lucene-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Grant Ingersoll <gsing...@apache.org>
Subject Re: [VOTE] merge lucene/solr development (take 3)
Date Fri, 12 Mar 2010 14:39:59 GMT

On Mar 12, 2010, at 9:15 AM, Dennis Kubes wrote:

> Yes railroading.
> 
> Many people don't want this to occur.  More than just minus 2

Only 2 committers on Lucene and Solr were against it.  Those are the people that do the work.


> . Underlying concerns are not being addressed.  

I don't follow.  We've talked up and down about it.  Sometimes every last issue can't be addressed,
but I think we've all worked reasonably hard to address them.  We've addressed the release
issue, we've addressed the duplication/poaching issue.

> Vetos count.  

No, they don't.  It's majority approval.

> Ignoring that is ignoring how Apache operates.  

No, it's not.  Nobody is ignoring anything.  There are some valid concerns.  I think they
will be addressed by all of us being vigilant.  I have plenty of Lucene only projects in my
stable, as do most other Solr committers as well as all of the current Lucene committers.
 Nobody here wants Lucene to be consumed by Solr and HTTP server to be rammed down their throats.
  

Please remember that this isn't just Solr committers who want this.  Most of the people who
do the daily lifting on Lucene who are not Solr committers (Michael, Robert, Uwe is +0) are
for it as well.

> Merging projects is definitely a code change.  Getting around it by saying this is a
goal is fundamentally wrong.
> 
> 1) What prevents functionality be moved over into Lucene within the current project structure?
 Nothing, so why are we even discussing this.
> 
> 2) Why is Solr getting special treatment?  Because there is a lot of committer overlap?
 Should I propose to merge Nutch in too, lets just have one big project, no distinctions.

I have no problem with you proposing to bring in Nutch's overlap.  The fact is, the Board
doesn't like subprojects anyway and we are likely headed for some consolidation/spinning out
anyway (see the December Board Minutes).  Mahout is going after 0.3 is out.  I could totally
see spinning out the crawling stuff from Nutch and taking the good bits of Lucene in there
into java-dev.  (We should be working together on that stuff anyway as we all end up writing
the same code and the solution in the end will be better.)  But that wasn't what this discussion
was about.  If you can convince all of the committers in Nutch to go for it and then convince
all the Lucene committers as well, then go for it.  I'm certainly not going to force it, as
that is up to the Nutch community.

At any rate, as Otis said, "it's just software".  If it doesn't work, we can split back off.

> 
> 3) Why the big push here to blur project responsibilities? Idk, I keep wondering that
myself.

Please read the thread.  The arguments have been made over and over.  There is still going
to be a set of Lucene JARs and there is still going to be Solr JARs.  IDK why it is such a
big deal right back when most everyone who actually does the coding on the two projects is
for it.
Mime
View raw message