Mailing-List: contact general-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: general@lucene.apache.org
Received-SPF: pass (nike.apache.org: local policy)
Date: Tue, 15 Jun 2010 12:42:55 -0700
To: general@lucene.apache.org
Subject: Re: [PMC] [DISCUSS] Lucy
Message-ID: <20100615194255.GA10968@rectangular.com>
References: <3A7AC73A-4448-42AF-875D-2D6D5A4762EC@apache.org>
 <alpine.DEB.1.10.1006141641180.15766@radix.cryptio.net>
 <2DD1574B-933A-4A09-9A73-25C5E7D38CAF@apache.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <2DD1574B-933A-4A09-9A73-25C5E7D38CAF@apache.org>
User-Agent: Mutt/1.5.13 (2006-08-11)
From: Marvin Humphrey <marvin@rectangular.com>

On Tue, Jun 15, 2010 at 07:52:12AM -0400, Grant Ingersoll wrote:

> OK.  So, in my mind, what this would take is:
> 1. Doing a release.
> 2. Showing some user list traction (i.e. real users)
> 3. Identifying and cultivating other contributors (via patches, discussions,
>    JIRA issues, helping others, etc.) who can then become committers.
> 
> (I agree w/ Hoss, it's less about code)

Lucy has struggled ever since Dave Balmain's effective withdrawal.  Had Dave
had been able to stay active longer, we would be in a totally different place
today.  Yet for obvious reasons, without users or releases it is difficult to
attract new developers to fill the void.

You could say that our vulnerability to the departure of a key contributor was
a weakness of the original proposal, but I don't think launching Lucy the way
we did, as a from-scratch development project, was a mistake -- many projects
start small.  We were knocked down but not out.

In the last year, I think we've done a pretty good job with the people we
have.  And certain technical milestones have been reached which have put a
Lucy release within reach.

There is a lot of promise for Lucy once the release comes.  Lucy's problems
getting a release out have been discussed, but KinoSearch's releases haven't
been handled well, either... which presents us with an opportunity.

If you go to CPAN and look for KinoSearch, you get the 0.1x branch that has
changed little since Lucy launched.  To get a version of KinoSearch that has
code extracted from Lucy (or more precisely, code designed and written for
Lucy then dual contributed), you need to get the dev branch, which is harder
to find.  There are various reasons that such a state of affairs has persisted
for so long, but the point is that when we release Lucy we will be able to
contrast it with KS 0.1x -- and that will make for good PR.

I fully expect that release to generate enough interest to satisfy Grant's
second criteria, "Showing some user list traction (i.e. real users)", and if
past is prologue, there will soon be sufficient contributors to satisfy the
third criteria as well.

Marvin Humphrey