relation for process and time period

[GoogleCodeExporter]

I have many use cases that need to link a process with time or date. For 
example, 
- blood specimen collected at 4:30 pm on May 1st, 2011
- cells were infected with T. curzi for 2 hours

In BFO, there is a relation 'exists at' with domain 'entity' and range 
'temporal region' (instance or interval). 

Is 4:30 pm on May 1st, 2011 an instance of 'temporal region' and 2 hours a 
temporal interval? Can BFO have a specific relation like 'has duration' (a 
subproperty of 'exists at') for a temporal interval?

For above two examples, is it correct to represent "blood specimen collected at 
4:30 pm on May 1st, 2011" as:
'specimen collection' exists at '4:30 pm on May 1st, 2011' and has_output 
'blood specimen'
and "cells were infected with T. cruzi for 2 hours" as:
'infection process' 'has duration' 2 hours and has_input T. cruzi  

Another question, what different of '4:30 pm on May 1st, 2011', '2 hours' as a 
'temporal region' form those as 'time measurement datum' in IAO and how shall 
we use them?  

Original issue reported on code.google.com by zhengj2...@gmail.com on 7 May 2013 at 4:25

[GoogleCodeExporter]

The hierarchy is 

Region
  Instant (zero dimensional)
  interval (one dimensional)

So both are regions. 

The "for two hours" refers to an interval, but only gives information about the 
difference between start and stop. 

the "at 4:30pm" sounds like a instant, but isn't, since collecting takes an 
extended period of time. It might be the time it started, or some time near the 
beginning, or when it was scheduled. So you would want to choose whether you 
are indentifying a point in time like a start time, or the time of the process, 
which is an interval.

has_duration wouldn't be a subproperty of exists_at. (baking has_duration 1 
hour does not imply baking exists at 1 hour.
It would be a property of a the interval, or could be defined as a shortcut for 
any occurrent that has a projection to a temporal region.

However, there is a bit of a problem using exists_at for particulars, in the 
sense that if you want to use OWL data types, you can't use them directly, due 
to exists_at being an object property. As a practical matter we would probably 
add object or annotation properties on (object) instances of temporal region. 

Original comment by alanruttenberg@gmail.com on 7 May 2013 at 5:06

[GoogleCodeExporter]

Stefan will review this, and Jie will hound him until he does.

Original comment by alanruttenberg@gmail.com on 5 Jun 2013 at 6:52