Can a "site" be part of a material entity?

[Hyunmin Cheong]

In BFO, I understand a site can be in a containment relation with a material entity (i.e., a cave contains a bear), but I was wondering if a site could be in a "part of" relation with a material entity. For example, is a nasal passage part of my body?

Motivating scenario: During engineering design, it may be necessary to create an arbitrary volume of space to indicate required clearance for a part being designed. For instance, when designing a side panel of a car, you could define a volume where wheels will be placed, and restrict your design from interfering with that space. My intuition would be classify such clearance volume as a "hole" (hence, a type of bfo:site), and make this hole part of the side panel. Does this seem reasonable?

[Barry Smith]

See the treatment of anatomical spaces in the Foundational Model of Anatomy

On Fri, Nov 3, 2017 at 5:47 PM, Hyunmin Cheong <hyunmin...@gmail.com> wrote:

In BFO, I understand a site can be in a containment relation with a material entity (i.e., a cave contains a bear), but I was wondering if a site could be in a "part of" relation with a material entity. For example, is a nasal passage part of my body?

Motivating scenario: During engineering design, it may be necessary to create an arbitrary volume of space to indicate required clearance for a part being designed. For instance, when designing a side panel of a car, you could define a volume where wheels will be placed, and restrict your design from interfering with that space. My intuition would be classify such clearance volume as a "hole" (hence, a type of bfo:site), and make this hole part of the side panel. Does this seem reasonable?

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss+unsubscribe@googlegroups.com.
To post to this group, send email to bfo-d...@googlegroups.com.
Visit this group at https://groups.google.com/group/bfo-discuss.
For more options, visit https://groups.google.com/d/optout.

[Melissa Haendel]

see also treatment in Uberon 

http://uberon.github.io/

lots of examples there

On Nov 4, 2017, at 10:39 AM, Barry Smith <phis...@buffalo.edu> wrote:

See the treatment of anatomical spaces in the Foundational Model of Anatomy

<image.png>

To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

To post to this group, send email to bfo-d...@googlegroups.com.
Visit this group at https://groups.google.com/group/bfo-discuss.
For more options, visit https://groups.google.com/d/optout.

Melissa Haendel, PhD
Associate Professor
Library & Dept. of Medical Informatics and Clinical Epidemiology 
hae...@ohsu.edu

503-407-5970

www.monarchinitiative.org

Appointments: Shanez De Silva 

des...@ohsu.edu

[Pierre Grenon]

According to BFO-1 -- sites are material entities of sort, aggregates
of a material part and an 'immaterial' part. Substances, however, are
either in a \contained in\ or in a \hosting\ relation with respect to
some sites, not a mereological relation. Also, a story about
delineation and boundaries. The nasal passage is hosted by your body.
It contains stuff that's not part of you.

According to BFO 2 -- sites are the immaterial parts in the previous
ones. While the doc on site doesn't really settle it, the ref doc on
material entity is explicitly saying that the FMA is followed in
taking sites to be parts of material entities.

While the difference between the two BFOs could be terminological
(both allow the same constructs), there is a potential issue. The
parthood take may perhaps apply cogently to internal sites (holes) but
perhaps only (or mostly) to anatomical ones --bodies are not purely
material (which is allowed by BFO2). Yet I'm not convinced the
treatment of external sites is tidy as a result. One way out might be
to say some sites can be parts of material entities (what's meant is
spatial parthood) and some sites can be hosted by material entities.

In the case of the car design, it really is not clear in the first
place what the material entity is. When you draw your car in your CAD
tool, you might actually be dealing with spatial regions only. If
that's the case, you have a lot of freedom and it might be easier to
use regions than sites.

The car when in assembly will be at some point a frame with a site for
the weels and at a later point have the weels as parts. There will
remain a site that will be only the difference between these
configurations. There is a material entity in the factory (and a
number of sites), your CAD tool describes spatial regions, ones where
material entities will be located, ones that will remain sites, one
that would be sites along the assembly process...

You can chose to hack this and say your design points at a material
entity (rather than a region that will be occupied by one at
production) and then you could still say that the other parts are
sites. Your clearance volume is still part of the overall volume. BFO
2 allows you to use parthood even if you don't follow this analysis.

Best,
Pierre

On Fri, Nov 3, 2017 at 9:47 PM, Hyunmin Cheong <hyunmin...@gmail.com> wrote:
> In BFO, I understand a site can be in a containment relation with a material
> entity (i.e., a cave contains a bear), but I was wondering if a site could
> be in a "part of" relation with a material entity. For example, is a nasal
> passage part of my body?
>
> Motivating scenario: During engineering design, it may be necessary to
> create an arbitrary volume of space to indicate required clearance for a
> part being designed. For instance, when designing a side panel of a car, you
> could define a volume where wheels will be placed, and restrict your design
> from interfering with that space. My intuition would be classify such
> clearance volume as a "hole" (hence, a type of bfo:site), and make this hole
> part of the side panel. Does this seem reasonable?
>
> --
> You received this message because you are subscribed to the Google Groups
> "BFO Discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an

> email to bfo-discuss...@googlegroups.com.

[Barry Smith]

On Tue, Nov 7, 2017 at 10:39 AM, Pierre Grenon <pierre...@gmail.com> wrote:

According to BFO-1 -- sites are material entities of sort, aggregates
of a material part and an 'immaterial' part. Substances, however, are
either in a \contained in\ or in a \hosting\ relation with respect to
some sites, not a mereological relation. Also, a story about
delineation and boundaries. The nasal passage is hosted by your body.
It contains stuff that's not part of you.

According to BFO 2 -- sites are the immaterial parts in the previous
ones. While the doc on site doesn't really settle it, the ref doc on
material entity is explicitly saying that the FMA is followed in
taking sites to be parts of material entities.

While the difference between the two BFOs could be terminological
(both allow the same constructs), there is a potential issue. The
parthood take may perhaps apply cogently to internal sites (holes) but
perhaps only (or mostly) to anatomical ones --bodies are not purely
material (which is allowed by BFO2). Yet I'm not convinced the
treatment of external sites is tidy as a result. One way out might be
to say some sites can be parts of material entities (what's meant is
spatial parthood) and some sites can be hosted by material entities.

​That is exactly my view, and I hope BFO is consistent with that view​

In the case of the car design, it really is not clear in the first
place what the material entity is. When you draw your car in your CAD
tool, you might actually be dealing with spatial regions only. If
that's the case, you have a lot of freedom and it might be easier to
use regions than sites.

​No. Because cars can move. And the holes inside them move with them. ​So those holes can't be regions. (They need not be parts of the car, of course.)

The car when in assembly will be at some point a frame with a site for
the weels and at a later point have the weels as parts. There will
remain a site that will be only the difference between these
configurations. There is a material entity in the factory (and a
number of sites), your CAD tool describes spatial regions, ones where
material entities will be located, ones that will remain sites, one
that would be sites along the assembly process...

​The CAD tool creates Information Content Entities pointing to some intentional analog of spatial regions.​

You can chose to hack this and say your design points at a material
entity

​

​the design is a funny kind of fictional ICE, about spatial and material things only in a very roundabout way​

​BS​

 

​

(rather than a region that will be occupied by one at
production) and then you could still say that the other parts are
sites. Your clearance volume is still part of the overall volume. BFO
2 allows you to use parthood even if you don't follow this analysis.

Best,
Pierre

[Alan Ruttenberg]

On Tue, Nov 7, 2017 at 11:20 AM, Barry Smith <phis...@buffalo.edu> wrote:

On Tue, Nov 7, 2017 at 10:39 AM, Pierre Grenon <pierre...@gmail.com> wrote:

According to BFO-1 -- sites are material entities of sort, aggregates
of a material part and an 'immaterial' part. Substances, however, are
either in a \contained in\ or in a \hosting\ relation with respect to
some sites, not a mereological relation. Also, a story about
delineation and boundaries. The nasal passage is hosted by your body.
It contains stuff that's not part of you.

According to BFO 2 -- sites are the immaterial parts in the previous
ones. While the doc on site doesn't really settle it, the ref doc on
material entity is explicitly saying that the FMA is followed in
taking sites to be parts of material entities.

This is something of a problem, in that we have agreed that terms should not change meaning.  BFO 1.1 definition

An independent continuant [snap:IndependentContinuant] consisting of a characteristic spatial shape in relation to some arrangement of other continuant [snap:Continuant] entities and of the medium which is enclosed in whole or in part by this characteristic spatial shape. Site [snap:Site] entities are entities that can be occupied by other continuant [snap:Continuant] entities

With comment:

An instance of Site [snap:Site] is a mixture of independent continuant [snap:IndependentContinuant] entities which act as surrounding environments for other independent continuant [snap:IndependentContinuant] entities, most importantly for instances of object [snap:Object]. A site [snap:Site] is typically made of object [snap:Object] or fiat object part [snap:FiatObjectPart] entities and a surrounding medium in which is found an object [snap:Object] occupying the site [snap:Site]. Independent continuant [snap:IndependentContinuant] entities may be associated with others (which, then, are site [snap:Site] entities) through a relation of "occupation". That relation is connected to, but distinct from, the relation of spatial location. Site [snap:Site] entities are not to be confused with spatial region [snap:SpatialRegion] entities. In BFO, site [snap:Site] allows for a so-called relational view of space which is different from the view corresponding to the class spatial region [snap:SpatialRegion] (see the comment on this class)

And, as is pointed out, is an immaterial entity in BFO 2.0

This doesn't look like a terminological difference, rather more like a substantive difference.

There should probably be a comment in the reference noting this difference. Ideally we would choose a different name for the current "site", and have a reason why bfo1:site is gone.

Alan

[Pierre Grenon]

I'm not sure whether axioms allow to decide what sort of move it has
been. I think bfo1:site may be 'gone' in terms of carving it out with
an identifier but it does not mean it is 'gone' from the material
branch in the sense of irreconciliable with bfo2's advertised and
named hierarchy (which is also noted to not be exhaustive).

will take it offline if you want but it'd be useful to look at this
and clarify, yeah

> --
> You received this message because you are subscribed to the Google Groups
> "BFO Discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an

> email to bfo-discuss...@googlegroups.com.

[Hyunmin Cheong]

Just to add my view on designs, particularly in CAD (computer-aided design) software:

In CAD, we deal with information entities that are digital representations of reality (either things that exist already or we plan to create). So the most of information entities in CAD have corresponding entities in reality. For instance, in typical CAD software, we start creating models in a given absolute "work space" or "work plane", in which digital models created can be uniquely located. They correspond to spatial regions (3D spatial region and 2D spatial region, respectively) in reality. Then, digital models created represent material entities - we have individual part models that correspond to objects, assemblies of part models that correspond to object aggregates, and some features of part models that correspond to fiat object parts. All these entities can have geometric entities (such as surfaces) that correspond to boundaries. Also, these entities can bear qualities (e.g., color) and relational qualities (e.g., contact between two surfaces). 

Essentially, I find that BFO's representation of reality is consistent with how things are organized in CAD software as well. I believe this is because (good) CAD software is also supposed to represent and deal with reality.

However, I did take one short cut when using BFO for my application - Strictly, all these representational entities that correspond to entities in reality are types of an information entity. However, I treated these representational entities as if they are entities in reality. For example, I would just assert that a "car assembly model" is an instance of "object aggregate" (rather than something like "digital representation of an object aggregate"). I'm hoping this is okay in my local application, and as long as we don't mix representational data and data from reality.

> email to bfo-discuss...@googlegroups.com.

> To post to this group, send email to bfo-d...@googlegroups.com.
> Visit this group at https://groups.google.com/group/bfo-discuss.
> For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

[Pierre Grenon]

How does it look like? Do you put references to ontological and BFO
concepts in text fields in your CAD drawings or do you use a
structured representation and the identifiers of elements to make
annotations?

I think this pretending stance (in the sense that we pretend that
these are entities) is fine in your context, if the goal is to attach
some sort of real world semantics. It might actually be necessary for
keeping it manageable if the goal is to generate new designs. Yet, it
is fine only because you are self relective about your use of the CAD
symbolism as a representation language.

At the risk of splitting hairs, you have to be careful though that a
design -- unrealised -- is not the same as a schematics of something
that exists, even though the schematics itself is something idealised
when it is normative and not descriptive.

For designs, we are in a make-believe universe in which it is in
practice easier than to say something insightful that takes into
account the nature of CAD objects (along Barry's and the lines you are
explaining too). We could do the same if we used BFO -derived
ontologies to represent Hamlet's plot. Implemeted knowledge bases
(Cyc, PLM) allow you to handle contexts in which these shortcuts are
made. Then you can say "this is a CAD context", "this is a fictional
context". It's on you to explain what that means and usually we don't
bother. It does not matter in practice and in applications until you
have to involve the nature of CAD representations themselves. So if
you want to use this stuff to annotate your designs or make a semantic
repo of them, you're good and standard in making that sort of
simplifications. When those shortcuts are not very well identifiable
(and you can't unpack them if you need), however, there is a general
sense of misrepresentation and this can lead to producing quite
useless annotations for more general purpose.

You would have to work through the ontology of CAD objects themselves
if you were to start from existing diagrams and tried to understand
them though (if you wanted to use ontology-based representation
through and through for doing this, that is).

Long story cut short, what you do is sensible practice. Make it more
contorted only if you need to.

[Alan Ruttenberg]

On Wed, Nov 8, 2017 at 7:25 AM, Pierre Grenon <pierre...@gmail.com> wrote:

I'm not sure whether axioms allow to decide what sort of move it has
been.

The logical axioms, at least from BFO 1.1 to BFO 2 do not. In BFO 1.1 OWL, site is a subclass of independent continuant, and in 2 it is subclass of immaterial entity. Equating  BFO 1.1 independent continuant to BFO 2 independent continuant, site -> site, material entity -> material entity, wouldn't yield an inconsistency. But it's common that the axioms don't say as much as the language. The text of the definition, and your earlier comments, seems to clearly indicate that in 1.1 site was a material entity. In 2 it's an immaterial entity. The reference says: "Immaterial entities are independent continuants which contain no material entities as parts". In the reference this isn't stated as an axiom, but in the BFO 2 Graz version there's an axiom to that effect. In the biodynamic paper there's this: "Niches are special sorts of sites [...] are typically made of a medium (of air, or water) enclosed by a mix of fiat and physical boundaries. Bottom line is that while there's no logical contradiction, it's clear the intent is different in the two versions. While it's not explicit that the word is an identifier pre-OWL, I think that would be the general expectation on reading the literature.

I think bfo1:site may be 'gone' in terms of carving it out with
an identifier but it does not mean it is 'gone' from the material
branch in the sense of irreconciliable with bfo2's advertised and
named hierarchy (which is also noted to not be exhaustive).

It looks like indeed, the older sense of site could be defined in the new BFO - it would be a subclass of material entity. 

That it *can* be defined is even more reason to choose a different name for what is currently called 'site' in BFO2.

 

Alan

> email to bfo-discuss+unsubscribe@googlegroups.com.

> To post to this group, send email to bfo-d...@googlegroups.com.
> Visit this group at https://groups.google.com/group/bfo-discuss.
> For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss+unsubscribe@googlegroups.com.

[Pierre Grenon]

I think I agree -- and believe I understand. :)

I was wondering about mereological and locational axioms and they may
not be in sight of the OWL implementation.

yes, 'site' was used to refer to something that is mereologically
constructed out of the variety of objects, object aggregates and so on
-- made of what's now called 'material entity' and not properly
speaking immaterial in the sense brought about with BFO2. Some sites
were not crisply bounded either, spatial vagueness is involved.

BFO2's 'site' could be named 'space', which works with FMA's
'anatomical space' (it's close to the common sense notion of space, as
in 'give me some space, mate, i wanna watch the game', yet there's the
risk of confusion with spatial regions maybe..) -- there's no real
immediate solution that will not bring some level of confusion and
raise terminological compatibility issues. I'm leery of using
'material site' and 'immaterial site' because there is a risk the
parallel could be stretched before we figure the proper analysis of
any parallel to draw.. Can an option be to keep 'site' in bfo2 and if
we ever need to name the former again use 'material site'? less
headache maybe

It needs to be figured out also how the parthood works between objects
and sites2 --but per Barry, it is cogent to approach this in terms of
spatial parhood and that there is an option for a 'hosting' rather
than parthood variant in certain cases..

It might be also useful to stress test the positive examples that were
carried over. If sites2 are a bit like spatial regions (but they
aren't because they move... as pointed out by Barry), then sites1 and
sites2 might also be linked in a manner of location.


On Thu, Nov 9, 2017 at 5:04 PM, Alan Ruttenberg

>> > email to bfo-discuss...@googlegroups.com.

>> > To post to this group, send email to bfo-d...@googlegroups.com.
>> > Visit this group at https://groups.google.com/group/bfo-discuss.
>> > For more options, visit https://groups.google.com/d/optout.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "BFO Discuss" group.
>> To unsubscribe from this group and stop receiving emails from it, send an

>> email to bfo-discuss...@googlegroups.com.

>> To post to this group, send email to bfo-d...@googlegroups.com.
>> Visit this group at https://groups.google.com/group/bfo-discuss.
>> For more options, visit https://groups.google.com/d/optout.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "BFO Discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an

> email to bfo-discuss...@googlegroups.com.

[Hyunmin Cheong]

On Wednesday, 8 November 2017 17:20:15 UTC-5, pierre...@gmail.com wrote:

How does it look like? Do you put references to ontological and BFO
concepts in text fields in your CAD drawings or do you use a
structured representation and the identifiers of elements to make
annotations? 

- It's work in progress, but I'd have a new data file that refers to CAD data entities (such as a particular surface definition in an OBJ file) and declare them as instances of an ontology class. This new file basically interprets CAD data entities through the view of the ontology.

 

 

I think this pretending stance (in the sense that we pretend that
these are entities) is fine in your context, if the goal is to attach
some sort of real world semantics. It might actually be necessary for
keeping it manageable if the goal is to generate new designs. Yet, it
is fine only because you are self relective about your use of the CAD
symbolism as a representation language. 

 

At the risk of splitting hairs, you have to be careful though that a
design -- unrealised -- is not the same as a schematics of something
that exists, even though the schematics itself is something idealised
when it is normative and not descriptive.

Yes, I think a design (well, in case of a CAD model) is an external representation of what a designer has in mind, which doesn't yet exist. It's interesting that you mentioned "to generate new designs". In fact, what we want is to enable computers to generate designs, given a particular problem. Now, the problem exists in reality and we need to represent this problem somehow in the CAD world; hence, my need to bring in real world semantics to CAD as you mentioned.

A specific example - you have a machine that is an assembly of parts in reality and you'd like the computer to re-design one of its parts. We represent this existing assembly of parts in CAD (so far, all the CAD entities would have real-world correspondence). This existing assembly also serves as a "problem", because the computer would need to design the new part that physically conforms to other parts. Then, we declare the part to be designed as "missing" or "non-existent" and the computer have to come up with new designs for this part, which would eventually become "real". Upon the creation of new designs, the computer would also have to perform physics-based simulation, which is also meant to represent a particular physical behavior (process) in reality.

 

For designs, we are in a make-believe universe in which it is in
practice easier than to say something insightful that takes into
account the nature of CAD objects (along Barry's and the lines you are
explaining too). We could do the same if we used BFO -derived
ontologies to represent Hamlet's plot. Implemeted knowledge bases
(Cyc, PLM) allow you to handle contexts in which these shortcuts are
made. Then you can say "this is a CAD context", "this is a fictional
context". It's on you to explain what that means and usually we don't
bother. It does not matter in practice and in applications until you
have to involve the nature of CAD representations themselves. So if
you want to use this stuff to annotate your designs or make a semantic
repo of them, you're good and standard in making that sort of
simplifications. When those shortcuts are not very well identifiable
(and you can't unpack them if you need), however, there is a general
sense of misrepresentation and this can lead to producing quite
useless annotations for more general purpose.

You would have to work through the ontology of CAD objects themselves
if you were to start from existing diagrams and tried to understand
them though (if you wanted to use ontology-based representation
through and through for doing this, that is).

Long story cut short, what you do is sensible practice. Make it more
contorted only if you need to.

I think I'd eventually need a "translator" that converts my "as-if-real" instantiations into more correct instantiations (e.g., X is_a object -> X is_a representational_artifact_of_an_object) if I wanted my data to interact with data from outside...