boundary of

[GoogleCodeExporter]

New relation pair requested:

has_boundary
boundary_of

The relationship between a material anatomical entity and the (2D) boundary 
that delimits it.

e.g. the relationship between me and the surface of my skin.

Given that 'part of' can apply between anatomical structures and anatomical 
spaces, it seems sensible for 'boundary of' to be a SubPropertyOf 'part of'.  
But what are the potential consequences of this for reasoning on overlap?

Original issue reported on code.google.com by dosu...@gmail.com on 3 Mar 2013 at 6:09

[GoogleCodeExporter]

Have added as root Object Properties for now.   Also worth considering domain 
and range.  The way I've defined here suggests has_boundary would have domain 
'material entity' and range '2D immaterial entity'.  But what is the 
relationship between a space and the boundary of a material structures that 
defines its limits?

Original comment by dosu...@gmail.com on 3 Mar 2013 at 6:18

[GoogleCodeExporter]

I'd like to update these terms, in keeping with the other ROCore revisions.

RO_0002000 boundary of
- revise textual definition (just matching the style of the other revised 
definitions): a relation between a 2D immaterial entity (the boundary) and a 
material entity, in which the boundary delimits the material entity
- add alternative term: is boundary of
- add alternative term: boundary_of
- add example if usage: the surface of my skin is boundary of my body
RO_0002002 has boundary
- revise textual definition: a relation between a material entity and a 2D 
immaterial entity (the boundary), in which the boundary delimits the material 
entity
- add alternative term: has_boundary
- add example if usage: my body has boundary the surface of my skin

Unfortunately ROCore does not yet have a general relation between material 
entities and immaterial entities such as boundaries and cavities.

Original comment by ja...@overton.ca on 27 Jan 2015 at 2:49

• Added labels: ROCore

[GoogleCodeExporter]

> Unfortunately ROCore does not yet have a general relation between material 
entities and immaterial entities such as boundaries and cavities.

On my understanding of BFO(2), 'part of' applies in these cases  (see Barry's 
word doc).

Original comment by dosu...@gmail.com on 28 Jan 2015 at 11:22

[GoogleCodeExporter]

I agree that the BFO2 Reference allows for immaterial parts of material 
entities, via the "has continunant part" relation. We can use RO "has part" for 
that too. That's probably good enough for boundaries and cavities, and I think 
we should leave the question there for now.

On the other hand (and getting off topic) I can think of cases where "has part" 
is not sufficient to capture what we mean by "delimits", e.g. these fence posts 
are the material entities that delimit this 3D site (immaterial entity), but 
the site is not part of these fence posts. I'd prefer something more general, 
but let's not worry about it for now.

Original comment by ja...@overton.ca on 28 Jan 2015 at 1:56

[GoogleCodeExporter]

RO_0002002 has boundary
- move domain from ro-edit.owl to core.owl: 'material entity'
- move range from ro-edit.owl to core.owl: 'immaterial entity'

Original comment by ja...@overton.ca on 28 Jan 2015 at 2:34

[GoogleCodeExporter]

Barry made this comment on the mailing list:

Some 2D boundaries of 3D entities are parts of the things they bound, e.g. the 
fiat object boundary of the apple
Some 2D boundaries of 3D entities are not parts of the things they bound, e.g. 
the fiat object boundary of the environment of the apple at the surface where 
the apple meets its environment

Note that these two relations are not candidates for being called 'boundary 
of', 'has boundary', etc., since there are many other varieties of boundary of 
relations, some of them documented in the BFO 2.0 Specification.

Original comment by ja...@overton.ca on 28 Jan 2015 at 2:50

[GoogleCodeExporter]

Re Barry's comments: We could add an explanatory note to the effect that the 
range of this relation is limited to  2D boundaries that are part of the the 
object they bound.

Original comment by dosu...@gmail.com on 28 Jan 2015 at 5:00

[GoogleCodeExporter]

I agree with Barry that there are other sorts of boundary relations, so our 
label 'has boundary' is too general. I don't want to deal with the full 
generality, but we need to do something with these terms.

The original request is focused on anatomical boundaries, but the definition 
and domain say 'material entity'. If we change the domain to 'material 
anatomical entity' we could break things.

Although the range is 'immaterial entity', the textual definition is clear that 
they are 2D immaterial entities. So I suggest:

- we relabel these to 'has surface' and 'surface of' (or something); maybe we 
can leave 'has boundary' and 'boundary of' as alternative terms
- we tighten the textual definition to say that the surface is also part of the 
material entity

Original comment by ja...@overton.ca on 2 Feb 2015 at 2:55

[GoogleCodeExporter]

To me, surface has the connotation of a boundary between material and 
immaterial entities - or  (depending on how you abstract things) between solid 
entities and a gas or liquid.  Examples include the surface of my skin or the 
inner surface of my gut or my aorta.  Morphological boundaries in anatomy are 
much broader than this - including boundaries that run through a connected 
structure.  e.g. the boundary between dermis and epidermis:

http://en.wikipedia.org/wiki/Dermis#mediaviewer/File:Normal_Epidermis_and_Dermis
_with_Intradermal_Nevus_10x.JPG

Note - I am talking about the connnotations of terms here - not strict 
definitions of ontology relations.  In particular, I'm using the term between 
in the way a biologist would use it, not to specify the range of the relation, 
which I agree should be 2D immaterial entity.

I can see that having a general relation has_boundary only applying to 2D could 
be too restrictive.  'has 2D boundary' would be in the spirit of this relation.

One other issue:  I'm assuming that 2D here applies to manifolds.  The surface 
of my skin or the boundary between dermis and epidermis have 3D extent, even if 
we abstracly treat them as having no thickness.

Original comment by dosu...@gmail.com on 2 Feb 2015 at 4:23

[GoogleCodeExporter]

I'm ok with "has 2D boundary".

The textual definition currently in core.owl does not mention parthood. I think 
it would be safe to add a parthood requirement, but we do then it has to be in 
the textual definition. I would also be good to include "part" in the labels, 
but now it's getting really ugly: "has 2D boundary as part"?

both terms:
- add editor note: Although the boundary is two-dimensional, it exists in 
three-dimensional space and thus has a 3D shape.

Original comment by ja...@overton.ca on 3 Feb 2015 at 2:40

[GoogleCodeExporter]

On the mailing list we've been discussing whether these 2D boundaries have to 
completely surround the bounded entity. Barry sent me this, making it clear 
that BFO allows for boundaries that have gaps. Since we've been trying to keep 
close to BFO, I'm now inclined to accept this position and add an editor note 
to make it clear:

both terms:
- add editor note: A 2D boundary can have holes and gaps, but it must be a 
single connected entity and not an aggregate of several disconnected parts.

From Barry:

​This issue (of gaps in boundaries) is addressed at length in 
<http://ontology.buffalo.edu/smith/articles/Material_Entities.pdf> e.g. in the 
following passage:

"A tomato is an object in this sense, not however the two halves of a tomato 
before separation. A human being, too, is an object. And as both of these 
examples show, the fact that the surface of an object must be self-connected 
does not imply that the surface does not contain holes—for example, pores, or 
your mouth—through which particles of matter can penetrate in one or other 
direction. An organ such as your heart or brain is an object in this sense, and 
so also is a fetus. Each of these entities is connected by physical conduits to 
its surrounding host organism. But these connections are relatively weak, and 
(as we know from experience) the object in question is able to survive its 
disconnection."

and also in the BFO 2.0 Specification Document, and in the forthcoming MIT 
Press book on BFO.

A boundary can be connected even if it has gaps. As your skin, of course, 
demonstrates.

Original comment by ja...@overton.ca on 3 Feb 2015 at 2:58

[GoogleCodeExporter]

Chris says he's OK with partial boundaries.

On second thought, I don't think that we should require the boundary to be a 
part of the material entity (for this relation). While this will usually be the 
case, it was not part of the previous textual definition, and so adding the 
restriction could break current usage.

Revised proposal, changed labels to be more specific:

RO_0002000 boundary of
- revise textual definition: a relation between a 2D immaterial entity (the 
boundary) and a material entity, in which the boundary delimits the material 
entity
- change label: 2D boundary of
- revise textual definition: a relation between a 2D immaterial entity (the 
boundary) and a material entity, in which the boundary delimits the material 
entity
- add alternative term: is 2D boundary of
- add alternative term: 2D_boundary_of
- add alternative term: boundary of
- add alternative term: is boundary of
- add example if usage: the surface of my skin is a 2D boundary of my body
- add editor note: A 2D boundary may have holes and gaps, but it must be a 
single connected entity and not an aggregate of several disconnected parts.
- add editor note: Although the boundary is two-dimensional, it exists in 
three-dimensional space and thus has a 3D shape.

RO_0002002 has boundary
- revise textual definition: a relation between a material entity and a 2D 
immaterial entity (the boundary), in which the boundary delimits the material 
entity
- change label: has 2D boundary
- add alternative term: has_2D_boundary
- add alternative term: has boundary
- move domain from ro-edit.owl to core.owl: 'material entity'
- move range from ro-edit.owl to core.owl: 'immaterial entity'
- add example if usage: my body has 2D boundary the surface of my skin
- add editor note: A 2D boundary may have holes and gaps, but it must be a 
single connected entity and not an aggregate of several disconnected parts.
- add editor note: Although the boundary is two-dimensional, it exists in 
three-dimensional space and thus has a 3D shape.

Original comment by ja...@overton.ca on 6 Feb 2015 at 3:18

[GoogleCodeExporter]

This issue was closed by revision r408.

Original comment by ja...@overton.ca on 10 Feb 2015 at 2:44

• Changed state: Fixed