Hi, On 5/11/07, J. Delgado <[EMAIL PROTECTED]> wrote:
I would like once again to open this can of worms, and perhaps think out of the box, without classifying DB and Full-Text as simply different, as we analyze concepts to further understand the real path for evolution of Lucene/Sorl
Another data point to consider are hierarchical content repositories like Jackrabbit and the myriad of custom "database + lucene" repositories out there. I had some interesting discussions during the last ApacheCon about the potentially converging storage and and feature requirements of various components along the DB/Full text index axis. At least from the Jackrabbit perspective it would be interesting to look at how to better integrate the Lucene search index with our native persistence model. I guess people from DB projects like Derby have similar interests. Looking further in the future it might even make sense to completely unify the storage model of these various projects. I.e. have set handling from Derby, hierarchies from Jackrabbit, and indexing from Lucene in a single extensible storage engine that could be used as a unified backend layer by various projects. I believe that many of the current storage data structures need to be redesigned in any case due to current trends in computing. The driving trends I see are increasingly cheap storage and the switch to more parallelism in computing. I believe that within the next 5-10 years we'll see many projects switching to append-only data structures that focus on massively parallel read operations with zero locking. I guess I'm sufficiently out of the box now... I'll crawl back in. :-) BR, Jukka Zitting
