opened this issue Mar 8, 2024 · 24 comments
Labels
discussion a question or comment followed by discussion

Comments

@ajnelson-nist
Copy link

I see that properties that could be used to associate a quality with a continuant exclude spatial regions, e.g., as with the domain of bearer of.

One concept I would personally consider a quality is, for a 3-dimensional thing, its volume, and likewise with 2-d things' areas, 1-d things' lengths. It appears if I wanted to relate spatial regions to, say, a threshold quality-value, I would not be able to do so with a spatial region directly, because I can't say this:

ex:Length
  a owl:Class ;
  rdfs:subClassOf obo:BFO_0000019 .
_:y
  a ex:Length .
_:x
  a obo:BFO_0000026 ;
  rdfs:comment "Some one-dimensional spatial region"@en ;
  obo:BFO_0000196 _:y .

Is there a reference somewhere of how spatial-dimensional qualities should relate to spatial region and its subclasses?

Possibly related and/or an answer to my question: Is BFO 210 (occupies spatial region) the "projection" mechanism alluded to in the elucidation for spatial region?

Related issues: Looking across the BFO-2020 tracker and BFO tracker, only this comment appeared to be relevant to this question.
@alanruttenberg
Copy link
Contributor

This is something to take up with Barry. His rationale is that qualities are what changes but that spatial regions don't change.
I've argued that some other qualities don't change, like the charge of an electron, but there's not been a resolution.
The workaround is to use information entities about the spatial regions.
That said, IMO if you are using spatial regions in your representations you probably should be using something else. Spatial regions are needed to hold the theory together but generally sites or boundaries are better choices. What's your use case?

@wdduncan
Copy link

wdduncan commented May 3, 2024

Based on this, and other issues, I am wondering what would be an instance of a spatial region. It seems that any region of space (such as the Earth's northern hemisphere) could be defined as a site instead of a spatial region. Moreover, the assertion that a spatial region cannot bear qualities makes me wonder what spatial regions are intended to represent.

@wceusters
Copy link

An example of a spatial region would be the one that was spatially-occupied by Earth's northern hemisphere yesterday at noon.
I am not sue what you mean by a 'region of space', but in any case, Earth's northern hemisphere is not a BFO:spatial region, neither is it a BFO:site. That hemisphere is clearly not a BFO:immaterial entity, while a BFO:site is immaterial.

@wdduncan
Copy link

wdduncan commented May 3, 2024

An example of a spatial region would be the one that was spatially-occupied by Earth's northern hemisphere yesterday at noon.

Would this be determined by the where the Earth was yesterday at noon? If so, then would it have been a BFO:site?

Does the spatial region that was spatially-occupied by Earth's northern hemisphere have any qualities? A particular height, width, volume, etc?

@phismith
Copy link

phismith commented May 3, 2024 via email

@alanruttenberg
Copy link
Contributor

@phismith that spatial regions have shapes and sizes either means those things are not qualities, or that you are disagreeing with the text you authored regarding specifically depends on:

DOMAIN: specifically dependent continuant
RANGE: independent continuant that is not a spatial region; specifically dependent continuant

Please clarify

@wceusters
Copy link

No site: occupies-spatial-region is time indexed and has domain: independent-continuant but not spatial-region and range: spatial-region [lzw-1]. A site can spatially-occupy a spatial region, but nothing can specially occupy a site. Things can be located in sites.

No: spatial regions don't have qualities. :-) As Alan said, you can use ICEs to say something about spatial-regions, or you can define subtypes of spatial regions, f.i. spheric(ish)-spatial-region.

@wdduncan
Copy link

wdduncan commented May 3, 2024

nothing can specially occupy a site

The distinction between spatially-occupy and located in seems quite technical. This can be a source of confusion.

No: spatial regions don't have qualities. :-) As Alan said, you can use ICEs to say something about spatial-regions, or you can define subtypes of spatial regions, f.i. spheric(ish)-spatial-region.

I think this need more clarification. An ICE that is about the shape of a spatial region is still about a quality. Similarly, an ICE about the height of tree is about a quality.

@ajnelson-nist
Copy link
Author

I was more meaning to approach the restriction on association of qualities, rather than drilling into the meaning of spatial region vs. site vs. boundary. Though, the discussion on that comparison is helpful too.

Some of my use cases about locations involve location parthood / nesting / sub-locations. E.g., if my graph denotes something was in a room of a building, I should be able to confirm through some mereological path that that something was in the building. Some use cases can go up through the time zone region, or the country.

There is one "Location" case I've been curious how to handle with a BFO basis. As I recall (apologies if I misremembered), in 2011, Japan had an earthquake that moved some parts of the country significantly (some feet). I'd normally want to treat some spot in that region as a "location." (I'm not currently sure if this term in this case clashes with BFO terminology.) If someone took a geo-tagged picture in 2010 standing at, say, some street intersection, and again in 2012 at the same street intersection, the GPS latitude and longitude (as qualities) would measure differently.

Aside from, say, the camera: in what entity could the latitude and longitude quality-instances inhere?

Feedback is also welcome if I incorrectly assumed latitude and longitude should be qualities.

@alanruttenberg
Copy link
Contributor

latitude and longitude don't inhere in anything, they are information about something.
During an earthquake, some things that used to be designated by a lat/long at t before are designated by a different lat/long at t2.
One might say that sites move around in that their material hosts move relative to other material entities.
There are a variety of relational qualities like relative distance between things on earth that will change value(meaning type in BFO) after an earthquake.

@wceusters
Copy link

nothing can specially occupy a site

The distinction between spatially-occupy and located in seems quite technical. This can be a source of confusion.

Not if one reads the axioms!

No: spatial regions don't have qualities. :-) As Alan said, you can use ICEs to say something about spatial-regions, or you can define subtypes of spatial regions, f.i. spheric(ish)-spatial-region.

I think this need more clarification. An ICE that is about the shape of a spatial region is still about a quality. Similarly, an ICE about the height of tree is about a quality.

But the ICE would not be about the in the current version of BFO not existing 'spatial-region-shape', but about the spatial-region itself. Or as alternative, like I said, the subtyping of spatial-regions. You cannot use s-depends-on to describe them, but you can make axioms to the effect that when a spherical material entity spatially-occupies a spatial-region x, x is an instance of spherical-spatial-region

@hoganwr
Copy link

hoganwr commented May 3, 2024 via email

@phismith
Copy link

phismith commented May 3, 2024 via email

@hoganwr
Copy link

hoganwr commented May 3, 2024 via email

@wdduncan
Copy link

wdduncan commented May 6, 2024

The solar system example is better. The hemisphere example wasn't as good. Sorry, I'm not not a skilled example crafter :(

True. It would be the shape of the Earth yesterday, and if it hasn’t changed at all, today.

So, the shape of the spatial region region that the Earth occupied would have qualities?

@alanruttenberg
Copy link
Contributor

@wdduncan hasn't this been asked and answered? Spatial regions can not have qualities.

@wdduncan
Copy link

wdduncan commented May 6, 2024

No.
Based on based on the previous response regarding the spatial region previous occupied by an entity, the mandate against qualities does not make sense. It is highly counter intuitive.

@alanruttenberg
Copy link
Contributor

Here's Werner answer and here I quote the relevant definitions

It may be an answer that's intuitive but it is nonetheless the answer.

@wdduncan
Copy link

wdduncan commented May 7, 2024 •
edited
Loading

Yes. I can read. I've made my reasons clear, and you haven't addressed the example I raised. Rather, you are insisting that very odd things (e.g., an IC that can't bear SDCs) are perfectly reasonable.

There is a lot I agree with in BFO. But, this is not one that I can agree with.

@wceusters
Copy link

@wdduncan: I addressed your example. Alan pointed out where. It is not because one can say things in a certain way, e.g. that the spatial region yesterday at noon occupied by Earth 'has the shape' of the Earth, that reality aligns perfectly with our phrases. That we can use words in this way, doesn't mean reality is that way. I assume you don't have problems in denying that the 'hotness' of a research topic is measured in degrees Celsius, or do you? So if you don't, there is no difference for accepting that talk about the shape of spatial regions does not mean spatial regions have shape.

@alanruttenberg
Copy link
Contributor

I am not insisting at a personal level. I am reading the specification. It's a formal language and the definitions answer the question, whether you like it or not. Unless I hear some new information I will close this as WONTFIX

@wdduncan
Copy link

wdduncan commented May 7, 2024 •
edited
Loading

whether you like it or not. Unless I hear some new information I will close this as WONTFIX

I've read the specifications. Using SHOUTY caps doesn't persuade me that this makes sense. But, this is the usual pattern discourse for BFO. Objections get ignored or shouted down. Close if you wish. I will stand firm in my objection and others can make their own decisions.

It is not because one can say things in a certain way, e.g. that the spatial region yesterday at noon occupied by Earth 'has the shape' of the Earth, that reality aligns perfectly with our phrases. That we can use words in this way, doesn't mean reality is that way.

@wceusters You have not addressed the issue at hand. Yes, sometimes we use analogies (e.g., "hotness of topic"), and sometimes we refer to things in reality (e.g., "the mass of a rock is 2 kg").

You have not provided clear criteria except by proclamation/stipulation that talking about spatial regions is merely a "way of talking" as opposed to talking about things in reality.

@ajnelson-nist
Copy link
Author

Re: @alanruttenberg

latitude and longitude don't inhere in anything, they are information about something. During an earthquake, some things that used to be designated by a lat/long at t before are designated by a different lat/long at t2. One might say that sites move around in that their material hosts move relative to other material entities. There are a variety of relational qualities like relative distance between things on earth that will change value(meaning type in BFO) after an earthquake.

I realize my example accidentally sidesteps one of my motivations for using qualities, and seems to have veered this conversation into meaning of spatial regions. But, that first line in your lat/long reply gets to something else I'm interested in with qualities: That the determined value could be in need of some process review, because different processes can find different values.

I wanted the Japan earthquake example because that makes a (dramatic) demonstration of the influence of time on latitude and longitude determination. But, I realize talking about the square of sidewalk veers the discussion into sites and fiat boundaries, and on those classes qualities are already fine.

But what about precision? Choice of instrument? Measurement strategy? Taking the example on BFO 28, "the region of space occupied by all and only the planets in the solar system at some point in time," I might be interested in the volume of that region (e.g., for a calculation of planetary vs. non-planetary density of the solar system). (Please forgive simplifications, I am not an astronomer, I'm just grabbing onto what the BFO-2020 encoding has.)

I can imagine several ways to determine that number in cubic meters, that could disagree on, say, atmospheric boundaries. Comparison of those found numbers, as quality-values, would entail contrasting the various measurement mechanisms seeking the true value. The volume of this region is "information about something," yes, but in order to compare various determinations about that "information about something" within one graph, I would want to use a quality.

I appreciate that, along the way, there is proxying the quality for the 3D spatial region (BFO 28) by qualities of either the planets as sites (BFO 6), and/or their sphere-ish boundaries as fiat surfaces (BFO 146). But I don't think this proxying and amalgamating disqualifies the 3D spatial region example in BFO 28 from being able to have a quality.

@wceusters
Copy link

@ajnelson-nist : All planets are material entities, most likely objects, thus not sites since sites are immaterial entities. A measurement procedure, when successful, results in the coming into existence of a GDC which is concretized as one or more qualities in some representational artifacts. But these qualities are not qualities of the measured things, they are ABOUT those things. There is no problem to put these 'representational' qualities in a graph as a graph itself comes with such q quality. But you can't put the volume of Earth in a graph. When you use different measurement procedures to measure the same entity, say the volume of Earth, then you obtain distinct 'representational' qualities ABOUT that one quality which is the volume of Earth. Now a representational quality about the volume of Earth, is about both the volume of Earth and Earth. You can use that representational quality to describe the spatial region that at that time was occupied by Earth, but if you wish to use BFO, not the volume of that spatial region because for BFO there are no qualities that INHERE in spatial regions. But BFO does not object against qualities that are ABOUT spatial regions.

@alanruttenberg alanruttenberg added the discussion a question or comment followed by discussion label Sep 26, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
discussion a question or comment followed by discussion
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
6 participants
@alanruttenberg @hoganwr @wdduncan @phismith @ajnelson-nist @wceusters