Re: VIEWS compared to Nodes as Windows into data
Date: Tue, 27 Apr 2004 14:20:31 -0400
Message-ID: <NKydneXQxeGlORPd4p2dnA_at_comcast.com>
<orthogonal>
> Yes and what just struck me with what you said is that with an RDBMS it is
> actually possible for a user to write a report using the query language
> against something that is NOT just a logical view of the data -- YIKES!
It depends. Most people would say that the tables themselves represent a logical view of the data.
In most senses of the word "logical", that's true. The user doesn't need
to know what disk the rows are on, or even whether they are all on one
disk. The user doesn't need to know whether the data blocks that contain
the table rows also contain rows from other tables or not. The user doesn't
need to know whether the data is compressed or not.
And, in particular, a query is independent of any columns it never mentions.
Those columns can be altered, or dropped or new columns added and the query
still works. It's also independent of any tables not mentioned. New tables
can be added, or existing tables can be altered or dropped, and the query
will still work.
That's a considerable amount of data independence.
Incidentally, by "users of the data", I probably mean something different than you do by "end-user". I'm using the term "users of the data" to include application programs that access the data, and then do things to it, before presenting it to the user. All of the good documentation I've read cautions against using SQL as an ad-hoc query tool for the "End user". You have to know too much to use it right. Something like "Business Objects" might be better. If there is a more standard term for what I mean by "users of the data", then I'd like to know it.
But I'm not going to adopt "standard terminology" just to be "politically correct". That reduces communication rather than enhancing it. Double plus ungood.
</orthogonal> Received on Tue Apr 27 2004 - 20:20:31 CEST