Comments on: We have OpenSearch, now what? http://www.givp.org/2008/05/02/we-have-opensearch-now-what/ It's a G thang Tue, 06 Jan 2009 10:55:03 +0000 http://wordpress.org/?v=2.5.1 By: Giv P http://www.givp.org/2008/05/02/we-have-opensearch-now-what/#comment-39 Giv P Thu, 30 Oct 2008 23:17:33 +0000 http://www.givp.org/?p=5#comment-39 Hi Jakob. I understand what you mean about the use of "federated" in this context and please feel free to use the material on this blog. Hi Jakob. I understand what you mean about the use of “federated” in this context and please feel free to use the material on this blog.

]]> By: Jakob http://www.givp.org/2008/05/02/we-have-opensearch-now-what/#comment-38 Jakob Thu, 30 Oct 2008 22:22:44 +0000 http://www.givp.org/?p=5#comment-38 Thanks for your inspiring experiment, can I use the image for my slides with reference to this article? The only thing that bothers me is that you call "federated search" what in fact is a meta search with additional result merging. In federated search, all particular sources have agreed on a common standard how to index documents, so you can easily combine the results without re-ranking. In your example there is no federation among the RSS-Feeds. This use of the term "federated search" always confuses people and in the end you have to explain in detail what kind of search you do - because people tend to use trendy terms without common definition. Thanks for your inspiring experiment, can I use the image for my slides with reference to this article? The only thing that bothers me is that you call “federated search” what in fact is a meta search with additional result merging. In federated search, all particular sources have agreed on a common standard how to index documents, so you can easily combine the results without re-ranking. In your example there is no federation among the RSS-Feeds. This use of the term “federated search” always confuses people and in the end you have to explain in detail what kind of search you do - because people tend to use trendy terms without common definition.

]]> By: Giv P http://www.givp.org/2008/05/02/we-have-opensearch-now-what/#comment-8 Giv P Fri, 02 May 2008 17:14:03 +0000 http://www.givp.org/?p=5#comment-8 Thanks, Nate. Yes I realise there is a scalability issue here for doing real-time indexing. In future experiments I will work on doing the indexing separately so the only thing the UI is interacting with is the index and that should be super fast. The main challenge there will be updating the index on a regular basis and speed. Thanks, Nate. Yes I realise there is a scalability issue here for doing real-time indexing. In future experiments I will work on doing the indexing separately so the only thing the UI is interacting with is the index and that should be super fast. The main challenge there will be updating the index on a regular basis and speed.

]]> By: Mike http://www.givp.org/2008/05/02/we-have-opensearch-now-what/#comment-7 Mike Fri, 02 May 2008 15:10:56 +0000 http://www.givp.org/?p=5#comment-7 Hi Giv Picked your site up via Nate's blog. I thought you might be interested in the stuff I've put together over on www.mashedmuseum.org.uk - I'm gently encouraging a "community of practice" around museum programmatic data access - lightweight, RESTful, RSS-feedy stuff. There is a very initial list of data sources there plus a Google Group and so on. tfn Mike Hi Giv

Picked your site up via Nate’s blog. I thought you might be interested in the stuff I’ve put together over on http://www.mashedmuseum.org.uk - I’m gently encouraging a “community of practice” around museum programmatic data access - lightweight, RESTful, RSS-feedy stuff. There is a very initial list of data sources there plus a Google Group and so on.

tfn

Mike

]]> By: Nate Solas http://www.givp.org/2008/05/02/we-have-opensearch-now-what/#comment-6 Nate Solas Fri, 02 May 2008 15:04:55 +0000 http://www.givp.org/?p=5#comment-6 I applaud the experiment: that's a very clever take on combining and ranking disparate result sets. I think speed would be an issue for any real-life usage, but overall this sounds similar to what Terry Makewell described for their system - although I don't think they were using Lucene? I'm in the middle of a project at the Walker that will eventually (soon!) expose an OAI-PMH repository of our collection metadata, I'll be sure to send you a link to play with once it's up. Hopefully something cool can come out of that! Great start to the blog! I applaud the experiment: that’s a very clever take on combining and ranking disparate result sets. I think speed would be an issue for any real-life usage, but overall this sounds similar to what Terry Makewell described for their system - although I don’t think they were using Lucene?

I’m in the middle of a project at the Walker that will eventually (soon!) expose an OAI-PMH repository of our collection metadata, I’ll be sure to send you a link to play with once it’s up. Hopefully something cool can come out of that!

Great start to the blog!

]]>