Terms - Proposals

[Ed Shaya]

Martin @ ROE wrote:

> Tony Linde wrote:
>
>>> data collection
>>> dataset (which I think is just my term for data collection)  
>>
>
> Perhaps we could consider a 'dataset' as the data from a particular 
> instrument/survey, and a collection as number of datasets.  For 
> example, SSA, USNO-B and 2MASS are three datasets in one data 
> collection at the ROE datacenter...?
>
For published tables, we refer to a published set of tables, plus 
metadata about the set, as a dataset.  If the same team works on similar 
science with similar instruments and publishes again, then the two 
datasets form a new collection that grows with further such 
publications.  This collection can be part of larger collection.

>>> database  
>>
>
> The storage medium.  A database might be SQL Server, or a collection 
> of FITS files.
>
Database is a slippery term used in several ways.  If you say that you 
are buying a database from Oracle, you are of course talking about 
purchasing an application with no data.  All of the data accessed by a 
given database application is also called the database.  And often  
that  data is  divided into  collections which are separately  called  
databases.   Sometimes a collection of data is called a  database even 
though there is no existing software to read it.  Webster's Ninth 
Collegiate Dictionary
says n (1967):   a collection of data organized esp. for rapid search  
and retrieval  (as by a computer).

Perhaps we need to modify this  definition  to
n (2004) a collection of datasets organized at a datacenter for rapid 
search and retrievable by a computer network.  This  explicitly leaves 
out applications, datasets that are not accessible and  differentiates 
it from a collection because it can not span over several datacenters.  
A new point, I am aware.

>
>>> mirror  copy  
>>
Mirror - a rapidly updating exact match to a collection or database
Copy - an exact match to a collection or dataset or table at a given time
I think we need another term to mean something very similar but 
different in some minor way.  Some possible terms that come to mind:
facsimile [L. fac simile make similar (1691)]
representation  [fr,MF  re presenter to present again (15c)]
replica [fr. replicare to repeat fr. L, to fold back]
reproduction
likeness

Of these I prefer likeness and representation because they signify an 
inexact copy more than the others do.  Then we need to discuss 
attributes of inexactness.

Ed