Search on lucene.apache.org

[Jukka Zitting <ju...@gmail.com>]

Hi,

As discussed at the ApacheCon, we're the search experts at Apache but
still most of our web sites at lucene.apache.org use a Google search
box. And the Google search isn't even that powerful, as it doesn't
cover our mailing lists, wikis, issue trackers, etc. Let's fix that
using our own search technology!

Lucid Imagination already implemented a search at
http://www.lucidimagination.com/search/, though that implementation
also contains stuff from their web site. Perhaps we could ask them to
set up a version that only returns stuff from apache.org.

Would people be worried about the Lucid branding on search result
pages? I don't think that's too much of a problem as the current
search results are already Google-branded. Anyway, if someone feels
this is a problem (and has the time and energy to fix it), we could
also ask for an infra zone where a similar ASF-hosted search service
could be set up. Or perhaps Lucid would be willing to give us an
ASF-branded version of the search with just a "search results brought
to you by Lucid Imagination" note attached?

PS. I can think of a number of Apache projects who'd love a similar setup.

BR,

Jukka Zitting

Re: Search on lucene.apache.org

[Grant Ingersoll <gs...@apache.org>]

We (Lucid) are interested and thought it would be best to submit some  
patches and details for people to review in the near future.

Cheers,
Grant

On Mar 31, 2009, at 6:54 AM, Michael McCandless wrote:

> +1
>
> This is a great idea: we should certainly eat our own dog food.  I
> think it'd be great to have the search powered by Lucid (Lucid being
> willing of course...).  I would also love to expand this to other
> Apache sites, but Lucid doesn't (yet?) search other Apache sites.
>
> There is some important risks involved... we'd (Lucid'd) need to do a
> good / responsive job fixing issues, keeping server up, data fresh,
> etc.  If not, then the average new possible user might go through this
> process: I heard of Lucene; should we use it?  Oh, Lucene powers
> Apache's site, so let's go play with that search.  Oh, it has some
> weird problems, and it's sluggish, and this other search site does a
> better job, and XYZ, etc...  Maybe Lucene isn't right for us.
>
> Meaning, if we do this, we really need to do it right.  If Lucid
> powers it, Lucid needs to fix issues quickly; or perhaps Lucid could
> make available (open source) the underlying configuration / code that
> powers their Solr instance, so that the community can jump in and
> produce patches to fix things.  EG there are some things about Lucid's
> search I'd love to fix -- I've interacted with them on these already,
> but if instead I could open a Jira issue, put a patch on, iterate,
> have it committed and in 30 minutes it's pushed to a dev instance
> which we all can test, and every so often production cuts over to that
> dev instance, then that'd be fabulous.
>
> Mike
>
> On Tue, Mar 31, 2009 at 5:42 AM, Jukka Zitting <jukka.zitting@gmail.com 
> > wrote:
>> Hi,
>>
>> As discussed at the ApacheCon, we're the search experts at Apache but
>> still most of our web sites at lucene.apache.org use a Google search
>> box. And the Google search isn't even that powerful, as it doesn't
>> cover our mailing lists, wikis, issue trackers, etc. Let's fix that
>> using our own search technology!
>>
>> Lucid Imagination already implemented a search at
>> http://www.lucidimagination.com/search/, though that implementation
>> also contains stuff from their web site. Perhaps we could ask them to
>> set up a version that only returns stuff from apache.org.
>>
>> Would people be worried about the Lucid branding on search result
>> pages? I don't think that's too much of a problem as the current
>> search results are already Google-branded. Anyway, if someone feels
>> this is a problem (and has the time and energy to fix it), we could
>> also ask for an infra zone where a similar ASF-hosted search service
>> could be set up. Or perhaps Lucid would be willing to give us an
>> ASF-branded version of the search with just a "search results brought
>> to you by Lucid Imagination" note attached?
>>
>> PS. I can think of a number of Apache projects who'd love a similar  
>> setup.
>>
>> BR,
>>
>> Jukka Zitting
>>

--------------------------
Grant Ingersoll
http://www.lucidimagination.com/

Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
using Solr/Lucene:
http://www.lucidimagination.com/search

Re: Search on lucene.apache.org

[Michael McCandless <lu...@mikemccandless.com>]

+1

This is a great idea: we should certainly eat our own dog food.  I
think it'd be great to have the search powered by Lucid (Lucid being
willing of course...).  I would also love to expand this to other
Apache sites, but Lucid doesn't (yet?) search other Apache sites.

There is some important risks involved... we'd (Lucid'd) need to do a
good / responsive job fixing issues, keeping server up, data fresh,
etc.  If not, then the average new possible user might go through this
process: I heard of Lucene; should we use it?  Oh, Lucene powers
Apache's site, so let's go play with that search.  Oh, it has some
weird problems, and it's sluggish, and this other search site does a
better job, and XYZ, etc...  Maybe Lucene isn't right for us.

Meaning, if we do this, we really need to do it right.  If Lucid
powers it, Lucid needs to fix issues quickly; or perhaps Lucid could
make available (open source) the underlying configuration / code that
powers their Solr instance, so that the community can jump in and
produce patches to fix things.  EG there are some things about Lucid's
search I'd love to fix -- I've interacted with them on these already,
but if instead I could open a Jira issue, put a patch on, iterate,
have it committed and in 30 minutes it's pushed to a dev instance
which we all can test, and every so often production cuts over to that
dev instance, then that'd be fabulous.

Mike

On Tue, Mar 31, 2009 at 5:42 AM, Jukka Zitting <ju...@gmail.com> wrote:
> Hi,
>
> As discussed at the ApacheCon, we're the search experts at Apache but
> still most of our web sites at lucene.apache.org use a Google search
> box. And the Google search isn't even that powerful, as it doesn't
> cover our mailing lists, wikis, issue trackers, etc. Let's fix that
> using our own search technology!
>
> Lucid Imagination already implemented a search at
> http://www.lucidimagination.com/search/, though that implementation
> also contains stuff from their web site. Perhaps we could ask them to
> set up a version that only returns stuff from apache.org.
>
> Would people be worried about the Lucid branding on search result
> pages? I don't think that's too much of a problem as the current
> search results are already Google-branded. Anyway, if someone feels
> this is a problem (and has the time and energy to fix it), we could
> also ask for an infra zone where a similar ASF-hosted search service
> could be set up. Or perhaps Lucid would be willing to give us an
> ASF-branded version of the search with just a "search results brought
> to you by Lucid Imagination" note attached?
>
> PS. I can think of a number of Apache projects who'd love a similar setup.
>
> BR,
>
> Jukka Zitting
>