Is facetted searching really useful for general purpose text searching?
Im am interesting in the possibilities of facetted sea开发者_Python百科rching using Lucene and perhaps Bobo but have a few q regarding just how practical it is for the user if they are just searching text and not data that has been broken up into many fields each which could be the target of a facet and tallying.
Whether faceting is useful depends a lot on your use cases -- though if done right, it tends to be very user friendly.
Indeed, as you point out, the quality of your facets is dependent on the quality of your data. If you have no way of classifying any of your Lucene documents in any way, then faceting is likely not for you.
However, if you can achieve even a simple classification of your data -- such as its source system, creation date, language, or size -- then that could certainly help users. Again, what to facet on depends on your use cases. Useful facets doesn't have to be based on field values that exist in your content source, they can be populated when processing data before indexing it.
If you have a lot of time for development on your hands (right), making use of classification systems such as Balie, OpenNLP, or similar, could provide good faceting options based on your user domain.
If your application allows it, an alternative could be to introduce Folksonomies, letting user tags be faceting options.
精彩评论