i411 is a faceted metadata search and browse engine, capable of scaling to very large deployments, such as the DexOnline yellow pages site, which uses it for both search results and browse navigation. The most recent version adds a web crawler to the local file and database connectors, a natural language module that can extract entities from queries and provide concept-based spellcheck, more flexibility in the search flow, and a SiteOptimizer analytics and reporting module to expose site dynamics and user behavior.(Disclaimer: I consulted with DexOnline and helped them choose the engine among a very strong field of candidates.)

Info Today Report: “Enterprise Search: Deployment, Usage and Trends”
A survey of 250 professionals connected to search in their enterprise has some enlightening results. They were a fairly wide variety of industries, organization sizes, departments and roles (described in detail in the report), so the results are generally applicable. This survey contradicts conventional wisdom by reporting that 62% of these enterprises have more than one search engine, with a 27% of having four or more search engines. In my view, this indicates the understanding that one search cannot solve all problems, and that some areas will require specialized, and usually more powerful, search solutions.The other response which surprised me was that 20% of respondents said they already provide search for audio and video, and 35% said they want to do so in the future. I suppose some of that is podcasts and training videos, and it’s a big challenge for search, although much easier if there are transcripts or textual captions.The report also covers integration with other applications (mainly CMS and KM), current search solutions, vendor support satisfaction, software vs. hosted search vs. appliance (only 17% reported using a search appliance), upgrade plans, and search features currently available and desired for the future. There’s a long section about the respondents’ relative emphasis on various criteria for selecting a search solution, covering ease of use, features, integration, cost, scalability, speed, vendor reputation, ease of installation, upgradability, and vendor support.This report is available on the Enterprise Search Center, at a cost of $495 US. The study was conducted by Shore Communications and Faulkner Information Services.

2007 Search Conferences list posted
The listing of search and search-related conferences is up on the Searchtools Conferences page. To suggest a conference for this page, please leave a comment here and we’ll add it.

Solr Open Source Enterprise Search and Faceted Metadata Server - new Searchtools Report
The Solr Java open-source search engine builds on the Lucene engine, adding more standard tools for indexing, query processing and sending back results. While Solr does not have a site indexing crawler, it can use Nutch or any other robot crawler, and accept content converted from any native format to a simple XML schema. The architecture provides powerful tools for analyzing and transforming text to create a very rich index with structured fields. It accepts a wide variety of query operators, and parameters to control retrieval and ranking, including sorting on specified fields. The default relevance ranking can be tuned to suit the needs of the users and content, and the search results provide the valuable match terms in context for each item. In addition to lists of results, Solr has faceted metadata displays dynamically calculated for search results, allowing users to drill down on topics, date ranges, price, brand or any other attribute. For scaling to millions of documents and high search traffic, the system offers caching configuration, index replication, autowarming for new starts. It’s installed on such high-traffic sites as CNET, shopper.com, and The Internet Archive, showing its scalability, and has active developer and user mailing lists.

This entry was posted on Wednesday, April 23rd, 2008 at 5:13 am and is filed under Tools. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.