ben.powell@CLA.CO.UK wrote: >Because I'm looking at this from a SAS perspective this is relevent here >although not directly involving SAS examples - I might work a couple in - >but also because of the experts who might be able to offer their thoughts >that are known to lurk on occasion. > >I have a problem with "metadata" - "data about data". The most common use >of >the term as far as I can see actually refers to "data about *things*", i.e. >data!. > >For instance a classic example is of a catalogue wherein the data refer to >books, for instance. I don't regard that as metadata but it seems the >bibliographic community does. > >It seems my conception of metadata is what is often referred to as meta >metadata. For instance in a m.metadata driven application the field names >(and their table references) AUTHOR, FAMLY_NM, GVN_NM might reside in a >table called identifiers, where that table would be the metadata. "Jane >Eyre", "EYRE" and "JANE" would not be. It might also include field length, >type, format, source, ETL rules, etc. The point being that you can alter >the >underlying schema easily and adjust the entire application in one step, >without identifying each instance of a var in a table, report, view, >query... > > >Is this just a popular nomenclature error or am I missing the point >somewhere? If all a database contains is data about stuff, then how can you >call that data "metadata"? Surely the database must contain data about >stuff >and only data defining that data could be referred to as metadata? Googling >"metadata misconceptions" (for ironical purposes) the following pdf is >indicative: > >www.cancore.ca/guidelines/CanCore_Guidelines_Introduction_2.0.pdf > >"A modern *metadata*1 infrastructure should be capable of ingesting, >merging, indexing, enhancing, and presenting to the user, *metadata*1 from >a >variety of sources describing a variety of objects." > >In the above example *m*1 is the correct useage and *m*2 actually refers to >presentation of *data*, not metadata? Another example: a DTD is metadata; >an >XML data transfer is not, though it may contain metadata.

Let me chime in. And remember how much you can buy for two cents. :-)

It's data if you need to analyze it or display it. So the same information can be data in one context and metadata in another. I don't see that you're mistaken anywhere, just too smart for your own good. :-) You've figured out that a lot of the data-vs-metadata blather is semantics.

If you have a large study on word frequencies of famous authors (there are hundreds of published papers on such issues), and you store all the text of "Jane Eyre" as one piece of your study, that is data. For the purposes of your study, the information on the authorship of the book, Charlotte Bronte, the Bronte sisters, the literary pseudonyms they used, and so on, are metadata. If I come along and use your data to try to prove that Jane Eyre was actually written by someone else, then your metadata may be crucial data for me. (Word frequency analyses are often used for the very purpose of attempting to figure out who wrote anonymous works, or whether writer X actually wrote something credited to writer Y.)

So call it what you will. Just use it well.

David -- David L. Cassell mathematical statistician Design Pathways 3115 NW Norwood Pl. Corvallis OR 97330

_________________________________________________________________ Don’t just search. Find. Check out the new MSN Search! http://search.msn.click-url.com/go/onm00200636ave/direct/01/