You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-dev@lucene.apache.org by Doug Cutting <cu...@apache.org> on 2006/02/25 00:18:33 UTC

The Lucene Web Service

Have folks seen this?

http://lucene-ws.net/

It's getting some attention:

http://www.intertwingly.net/blog/2006/02/24/Lucene-Web-Service-API

Doug

Re: The Lucene Web Service

Posted by Chris Hostetter <ho...@fucit.org>.
: > Have folks seen this?

: Not until now.

Nor I.

: The standards it uses are interesting, but I don't know enough about
: them to know if the benefits outweigh the costs.

In general, it seems like it's more targeted at really simple indexes.  IN
paticular, as a way to have a single server supporting many simple indexes
all completely administrated via the web -- in much the same way a lot of
blog software is targeted at being able to run many blogs, allowing users
people to create blogs on the fly.

>From what i can tell looking at the API, you can't define much of a schema
-- just tell it which field is your default searchfield, and which fields
to use for the id, title, and summary -- neccessary concepts for it to
know about to support OpenSearch.

: The most interesting thing is perhaps OpenSearch...  it enables
: syndication (wasn't an original goal of Solr, but it's tough to say
: what others might be interested in.)

I looked into OpenSearch back when A9 first announced it.  It's basically
just a spec for returning search data in either RSS or Atom formats --
using a custom namespace to add metadata like the totl number of records.

The only part i thought was really intersting was the "OpenSearch
Description" which is their XML files for
publicizing/identifying/autodiscovering your OpenSearch service.

In short: It seemed to me at the time, that if we ever wanted to power an
opensearch on Sol(a)r it would be easy to do using the standard query
handler and a server-side XSLT to generate the RSS output.

There may be more features to OpenSearch now, or I may have missed some
when i looked into it before, but I don't think supporting it "natively"
would be hard if there was interest.

: >From a developer with prior Lucene exposure, I think Solr should be
: more intuitive and straightforward.

agreed.

: I'd like to hear if anyone else has perspectives on this...
: What are the usecases where one would want to use the Lucene Web
: Service API rather than something like Solr or Nutch?

I imagine that if i'm an web hosting provider who wants to support easy to
use webservices for my clients (like WordPress, and Twiki)  without
needing to manually setup a lot of stuff per customer -- running a single
instance of lucene-ws that they can create indexes in self serve might be
apealing.


-Hoss


Re: The Lucene Web Service

Posted by Yonik Seeley <ys...@gmail.com>.
On 2/24/06, Doug Cutting <cu...@apache.org> wrote:
> Have folks seen this?
>
> http://lucene-ws.net/

Not until now.
The standards it uses are interesting, but I don't know enough about
them to know if the benefits outweigh the costs.

The most interesting thing is perhaps OpenSearch...  it enables
syndication (wasn't an original goal of Solr, but it's tough to say
what others might be interested in.)

Compared to solr:
  - richer in standards support
  - it might be a natural fit for indexing blogs given all the atom support?
  - not as much functionality
  - couples the client much more tightly with the server
    - client specifies analyzer to use during a query
    - client specifies  what fields are indexed, stored, etc.

>From a developer with prior Lucene exposure, I think Solr should be
more intuitive and straightforward.

I'd like to hear if anyone else has perspectives on this...
What are the usecases where one would want to use the Lucene Web
Service API rather than something like Solr or Nutch?

-Yonik