Full text search in codeBeamer has served well during the years. It was flexible, it was stable, but it was a little inconvenient to use, result lists were hard to browse and sometimes it was hard to find information in a blink of an eye. In the upcoming codeBeamer version 5.4, search is going under a major redesign.
Improvements revolve around the followings:
- More relevant hits. Regarding search hits, ultimately nothing else, but relevance matters. Especially when you have large data sets, and some of our customers do have. Our "ranking formula" is now being reviewed and tuned to deliver better results.
- Query syntax enhancements. Our query syntax will be extended to support search terms like "created_at:20090805" (matches any entity created at 2009-08-05) and "owner_id:123" (matches any content which is owned by that particular user).
- Aggregated search result list. Until now, codeBeamer delivered result lists that were separated by entity type. That means document type hits were not mixed with wiki type hits. This was useful when you actually know beforehand what entity type you are looking for, but it made browsing the results very inconvenient otherwise. And if you knew the type, you filtered by that any way, right?
From now, codeBeamer returns a single unified Google-like result list, with titles, interwiki links, excerpts. This single list includes every type of hits, and is sorted by relevance. - "Did you mean?" We have a prototype implementation of the search suggestion thing. There is good chance that it makes to 5.4. Question: would this be important for you?
Interesting? If you want to receive similar quick updates about our products, follow us on Twitter: http://twitter.com/intland
