Re: "thou shalt not conflate meta-data with data"

From: Paul <paul_at_test.com>
Date: Thu, 03 Mar 2005 22:31:41 +0000
Message-ID: <4227904d$0$25330$ed2619ec@ptn-nntp-reader02.plus.net>

Neo wrote:

>>>When modeling "John is a person & John is male
>>>& male is a gender" which data is meta data?
>>>Which data "step[s] outside the language to talk
>>>about the language itself" ?
>>
>>the meta-data would be things like "gender G is of
>>varchar(10) type" or  "the table representing the predicate
>>'person P has gender G' is called T_Person".

>
> Thus it seems, Morgan's definition that meta data is "data about data"
> was along the right track, however it was too broad since male is data
> about John, but male is not considered meta data. Frosty's response
> that attributes of a table (ie Gender & Age) are examples of meta data
> was basically correct. And based on your example above, it seems a
> narrower definition might be: meta data is data that imposes some
> condition (structure, constraints, significance, etc) upon other data.
> For example, data for a relation's header might impose the condition
> that data in tuples represent persons and each value corresponds to
> predesignated attributes.

Maybe an alternative definition might be "meta-data is data about the representation of data".

You have data, but then you have to describe it using symbols in some way - to write it down for example, or to describe relationships between the data elements. Those symbols give a representation of the data which is distinct from the actual data itself.

But then you get into questions of how you define "data" and "representation". Ultimately the only rigorous definitions are provided by mathematics but these aren't very untuitive. I suppose if you *really* wanted to understand what metadata was, you'd have to study heavy-duty mathematical topics like set theory or type theory and probably some other theories I'm not even aware of.

Paul. Received on Thu Mar 03 2005 - 16:31:41 CST