In-Reply-To: <200509061802.51089.josh@agliodbs.com>
References: <20050906135057.GA24136@granicus.if.org>
	<36e6829205090612575b954f38@mail.gmail.com>
	<3F60E55D-B558-4929-B614-C230EACFF473@redivi.com>
	<200509061802.51089.josh@agliodbs.com>
Mime-Version: 1.0 (Apple Message framework v734)
Message-Id: <E2E7828D-8F3B-457D-9E62-5CC92FB4E2B1@redivi.com>
Cc: jonah.harris@gmail.com, pgsql-hackers@postgresql.org,
	mark@mark.mielke.cc, nathan wagner <nw@hydaspes.if.org>
From: Bob Ippolito <bob@redivi.com>
Date: Tue, 6 Sep 2005 18:15:21 -0700
To: josh@agliodbs.com
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: uuid type for postgres


On Sep 6, 2005, at 6:02 PM, Josh Berkus wrote:

> Bob, People,
>
> Let me clarify my stance here, because it seems to be getting
> misrepresented.
>
> Mark (and Nathan) pushed at repaired UUID type for possible  
> inclusion in
> the core PostgreSQL distribution.  I'm not opposed to that,  
> provided that
> the portability, licensing, and bugs are worked out.  Why not?  We  
> have
> ipv6 data types, after all.
>
> However, Mark went on to suggest that we should recommend UUID over  
> SERIAL
> in the docs, and that we could consider dropping SERIAL entirely in  
> favor
> of UUID:
>
> ---quoth Mark------------------
> I suggest that UUID be recommended in place of SERIAL for certain
> classes of applications, and that it therefore belongs in the core.
> UUID and SERIAL can be used together (although, once you have a
> UUID, it may not be useful to also have a SERIAL).
> ---------------------------------
>
> This was what I objected to; I believe that the use-case for UUIDs is
> actually quite narrow and assert that it's a very bad idea to  
> promote them
> to most users.

I agree with you (Josh) completely, which is why I said:
"If the documentation gives the user a good idea of when to use UUID  
and when not, I think it would be a good addition."

.. the fact that the use-cases are narrow was implicit :)

Everything else I talked about was just implementation details.   
Summary: there are (several) UUID implementations out there that are  
appropriately licensed and easy enough to use, and a lot of OSes ship  
with pretty good implementations already.  Creating a decent UUID  
type should be relatively trivial, as far as those things go.

-bob