Ontology for Devices: Frequently Asked Questions

DOWG Home || Next Telecon || FAQ || Minutes || Collaborations || Participants || Sign Up || Device Info || Use Cases || TRAC || SVN

(For logistical questions, see the Logistical FAQ.)

1. What kind of terms are you trying to put in your ontology?

   The first product we'd like to create is on ontology of device types, useful for classifying sensors and samplers into broad categories (e.g., for discovering sensors or the data they produce). This is the type of information people often want to use when they refer to groups of devices.

2. What do you mean by 'device type'?

   A very good question, and not easy to answer in a simple way. Typically people think of 'device type' as a class of words that can be used to group devices, or their data, for discovery. Some examples that come to mind are CTD, fluorometer, or pressure sensor.

   Unfortunately, these terms are difficult to fit into a rigorous lexicon -- a 'CTD', for instance, has multiple sensors, may include a depth or pressure sensor (not all do), and may be configurable with custom sensors that report voltages (so a CTD also can serve as a platform!). And while some people focus on the function of a device, others focus on specific items it measures or collects, how it does the measurement or collection, in what medium it is deployed, or what post-processing can be applied to it. Our ontology may have to reflect many or all of these qualities to be most useful to the community.

3. What science domains will the ontology address?

   The principal focus will be marine sensors, a domain which typically overlaps with environmental sensors. The intent will be to make the ontological framework suitable for describing other kinds of sensors beyond marine or environmental sensors. That way other groups can easily extend the work beyond the marine community.

4. Will you be defining terms to model and describe the internals of devices?

   Some terms useful to modeling sensor internals will probably be created, but initially we will not be trying to create an ontology of sensor-related terms. We would likely support extending the work in that direction, either through parallel projects or follow-on work.

5. How will you do this work and what will your schedule be?

   We intend to discuss those questions with the interested community. Many community projects requiring sensor vocabularies are just beginning in 2007, so quickly producing a list of terms (e.g., for "device types") would be valuable, but possibly difficult.

6. How much of a time commitment is required to sign up?

   No time commitment is required to get on the mailing list, other than the time you spend reading (or ignoring) your email messages. To be an effective contributor, you should expect to spend one to three hours every few weeks. This number may change depending on the work we undertake, and you are free to drop out at any time.

7. What do I need to know to be helpful?

   Initially, very little. The initial meetings will discuss goals, strategies, schedules, and tasks, and solicit volunteers for tasks. Some of the tasks will be fairly simple (taking notes or updating the site), not requiring special knowledge. These early meetings will also give you a sense of the type of work we will be doing.

   To actively contribute to the ontology development, you should have some knowledge of marine or environmental sensors and possibly other marine science devices (e.g., samplers, analysis equipment), and some understanding of the goal and process of ontology development. We can provide references to tutorials as needed; a classic tutorial is Ontology 101, and a simple graphical presentation is the Ontologies section (p44) in the workshop training presentation (PDF) from the 2005 vocabulary workshop.

   We are keeping a list of reference materials, but we do not expect you to have read this before the first meeting! We do ask that you help us add relevant references materials to this folder.

8. You seem to think it will be hard to create an ontology sensor types, do you? If so, what makes you think that? Can't you just make up a list?

   We think it may be hard, because we have tried it several times and identified some challenges. It is certainly possible to make up a list, and our sensor types vocabulary references page contains references to existing lists of sensor types. The hard part is making up a list that is semantically precise and useful across many data systems (read: interoperable).

9. How will you tell if you have a good ontology?

   Above all, the ontology should be valid across as wide a range of potential uses, and users, as possible. Our practical and usability goals for the terms in this ontology include the following:

   • terms are unambiguous: it is clear which term(s) apply to each instrument instance
   • terms are as close as possible to common usage
   • syntax (capitalization, underscores, word order) is consistent
   • completeness: at each level of the ontology, all possible cases are addressed
   This list is subject to change by the members of the working group.

10. Are you using a standard ontological approach?

    This depends on the discussions with the participants. Some of the initiators have some suggestions on how to proceed, but this will be considered at the first few telecons.

11. What classification properties will you use?

    This is not clear; depending on community input, we may try to 'discover' the classification properties as we go through the process. Properties that have been proposed include:

    • purpose(s) of the device
    • location where the device is deployed
    • method of operation of the device (perhaps including whether it is remote or in-situ)
    • real-world properties sensed by the device
    • variables output by the device
    • post-processing algorithms suitable to the device