Horton

Re: WIP: extensible enums

Peter Eisentraut <peter_e@gmx.net>

View thread
From: Peter Eisentraut <peter_e@gmx.net>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Andrew Dunstan <andrew@dunslane.net>, Bruce Momjian <bruce@momjian.us>, Dean Rasheed <dean.a.rasheed@gmail.com>, Greg Stark <gsstark@mit.edu>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2010-11-13T17:30:03Z
Lists: pgsql-hackers 
On fre, 2010-11-12 at 17:19 -0500, Robert Haas wrote:
> If we allow users to name objects, we ought to make every effort to
> also allow renaming them.  In my mind, the only way renaming is too
> marginal to be useful is if the feature itself is too marginal to be
> useful.

The bottom line is, any kind of database object needs to be changeable
and removable, otherwise there will always be hesitations about its use.
And when there are hesitations about the use, it's often easiest not to
bother.

I remember ten years ago or so we used to send people away who requested
the ability to drop columns, claiming they didn't plan their database
properly, or they should load it from scratch.  Nowadays that is
ludicrous; databases live forever, development is agile, everything
needs to be changeable.