Re: Pluggable Indexes (was Re: rmgr hooks (v2))

On Wed, 2009-01-21 at 18:24 +0200, Heikki Linnakangas wrote:

> If we allow them to develop as separate projects, then whenever they
> are ready they can be used with particular releases.

> Developing a new indexam is not something you do over the weekend.
> It's a long way from design to an implementation robust enough that
> anyone cares about crash recovery. Short-circuiting the release cycle
> with a plugin won't get you a production-ready indexam much sooner.

You're assuming that somebody is starting from scratch and that they
don't have access to index and/or Postgres experts.

There are already research projects in various forms of new index. This
would further encourage that. There are also companies such as CopperEye
that sell indexes for use in other RDBMS, that would be easily able to
adapt their technology to Postgres.

They could also be adapting one of the existing index types for use in a
particular application. Various ideas present themselves.

I'm not trying to persuade you to personally work on indexes. I'm trying
to persuade you to let others work on indexes without your approval.
They already can, though they cannot make them production ready without
this and I see no reason to prevent them. We're not talking about
including their code in Postgres, we're talking about allowing them not
to.

Bruce Lindsay, IBM Fellow and long term DB guru was interviewed in 2005:
Q: If you magically had enough extra time to do one additional thing at
work that you're not doing now, what would it be?

"I think I would work on indexing a little harder".

(He mentions XML indexing, multi-dimensional indexing etc)
[Taken from SIGMOD Record, June 2005]

--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Training, Services and Support