meetingnotes20040422 - Marine Metadata Wiki

meetingnotes20040422

Standard Naming Exercise

meetingnotes20040422

Initial Notes and Issues

AOSN goals:

  • Ability to work with data set in common language

  • Consistency: Common look and feel

Other goals:

  • arrive at a list of variable names that we all agree upon

  • determine the process through which we identify this list (do we build on what a majority of people use or let individuals add their terms indefinately?)

Overall priorities:

  • common language of standard variable names

  • ability to differentiate between varaibles and group similar measurements

  • ability to manage units

  • ability to search

Possible Breakout Tasks

  • Define the construction of a name

    • what's in it

    • what order

  • Define the elements of interest about a variable

    • describe each parameter

    • synonyms or equivalencies

    • what elements must follow a content standard and what's not?

  • How do we pick which names to choose from (for 'Characteristic' and 'Domain'?

    • what are mature standards?

    • what level of detail is appropriate?

No breakouts at this first meeting, but the group discussed many of these questions generally.

Questions

  • How to make it so we don't need more meetings like this?

  • Do we even want to have standard names?

  • How to reconcile what other people are used to?

    • How do we handle synonyms?

  • How does end user community affect the solution?

  • How do you deal with metadata when it comes to tools that can't see it (e.g., ODV)?

  • How do we get all this information into the variable name?

  • How does the proposed solution allow for search?

  • Which elements require standardized content?

  • Can we construct a naming convention that conveys as much as possible (i.e., the object and the variable being measured?)

  • How to ensure that people will actually use a naming convention? (e.g., JGOFS has one but this is not necessarily reflected in their files)

Examples

  • Are "currents" ocean currents or electrical currents?

  • Do we specify the information for the raw measurement, or the physical quantity it represents?

  • "depth" is the distance below sea level, while "altitude" is the distance above sea level; also have heights relative to the sea bottom...many terms are context-dependent

  • parameters = variables = characteristics

  • object = domain (e.g, air, water)

  • issue of users' name vs. variable name vs. modifiers

General Comments

  • there is a complication with NetCDF programming in that the programmer needs to use unique names for each variable to differentiate if there are different kinds of temperatures or the same measurement at different depths. NetCDF doesn't offer file structure like a database (where you could use a composite key to differentiate records)

  • the name is a handle that the user can use to locate information (variable name -- no spaces for MatLAB)

  • perhaps there are two different challenges: 1) how to access variables when programming and 2) how will the user know what these are?

  • DMAC is advocating that metadata and data sets will be bundled together

  • it is impossible to create the perfect standard variable name

Development of Spreadsheet Template

  • include information we will need to know (this exercise expanded beyond the standard variable name question into an identification of other metadata elements that promote interoperabiity):

    • field/column name/differentiator/modifier/index/subclass (e.g, daily temperature and average temperature)

    • domain/object/entity

    • parameter/characteristic/variable/standard variable/attribute

    • units

    • reference to a standard for the units (e.g., different scales for degrees C)

    • definition

    • maybe also add instrument, long name (for plotting), and user's name

Next Steps

  • complete spreadsheet with the above information from your programs and send to John (suggestion not to generate standard names by instrument but alphabetically)

  • John will compile the spreadsheets into a master spreadsheet; cross-reference between the programs

  • Stephanie and ?? will include external references for the spreadsheet

  • should try to include representatives of NEOCO, CI-CORE, CIMT, etc. if they wish to participate

  • the group will reconvene on May 18th, 8:30 - 11:30 in Building G to use the master spreadsheet to cross-reference between the programs, deal with issues such as abbreviations...

Attachments

Posted April 15th, 2005 by graybeal