DomainLessons - Marine Metadata Wiki
Standard Naming Exercise
| exercisegoals | detailedplan for task | externalstandards we can use | ourdocuments | lessonslearned |
| Home / Introduction | (old home page) | participantslist | ||
Domain Element Lessons
This page contains lessons learned about the 'Domain' element of our variable metadata.
One of the original notions was that a variable name could be made up of a domain, e.g., 'ocean', and what's measured there ('temperature'). So the selection of domain names was biased in the direction of domains that would clarify variable names.
What's a Good Domain?
-
Atmosphere felt like a "good" (useful, clear, understandable) domain
-
-
Corresponding concepts are hydrosphere, lithosphere
-
-
Ocean feels like a good domain
-
-
but should the domain change as the sensor goes from the ocean to the estuary to the river to the lake?
-
in some cases we are definitely measuring the ocean
-
-
that's what people would search on
-
but what to do about the other issues? (see next section)
-
-
What Level of Detail?
-
Giving everything the same level of domain (e.g., sphere) is too broad for some purposes
-
But it's important to recognize domain relationships
-
A hierarchy of domains can be effectively exploited
-
-
see bottom of document for hierarchy so far
-
-
Specify the most precise domain you can
-
-
we understand that the more precise name implies the higher-level domains
-
Odd Domains
-
What's the domain when the measurement applies to some man-made thing?
-
-
the thing itself -- an instrument, platform, or sensor
-
-
What's the domain for quality control flags?
-
-
the data they refer to
-
-
What's the domain for latitude and longitude?
-
-
some argue for no domain, others for earth or globe or geosphere
-
-
(lat/lon exists also on other planets)
-
geo recommended as a generic term that is in common use
-
-
we settled on geo since it had the least amount of complaining
-
-
What about time and time zones?
-
-
this is just lat/lon, rotationally speaking
-
decided time is unique unto itself, gets its own domain
-
time zone is really a longitudinal reference, so geo for it (could also be a setting)
-
Is it a Domain?
-
wind: no
-
-
this is the measurement being taken, more than the component of atmosphere being measured
-
comparative example:
-
Which Thing is the Domain?
-
When measuring distances
-
-
the material you're measure through? (more clear in English usage) or
-
the material you're measuring to (more uniform/rigorous)?
-
For example, altitude above ocean floor -- is the domain 'ocean', or 'benthos'?
-
we thought that the To usage was clearer
-
-
the Domain for altitude above sea level was sealevel (hydrosphere.ocean.sealevel)
-
the Domain for altitude above ocean floor was benthos (lithosphere.benthos)
-
-
-
When measuring differences across domains (e.g., temperature difference)
-
-
a joint domain
-
the higher-level domain
-
a new domain
-
just one or the other
-
no domain
-
so far, we have used the higher-level domain (I think)
-
-
When measuring locations
-
-
the substance you're In (ocean)?, or
-
the platform you're on (platform)?, or
-
the sensor making the measurement (sensor)?
-
we thought it's the substance you're in, except
-
-
the derived value MidDepth, the midpoint of a dive, was a deployment domain
-
-
Domain List
-
hydrosphere
-
-
ocean
-
-
sealevel
-
seasurface (same meaning as sealevel but used in different context)
-
-
-
atmosphere
-
lithosphere
-
-
benthos
-
-
observatory
-
-
platform
-
-
instrument (definition TBD; typically more complicated than sensor; reporting center?)
-
-
sensor (definition TBD; makes single measurement?)
-
-
-
-
deployment (some measurements are not about the device itself, but about its usage)
-
data
