OBI device tree insufficient

[rduerr]

I am working on an ontology for all of NASA's data which spans the fields of Astrophysics, Planetary Science, Heliophysics, Biology, Physical Science and Earth Science. Obviously I need terms for instruments, observing systems, etc. that span all branches of science.

OBI currently has the device branch under processed material which is under material entity. However, that is not nearly general enough for all of the fields I am dealing with, including:

1. those where humans are part of the observing system (i.e., not a subject of the observation, but part of the system doing the observing and without whom there would be no observation - e.g., Apollo landers, space station, space shuttles, etc.), and
2. those where the observing system is observing multiple kinds of things with its many instruments where each instrument is also capable of producing data about many kinds of things (e.g., temperature and wind speed and vegetation index and 200 more kinds of measurements from one instrument as the MODIS instrument that flies on the TERRA and AQUA satellites does ) and
3. where the observing system may have parts that are attached at one point in a mission and not attached in others (e.g., CASSINI and the Huygens probe it dropped on Titan or a Mars lander, its rover and its helicopter).

I note that the ENVO ontology has a nascent Observing System branch under the System branch which is also under the material entity branch as is OBI's device branch.

It does not make sense for some kinds of observing systems or measurement devices to end up on one subtree in one ontology while other observing systems and measurement devices end up in a different subtree in a different ontology; especially in the case I am working on where the use cases really do cross disciplinary boundaries (e.g., biology and energetic particles from the sun/cosmic radiation; and working out where organic exomoons are likely to be discovered in all of the thousands of extrasolar planets that have been discovered are our first two use cases). I think users would expect a single source for instrument types for example.

I think there are several options here, including:

1. Pulling both the ENVO tree and the OBI tree out into a science-wide tree encompassing all of the terms from all the scientific disciplines. Both ENVO and OBI would deprecate their existing trees once all the terms have been moved.
2. Either OBI or ENVO could change their scope to accept responsibility for the entire scope of measurement devices used in science. The ontology that doesn't step up to this would then deprecate their branch... This option doesn't really make sense to me, given the existing scopes of each.
3. I'm not sure if there is a third option other than making folks who are dealing with interdisciplinary science learn to deal with multiple conflicting lists of terms.

In any case there is a related issue https://github.com/OBOFoundry/COB/issues/190 on this topic. It was suggested to me that this should be a topic for the upcoming OBI meeting which I understand is in January?

[bpeters42]

Thank you for raising this. It would be great to coordinate the creation of a 'device' class that fits the broader needs of ontologies within COB, 'observation devices' or whatever you want to call it should be covered by it. I believe we are already working on that.

I do want to raise that we do have some limitations on what OBO / COB will cover. One is that we stick to non-quantum, non-relativistic physics. So at some point, things that NASA, or other scientists will want to describe will go beyond what OBO/COB will want to describe. Which is perfectly fine, and physics/science has always dealt with that. We just have to be clear where we deviate from each other.

• Bjoern

[DanBerrios]

@bpeters42 How exciting! I agree with both of you that we should have the proper subtrees of devices located in the proper area ontologies, and they should have common roots in a high-level ontology like COB (for now) or maybe the SDDO, the ontology Ruth mentions above for all Science Data Discovery (once we get it into OBO).

[ddooley]

Note that there has been discussion by some ENVO curator group people about the idea of cleaving off manufactured product branch into its own ontology. Devices of all kinds would in theory sit under that, but perhaps an upper level categorization of devices could organize further delegation to other ontologies. E.g. mechanical device (lever, scissors), electronic device (LED, CPU), biological device (eye, organ ... ok, maybe only artificially grown!), multi-component device (robot, sequencer, ...) ...