Re: Name for new VACUUM

From: mlw <markw(at)mohawksoft(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Name for new VACUUM
Date: 2001-08-03 11:42:40
Message-ID: 3B6A8E30.719C4F1D@mohawksoft.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Tom Lane wrote:
>
> mlw <markw(at)mohawksoft(dot)com> writes:
> > Why rename VACUUM, why not create a new command RECLAIM, or something like
> > that. RECLAIM does the VACUUM NOLOCK, while vacuum does the locking.
>
> Um, that gets the default backwards IMHO, where "default" = "what
> existing scripts will do".

Changing how the default works is always tricky. Even if you improve something
dramatically, someone still will still gripe about the change.

>
> > The term RECLAIM will make more sense to new comers than VACUUM,
>
> What's your basis for claiming that?

I am so used to "vacuum" and postgresql, it makes perfect sense to me. Yet, I
gave a brief discussion at work a week ago about PostgreSQL and how we can use
it to offload SQL queries from Oracle. In the pros and cons part of the
discussion, people looked at me like I had two heads when I told them about
"vacuum." It wasn't obvious to them what it did.

The term "reclaim" may be a little more obvious, but I could be wrong. It is
just that the name vacuum, from the perspective of someone new to PostgreSQL,
is a bit obscure.
>
> In any case, VACUUM is the term already used in all our documentation.
> I have no appetite for trying to teach people and documents that
> currently know "you must do VACUUM periodically" that the new truth is
> "you must do VACUUM or RECLAIM periodically". All these discussions
> about which should be default aside, the bottom line is that the two
> pieces of code do more-or-less the same thing from a high level
> perspective. Calling them completely different names isn't going to
> make things easier for novices. Calling them different options of the
> same statement seems like the right thing to me.

I understand the documentation issue completely, and it is a very strong point.
However, saying that VACUUM NOLOCK and VACUUM LOCK do "more-or-less the same
thing" really isn't so. Think about it, the VACUUM LOCK, practically rebuilds a
tables representation, in older versions of Postgres didn't it actually rewrite
the table? The new behavior of vacuum doesn't do that at all.

Perhaps VACUUM gets changed to the new behavior, and the old behavior gets
renamed to DEFRAG or COMPRESS? Win/DOS users will find those names completely
obvious.

VACUUM DEFRAG?
VACUUM COMPRESS?

>
> regards, tom lane
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo(at)postgresql(dot)org

--
5-4-3-2-1 Thunderbirds are GO!
------------------------
http://www.mohawksoft.com

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message mlw 2001-08-03 11:52:14 Re: OID wraparound: summary and proposal
Previous Message Zeugswetter Andreas SB 2001-08-03 08:17:11 Proposal: OID wraparound: summary and proposal