On Mon, Nov 3, 2008 at 1:02 PM, Tobias Marx <[EMAIL PROTECTED]> wrote: > Doesn't it use Lucene in the background?
Yes, for full-text searching, which is a major standardized feature of JCR. But please note that with JCR you typically build all your reads and writes using the API and node navigation. Queries should only be used for things where end-users actually do queries. This requires a good (hierarchical) content model, but which maps very good to the web. > What if there is many simultaneous users on a website? No problem, Jackrabbit (and Sling) with a proper content model are very RESTful in the sense that they allow easy caching by using a proxy-cache in front of the server. > How fast is it? Around 200km/h! No, sorry, but this question is too broad to be answered. Some things can be faster when compared with a database (especially the development time, since you can don't have to stick to a database design made upfront and you don't need any DAO frameworks complicating your architecture), some things might be a bit slower. And please note that it also depends on the persistence manager you use in Jackrabbit (http://wiki.apache.org/jackrabbit/PersistenceManagerFAQ) Regards, Alex -- Alexander Klimetschek [EMAIL PROTECTED]
