While it is tempting, there are other big benefits from the Lucene 
port, not least of which is sharing indexes with a Java (or python, or 
perl, or...) app (which I've seen being required). Making use of the 
knowledge built into the various analyzers in lucene is also something 
I am too lazy to reimplement myself).

That said, I'm sure as heck going to use Odeum in the interim (and when 
fighting with gcj just ain't worth it).

-Brian

On Apr 12, 2005, at 9:00 PM, Zed A. Shaw wrote:

> I'd say it's a little lower level than Lucene, but much HIGHER level
> than a Lucene+Java+GJC+CNI+Ruby implementation.  :-)
>
> Seriously though, I'd say it's just a little lower than Lucene and 
> would
> be the major building block for a similar system.
>
> If you're interested in dropping your labor of love and working on
> something more extensive then I'd contribute as I'm going to do it
> anyway.  My focus though is on code searching for specific analysis
> needs.
>
> Zed
>
> On Tue, 2005-04-12 at 21:58 +0900, Brian McCallister wrote:
>> Awesome!
>>
>> I've not used Odeum, but how does it compare to Lucene?
>>
>> This may take some pressure off of me to get the Lucene bindings out
>> ASAP, though I am definitely still doing it -- too much bloody time
>> spent playing with GCJ and CNI already. ;-)
>>
>> -Brian
>>
>> On Apr 12, 2005, at 8:12 AM, Zed A. Shaw wrote:
>>
>>> Hi,
>>>
>>> This is just a quick announcement of a Ruby extension to Mikio
>>> Hirabayashi's Odeum Inverted Index library which is part of both
>>> Estraier (http://estraier.sourceforge.net/ ) and QDBM
>>> (http://qdbm.sourceforge.net/ ).
>>>
>>> You can download the source package from:
>>>
>>> http://www.zedshaw.com/projects/ruby_odeum/
>>>
>>> Ruby/Odeum lets you index a set of documents (including meta-data) 
>>> and
>>> then use the resulting index to do fast searches by words.  The
>>> extension supports most of the features and should become more
>>> idiomatic
>>> as I work on it.
>>>
>>> Comments are welcomed.
>>>
>>> Zed
>>>
>>>
>>
>>
>>
>>
>
>