Thread: simple or global column names?

[PostgreSQL 7.4RC2 on i686-pc-linux-gnu] [soon to upgrade to 8.x]

I have a simple schema design question.  I'm torn between:

   create table steps(step text, step_version int, substep text, substep_version int);

and:
   create table steps(step text, version int, substep text, substep_version int)

I.e., should a field in steps be "version" or "step_version"?  On one hand,
the "step_" prefix is redundant noise in this context, but for doing joins,
it seems like globally distinct names might make things clearer.

Are there other advantages/disadvantages to these naming schemes?

My goals (in this major schema reorganization) are simplicty, clarity, and
in particular, to facilitate nieve users' read-only ODBC access through
Excel or other GUI clients.

-- George Young

--
"Are the gods not just?"  "Oh no, child.
What would become of us if they were?" (CSL)

On Nov 8, 2005, at 10:28 AM, george young wrote:

> I.e., should a field in steps be "version" or "step_version"?  On
> one hand,
> the "step_" prefix is redundant noise in this context, but for
> doing joins,
> it seems like globally distinct names might make things clearer.
>
> Are there other advantages/disadvantages to these naming schemes?
>
> My goals (in this major schema reorganization) are simplicty,
> clarity, and
> in particular, to facilitate nieve users' read-only ODBC access
> through
> Excel or other GUI clients.


I find it annoying to read and write column names that are prefixed
with table names or some variant. It is easy enough to deal with
joins. Something like

select s.title as "Song Title", b.title as "Book Title" from song s,
book b ...;

And for read only access, views are an easy way to present whatever
column titles are the most user friendly.


John DeSoi, Ph.D.
http://pgedit.com/
Power Tools for PostgreSQL