[seek-kr-sms] OBOE clarifications and questions

[Shawn Bowers]

>> 3) How to deal with multiple relations with integrity constraints?  For
>> example, a 'site' table, and a 'tree measurement' table that has a
>> foreign key into the site table.  Can we create annotations that refer
>> to attributes in both tables?
>> 
>
> I'm not 100% sure what you mean here.  I hope that we can do this.  Shawn 
> might have a better sense for this question.

Matt, we have typically been defining a semantic annotation as a mapping 
from relation (database) instances to ontology instances. These mappings 
have signatures of the form (where a is the annotation)

   a: R1 x R2 x ... x Rn -> O1 x O2 x ... x Om

such that R1 to Rn are relations (tables) and O1 to Om are ontology 
classes and properties.  For example, the annotation

  a: Site(x) & Tree(x, y) -> StudyArea(x) & TreeMeasure(y) & measuredIn(y,x)

asserts that if x is a value in the Site table, and x,y are values in the 
Tree table, then x is an instance of a study area concept, y is an 
instance of a tree measure concept, and there is a property 'measuredIn' 
from y to x.

OBOE is only concerned with providing a useful vocabulary for the 
right-hand side of these rules. Not for specifying the left-hand side, and 
not for specifying the annotation logic itself.

>> 4) context doesn't seem to be enough to handle experiments -- it
>> captures some information, such as spatial nesting of experimental
>> units, but it doesn't fully capture the dependency information in
>> tuples.  In particular, it seems to me that experimental manipulations
>> are different from spatial nesting. See example below. Can you clarify?

This is not true: We do *not* define "context" as spatial nesting. We 
currently leave the semantics of "context" open -- with the intention of 
elaborating various context properties as needed.  Like entities, 
characteristics, and units, context is another "extension point" in OBOE, 
i.e., one can define various domain-specific vocabularies for context.

We have, however, talked about placing some limitations on how measures 
related through context are interpreted.  In particular, if a value x is 
related to a value y s.t. x is "within the context of" y, then we assume 
that x values can only be interpreted w.r.t. their corresponding y values.

That is, a given x value (lets say 4) for a given y value (lets say 1) is 
assumed to be in a different context as the same x value for a different y 
value. This means, e.g., that a sub-plot id of 4 within a plot id of 1 is 
a different sub-plot than a sub-plot id of 4 but wihtin a plot id of 2. 
I don't think we have agreed yet, however, on this definition/restriction 
for context.

-shawn