Re: A query planner that learns

From: AgentM <agentm(at)themactionfaction(dot)com>
To: pgsql general <pgsql-general(at)postgresql(dot)org>
Subject: Re: A query planner that learns
Date: 2006-10-16 20:33:10
Message-ID: 0F0C580B-7ECD-4333-B387-22CC7A0D0959@themactionfaction.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general


On Oct 16, 2006, at 16:17 , Madison Kelly wrote:

> Alvaro Herrera wrote:
>> Jochem van Dieten wrote:
>>> Scott Marlowe wrote:
>>>> While all the talk of a hinting system over in hackers and
>>>> perform is
>>>> good, and I have a few queries that could live with a simple
>>>> hint system
>>>> pop up now and again, I keep thinking that a query planner that
>>>> learns
>>> >from its mistakes over time is far more desirable.
>>>> Is it reasonable or possible for the system to have a way to
>>>> look at
>>>> query plans it's run and look for obvious mistakes its made,
>>>> like being
>>>> off by a factor of 10 or more in estimations, and slowly learn
>>>> to apply
>>>> its own hints?
>>> Technically it is very feasible. But I think you might want to
>>> check US Patent 6,763,359 before you start writing any code.
>> I think it would be a very good idea if you guys stopped looking
>> at the
>> US patent database. It does no good to anyone. There's no way we
>> can
>> avoid stomping on a patent or another -- there are patents for
>> everything.
>
> Hasn't IBM release a pile of it's patents for use (or at least
> stated they won't sue) to OSS projects? If so, is this patent
> covered by that "amnesty"?
>
> Simply ignoring patents because "there is a patent for everything"
> is a recipe for disaster. Companies like MS are running out of ways
> to tear open OSS and they are certainly not above (below?) suing
> the heck out of OSS projects for patent infringement.
>
> What's needed is reform in the USPO. Call you congress (wo)man and
> complain, but don't flaunt the law; you will lose.

Alvaro's advice is sound. If the patent holder can prove that a
developer looked at a patent (for example, from an email in a mailing
list archive) and the project proceeded with the implementation
regardless, malice can been shown and "damages" can be substantially
higher. You're screwed either way but your safest bet is to never
look at patents.

Disclaimer: I am not a lawyer- I don't even like lawyers.

-M

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Martijn van Oosterhout 2006-10-16 20:33:53 Re: datatype preceded by underscore creates array
Previous Message Carlos H. Reimer 2006-10-16 20:30:11 RES: RES: Dates rejected