Over the last week, I've completely rewritten the KBCafe search engine. I found that everybody was using the default search and that I wasn't getting any results on the alternate search engines. I think the reason was that the number of search engines was simply overwhelming. In place, I grouped all the searches into 9 buckets, which can each be searched independently.
- Blogs (searches BlogPulse, blogWatcher, Feedster and IceRocket)
- Jobs (searches Feedster, Hotjobs, Indeed and Simply Hired)
- Links (searches BlogPulse, Feedster, IceRocket and MSN)
- News (searches IceRocket, MSN and Yahoo!)
- Podcasts (searches AllPodcasts.com)
- Shopping (searches Amazon)
- Souce Code (searches Koders)
- Tags (searches del.icio.us, Flickr, Technorati and Yahoo!)
- Web (searches MSN)
The search results are grouped together for each search type. This is more inline with traditional meta-search engines. Hope you like!
lazytom, I want to make the other Web results available, but I haven't decided how to do that yet. Splicing two feeds that are sorted by the same value makes sense. Since most of the feeds are sorted by time, splicing them is easy. The Web feeds are sorted by relevance and there's no common relevance value that can be used to splice the feeds. I hope that makes sense. Do you have a suggestion on how they could be spliced?
Randy