On 09.09.2015, at 09:19, Flavel Heyman <[email protected]> wrote: > 2. You really really want a folder hierarchy. > (I say #2 as .5 because I believe tagging fills this need, so who cares > about folders?)
That is what a content repository is all about, compared to an RDBMS. This paper might give some more insights [1]. Content repositories are great for CMS etc. where end users want to control the hierarchies (page, navigation and asset/document folder structures). > The other things Jackrabbit brings to the table like versioning, locking, > restrictions > can be implemented in very little time from scratch. In fact if you are > implementing all 3 it takes less time > to do it from scratch than it would to do it in Jackrabbit because the > order in which you do the above things together matters. That sounds a bit brave to me :) > Obviously you have to use Jackrabbit Oak to get ³real" performance, going > with plain Jackrabbit will leave out a lot of the > the indexing performance enhancements. Oak is basically version 3 of Jackrabbit, but was a complete rewrite for better performance and flexibility, hence the different name. Jackrabbit 2 is in maintenance mode (Oak also does not implement all JCR parts and leaves out some rarely used features in favor of performance, so Jackrabbit 2 still is the JCR reference implementation). There is no "plain Jackrabbit". Anyone new to JCR/Jackrabbit should use Oak and existing Jackrabbit users should look into upgrading, if possible. [1] http://dev.day.com/content/ddc/blog/2009/01/jcrrdbmsreport/_jcr_content/images/jcrrdbmsreport/jcr_rdbms_report_chapuis.pdf Cheers, Alex
