On Fri, Dec 12, 2008 at 10:19 PM, Ryan McKinley <[email protected]> wrote:
> I have not looked into either DIH or Solr Cell, so forgive me if this is > way off base. I understand the desire to have these baked into the server > -- it simplifies things dramatically for anyone wanting to use them. > However is it possible to write them so they could be easily extracted to > the client? > > (IMHO) ideally these could be written against the solrj API and then work > either with the EmbeddedSolrServer or a remote one. For example the core of > Solr Cell and perhaps DIH would be great to plug directly into droids. > Look at SOLR-853 which aims to make DIH available as an API to be used from SolrJ or plain old Lucene itself. There can be other uses too. > > If that seems reasonable, are there things we should consider about the > existing API before it gets baked into 1.4? I feel that we should try to get 1.4 out of the door and attempt this in the next iteration. We have enough features for 1.4 which are needed by most users (think replication, faceting improvement, cell). Let us not delay 1.4 for things that a smaller section of users would need. Thoughts? -- Regards, Shalin Shekhar Mangar.
