Re: [HACKERS] Proposal for new SET variables for optimizer costs

At 18:31 5/02/00 -0500, Tom Lane wrote:
>Philip Warner <pjw@rhyme.com.au> writes:
>> Any chance of prefixing the 'set' variable names with 'PG_' or 'PG_OPT_' or
>> something similar? Or doing something else to differentiate them from
>> user-declared SQL variables?
>
>I see no need to do that, since the *only* place these names exist is
>in the SET command (and its friends SHOW and RESET), and SET exists only
>to set system control variables.  There are no user-declared SQL
>variables.
>
>The names are quite long and underscore-filled enough without adding
>unnecessary prefixes, IMHO ;-)

I agree, given their complexity, they are unlikely to conflict with future
SQL names, but the SET statment *is* part of the SQL standard, and I
thought it would be good to be cautious in the names you choose. This would
avoid any possible future conflict, as well as make it clear from the
outset that they are *not* standard SQL names.

Another option would be to add another command, eg. 'PG', which is used for
all non-SQLxx commands:
  PG SET somename = somevalue  PG VACUUM

...etc. But this has the disctinct disadvantage of being more work, and
being cumbersome in comparison to changing names. The transition could be
managed by supporting old commands until version 8.0, with an appropriate
notice.

Just my 0.02c worth.

----------------------------------------------------------------
Philip Warner                    |     __---_____
Albatross Consulting Pty. Ltd.   |----/       -  \
(A.C.N. 008 659 498)             |          /(@)   ______---_
Tel: +61-03-5367 7422            |                 _________  \
Fax: +61-03-5367 7430            |                 ___________ |
Http://www.rhyme.com.au          |                /           \|                                |    --________--
PGP key available upon request,  |  /
and from pgp5.ai.mit.edu:11371   |/