Main Navigation

User login

Devices Ontology WG Mtg 2008.12.09

Agenda/minutes for Device Ontology Working Group meeting 2008.12.09

Agenda

Logistics: Welcome, Minutes Review, Agenda Review
Search Facets
End Logistics: Next meeting, Action items, Attendees

Minutes

Logistics: Welcome, Minutes Review, Agenda Review

Minutes review Agenda approval

Search Facets

We have recently started to focus our attention on the definition of search facets with the goal of determining an appropriate organization of categories that can support a flexible, yet effective mechanism to find marine devices according to multiple criteria. This approach should play a synergistic role to better specify the use cases and, ultimately, the device ontology itself.

The list of search facets as well as related information and links will be maintained in the Search Facets page.

In today's meeting, we will also be looking at the Advanced Search page of the Earth System Grid as a very relevant example of faceted search. Note that the search facets can be rearranged to apply a desired search hierarchy.

Discussion:

Note the list of facets that has been created separately. That is the maintained list.
Question about what is the facet, in the case of operational restrictions: all the operational restrictions are one facet, or each are their own?
Now let's take a look at the Advanced Search page of the Earth System Grid, to appreciate what we mean by faceted search. Each button is a facet, so should operational restriction be one button?
What does one do with categories that are numeric? Create subcategories for each range? (What to call them? Subfacets, Named facets, names. Does it matter?)
What are the facets we should concentrate on? Some are more critical to the group, like sensor type. (Some still think 'sensor type' doesn't really mean anything....)
What's the big picture goal? Take into account use cases, users that want to do searches. Goal for today is to have agreement that this is possible approach.
One of more important things in the list is the concept of operational restriction. Platform, medium are types of operational restrictions. Capability might be better term. Some are more descriptions of potential use. Can we take out "Operational XXX" from the list and still have it make sense? (Maybe what is a restriction and what is a characteristic is in the eyes of the beholder.) Manufacturers tend to think of temperature, depth as the two restrictions they document.
Regarding the big picture goal, the reason we are trying out 'search facets' as a way to proceed is that almost all the existing use cases are really searches, dressed up. Or else, they can be achieved with an ontology built around concepts for search.
The email words were helpful in explaining the concept in a way that is attractive. Suggest we put the email words in the minutes ("find marine devices by multiple criteria").
It seems the idea of identifying facets or categories for search is a good one, and a good basis for designing the ontology.
Additional criteria: Availability status can be more general ('is device still made?'). Also: cost. Disposability of system or components (do you ever recover the instrument, or is it used up by the measurement?).
More criteria: Manufacturer, model. What I/O interfaces are supported (e.g., what can I attach to my system, by what method does it connect)?
What about depth, do we need all those terms for depth? Not really, just the property to support search. And maybe an additional set of terms.
These facets [Earth System Grid terms? or our list of facets? not sure. -ed.] are discovery vocabularies, used for searching. They are not necessarily the same thing that are in the core ontology, the ontology elements may be quite different. Yes, but the more detailed list is necessary to support all the broader cases.
The details of the device interfaces gets pretty hairy, no one search will require all the atomic-level interfaces -- do we really want to get into all that? Someday, maybe, but we have to prioritize. Breadth vs depth -- cover everything a little bit, or a few things well?
How to prioritize? Could look at ESG example, Alliance for Coastal Technologies example for how they categorized. Could look across all the search engines that mention instruments, whether for data search or other purposes. (Will be heavily weighted toward data search though. Need to allow for other use cases too.) GCMD is an example too.
Once we're prioritized what's next? Could probably build an ontology pretty quick with the given properties. But we'll need controlled vocabularies (terms) for 'filling out' the various categories/properties. Maybe could assign those to participants, one category per person.
Send one or two links of search engines/faceted searches that give us some indication of how devices are searched (what categories are important). (Announce this request in email.)
For next time improve upon the definitions of several of these. Sensor Type, Phenomenon, Instrument Functionality, Output Variable, Feature of Interest, Method, Gross Method, and Measurement are potentially conflicting/confusing/mergeable.
Proposed to find a concrete portal that we could apply these categories against, the OOSTethys portal could be a good choice.
Agreed not to meet in two weeks, next meeting is in 4 weeks.

End Logistics: Next meeting, Action items, Attendees

Action Items

[Carlos] put the email words in the minutes to explain what we're trying to do ("find marine devices by multiple criteria")
[all] Send one or two links of search engines/faceted searches that give us some indication of how devices are searched (what categories are important).
[Carlos] Announce this request in email.
[Carlos] Add new categories to the facets list:
- More general availability status (e.g., 'is the device still made?')
- cost
- whether device includes consumables (or is itself consumed during the process)
- manufacturer, model
- what input/output interfaces are supported (what can I attach to my system, and by what method?)
- what commands are supported