lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Miller <>
Subject Re: modularization discussion
Date Tue, 03 May 2011 17:46:54 GMT

On May 3, 2011, at 1:29 PM, Shai Erera wrote:

> I don't like that approach. Two years from now, if indeed your vision becomes the reality
(obviously, not everyone think like you), what would o.a.solr mean? Who will remember that
'suggest' (just picking an example) came from Solr? Who'd care?
> Why, when I will integrate several modules together, will I need to see o.a.lucene on
some, and o.a.solr on others, when both come from the same distro (even same tar.gz file,
e.g. modules)?
> What makes sense, at least to me, is that either we call everything o.a.lucene and solr
becomes o.a.lucene.solr (I know I've probably pissed off some people with that, sorry), or
we come up w/ a new namespace (proposed by Grant I think) o.a.lusolr. If we go with the second,
then we'll have 3 namespaces:
> * o.a.lucene for core Lucene stuff (e.g. Lucene core, benchmark?)
> * o.a.solr for pure/core Solr stuff
> * o.a.lusolr for shared modules.

Honestly, I could go for any of those. I can't bring myself to get caught up caring long term
what the package names are. You can't even make rules about that - they won't and shouldn't
stand over time.

> Picking a good package name is important. And deciding to call everything that came from
Solr o.a.solr, just to not offend someone, is not the right way to do things, at least IMO.

Yeah, its just not a sustainable idea for an open source project anyway.

> Mike, I do share with you the vision you outline, and I believe many of us do. It will
become a reality if we factor out modules from Solr and Lucene under /modules. It can also
become a reality if someone simply contributes under /modules alternative packages for e.g.
faceting, suggest, spellcheck etc. If those are good packages, I doubt "Solr" would be reluctant
to adopt them.

> Either way, it's the community that will dictate the future of itself, and not individuals.
Perhaps we should stop discussing what can possibly happen, and start doing things. Actions
get more results than endless threads. This have been stated on this thread numerous times
-- if a contribution is good, well coded, designed, thought of, it will go in. Whether it's
a refactoring of something, or a completely new code. I doubt there are people on this community
that can stand in the way of it.

This is really the crux of it. IMO, people should be much less concerned with how they perceive
others, and more concerned with just doing things. The Apache rules are set up to deal with
this type of thing. Those rules can get tricky, and nobody likes to fall back on them - but
when you have strong disagreement, that is what they are there for. Not everyone on a project
has to agree - nor do they have to have "pure open source" motives. That's just normal and
expected. We are a very varied group. The more differences the better IMHO.

Just as a reminder - a couple things I see repeatedly at Apache:

"community over code"
"merit does not expire"

Other than that, the doers do, occasionally we vote, and in general things move along.

> Shai

- Mark Miller

Lucene/Solr User Conference
May 25-26, San Francisco

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message