On Wed, 14 Feb 2007 15:23:57 +0100, "Wolf Siberski" <[EMAIL PROTECTED]> said:
> With respect to the general evolution of Lucene, IMHO we have > at least two types of users, with different needs: > a) Lucene experts which use it as low level building block > within their highly customized information retrieval systems. > b) Lucene non-experts which use it as out-of-the-box component > to just add full-text search to some application. > IMHO the API evolution should keep both types of users in mind, > instead of just focusing on type a) (as you seem to suggest). > Besides, Lucene users typically start in b), because it is > so easy to get it up and running, and then gradually move > towards a) (if and when required). > Therefore, it helps the Lucene community as a whole to remove > API obstacles for newcomers. There is a great quote from Alan Kay that applies here and applies to most software development: "Simple things should be simple. Complex things should be possible." I find myself relying on this very often when making design tradeoffs (and elsewhere!). Most of our users (category b above) will use Lucene for the simplest of use-cases. For these users we should strive to have Lucene be incredibly simple / intuitive to use. Then some of our users (category a above), I think the shrinking minority (with time), will have advanced / expert use cases. For this group we should clearly make possible all of these advanced use cases, but, not to the detriment of keeping simple use cases simple, when possible. On this particular issue I think we have achieved exactly that: made Lucene simpler for simple use cases, while not affecting the complex use cases. Mike --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
