Obsolete '' Animal manure'' and create ''animal dung''

[celineaubert]

Hi,
As we discussed on Skype, the ''Animal manure'' class may be obsoleted in ENVO and created in AgrO.
ENVO could create the class ''animal dung'' that will be imported in AgrO, and used as follows: "animal manure" is primarily composed of "animal dung".

[pbuttigieg]

Is there a difference between dung and feces? If not, we can subclass feces.

[pbuttigieg]

Also, is there any difference between manure and feces? If not, we can obsolete animal manure and move its subclasses under feces and add synonyms dung, manure, etc.

In this case, AgrO can just import it.

[celineaubert]

There is a slight difference. There are different types of animal manure : feces only, farmyard manure (feces+urine+vegetal like straw) or farm slurry (urine only).

[pbuttigieg]

Okay, got it.

So ENVO will keep the classes feces, urine, straw and AgrO should then define manure as

(feces OR (feces AND urine AND straw)) AND has_role some resource role

You could also split that up if you need to.
Sound good?

[pbuttigieg]

PS: we need resource role somewhere, but defining that is tricky (it implies usefulness to something), perhaps @cmungall has and idea on where that could go? We have waste role in ENVO, so perhaps we can create it there (to be moved, improved later)?

[cmungall]

we also have feces here:

UBERON:0001988 ! feces

But an instance of this is a portion from a single animal. animal manure is combined portions, possibly with other stuff mixed in.

as for

(feces OR (feces AND urine AND straw)) AND has_role some resource role

I am not a fan of disjunctive definitions, often a bad smell (in the coder's sense of bad smell, not the manure sense).

Also urine AND straw is unsatisfiable (or should be, with sufficient axiomatization). I think you mean

(has-part some X) and (has-part some Y) ...

My definition would be:

'animal manure' = 'enviromental material' and derives-from some feces and capable-of some 'soil fertilization'

I generally favor replacing roles with capable-of to the realizing process, but if we really want to name roles:

'animal manure' = 'enviromental material' and derives-from some feces and has-role some 'soil fertilization role'

maybe fertilizer is too specific, it can also be a soil amender. the crucial point here is that 'resource' is unsufficiently specific.

There is an argument for replacing derives-from with composed-primarily-of. It depends how we view the feces lifecycle. When does it stop being feces? I don't have strong opinions.

For the disjunctive clause, how about instead just having subclasses:

 animal manure
  slurry
  animal manure with plant material
    straw-based manure

lastly, regarding which ontology things go in. I want to caution against the situation where we have reciprocal dependencies. This can be a major headache.

[pbuttigieg]

we also have feces here:

UBERON:0001988 ! feces

But an instance of this is a portion from a single animal. animal manure is combined portions, possibly with other stuff mixed in.

Good point, didn't think about restrictions on instances. ENVO's excreta related to UBERON the same way as its water relates to CHEBI - import UBERON term, create ENVO term which is composed-primarily-of the UBERON entity, but allows for other stuff in there.

I generally favor replacing roles with capable-of to the realizing process, but if we really want to name roles.
Are these capable-of relations main streamed? We're at a stage where we don't have too many ENVO roles so we could switch, but we're working with CHEBI roles and some BFO roles and I'm wary of diverging too strongly. I really like the direct connection to the process, however.

There is an argument for replacing derives-from with composed-primarily-of. It depends how we view the feces lifecycle. When does it stop being feces? I don't have strong opinions.

I think this will have to reviewed by the agronomy crowd, we should capture this in a comment.

For the disjunctive clause, how about instead just having subclasses ...

Fine with that. @celineaubert we'll keep this in ENVO for now while these semantic issues are worked on. AgrO can import them and not worry about this for now. We'll surrender them over when they are stable. This also avoids the reciprocal dependencies at too early a stage.

[mark-jensen]

we need resource role somewhere, but defining that is tricky (it implies usefulness to something)

We have it in SDGIO, along with the realization process.

resource role = a role which is realized by conferring some benefit to its consumer.

I think it's ok to leave the nature of "benefit" open. But 'consumer' may not be the best choice, since it implies the actual use of the resource, whereas a body of water, stand of trees, money in the bank can all be resources. Putting it aside for future use/consumption makes it a resource, so consumption is not necessary.

[Public-Health-Bioinformatics]

@cmungall about "I generally favor replacing roles with capable-of to the realizing process,". Is the primary idea that we can claim that an entity be capable-of realizing some objective (or engaging in some process), e.g. 'doctor: person and capable-of some "provision-of medical diagnosis and advice" '.

Then one could say an entity "has role" some [role name] during the time it is realizing the aforementioned objective or process participation? "Though I have many hats, I'm wearing my doctor hat when I say to you bla bla ...". (I might be messing with BFO though in reducing semantics of "has role" from its previous scope of a) being the way one defined what an entity was capable of; and b) being attached to an entity even when it was not being realized.)

Hmm. Underlying many definitions is a seeping "intentionality", a kind of future contract. Our use of crap vs manure is about a possible or intended future of the entity, not anything an instrument can distinguish in the now. So I'm seeing an ontology pattern in which a class definition is stripped of intentionality, and where its subclasses DO enable intentionality differentia e.g.

feces    (Uberon "excreta and output_of some defecation"; would be ok to add "waste" since it is relative to the organism, the "upstream" semantics.  But how to logically imply only 1 organism is source?)
feces mix    (rename ENVO's "feces"? Allow for 1+ source organisms or 1+ defecations)
    animal manure: 'environmental material' and derives-from some "faeces mix" and capable-of some 'soil fertilization'
    fecal transplant material: 'transplant material' and derives-from some "faeces mix" and capable-of some 'gastrointestinal disease eradication'

the capability-of part is the intentionality part. In a conversational dialog, use of "animal manure" cues the intention of its use, even though many other logical capabilities (e.g. fecal transplant material) are inherent in the entity.

So to set the stage for "resource role"

resource: entity and capable-of some "positive input to some system"
liability: entity and capable-of some "negative input to some system"

then maybe
resource role: role which an entity partakes in while it causes some "positive input to a system"
(positive = beneficial)
I'm appreciating the precision y'all are bringing to this, and hoping to understand every last detail.