Given the votes expressed so far by other jakarta commiters - I can only vote +1.
This seems like a usefull tool ( I spent few years working on OODBs and OQL ), with a modular design. My only comment: I would be much happier with a wider top-level project, where the various db-related technologies could find the expertise and community that is needed ( instead of ending up with 3-4 smaller projects ). We do not have ( AFAIK ) any top level project with the primary goal of doing database-related stuff. In time ( and if we are lucky ) this project can incorporate various other db-related components - like the various connection pools, portions of torque, etc. Of course, the internal organization is your choice, but I think it would be good if various database related proposals and technologies share the same top level porject You have my +1 on the proposal anyway. Costin On 29 Apr 2002, Jason van Zyl wrote: > Hi, > > I would like to propose ObjectRelationalBridge > (http://objectbridge.sourceforge.net/) as a top level subproject of > Jakarta. > > For those not familiar with ObjectBridge it is arguably one of the most > advanced persistence layers available, commercial or otherwise. It is > accompanied by an extensive, current documentation set which includes a > quick start guide, tutorials, a FAQ, design documentation describing how > certain features of OJB have been implemented, and deployment guides. > > The developer community is incredibly strong and currently consists of > 17 inviduals: three of whom are Jakarta committers, and one of the core > Castor developers. So the project has the numbers and has displayed some > collaboration with other projects. There are developers from the Torque > team (the simple table->object persistence tool within the turbine > subproject) too so there is obvious interest in OJB. The current list of > developers can be found here: > > http://sourceforge.net/project/memberlist.php?group_id=13647 > > I would also like to note that David Taylor, a Jetspeed fellow, also > contributed to the internal transaction mechanism. So again, another > point of interest within Jakarta. > > OJB is currently being used in the Jetspeed project, and integration is > well underway in the Turbine project and Thomas Mahler, the author of > OJB, uses OJB in conjunction with Struts as part of some of the > solutions his company provides for clients. Thomas is also a user of > TopLink, which is the only product that is even remotely comparable with > OJB, so he is very familiar with both and reports that OJB is on par > with TopLink with to respect to performance and available features. > > I won't go into a complete list of features, but here are some of the > features that set OJB apart: > > o Pluggable APIs: Currently there is the native PersistenceBroker API, a > full ODMG API (which provides enhanced transaction isolation) and a JDO > implementation is in the works. OJB has been designed to allow different > front-end APIs for maximum flexibility. The ODMG API, for example, is a > small set of classes layered over top the core of OJB. The JDO > implementation will be very similiar in nature. > > o Pluggable query APIs: currently supported are a criteria based API > (AST based mechansism), OQL and SODA. But again they are pluggable, so > for example the query mechanism in Torque could easily be made to work > with OJB. > > These two features alone make OJB attractive as different APIs can be > made so that existing users of different systems can use OJB without > forcing clients to change code. Trying this with Torque is going to be > one of my first exercises to see how well this mechanism works. There > are many tools like Torque and OJB can be made to work with the APIs of > these projects so that greater collaboration can occur within OJB > itself. One can take a look at the source and design of OJB and quickly > determine that OJB stands in a class of its own, is very reliable, very > flexible and very performant. > > The greatest feature with respect to development is the extensive > regression testing features and the testbed. There are currently 130+ > test cases and a regression test that compares the performance of OJB > with native JDBC calls. > > A full list of features can be found here: > > http://objectbridge.sourceforge.net/features.html > > OJB also makes use of many Jakarta packages: Ant, Maven, Crimson, and > Log4j. There are also plans to use more of the commons utilities where > possible so the project is already Jakarta friendly :-) > > Another interesting note is that OJB is one of the top 100 projects on > SourceForge (rank 89) with about 15,000 hits and 3,500 downloads per > month. So there is a very healthy user community that complements the > strong developer community. > > Currently the license of OJB is LGPL but in discussion with Thomas he > feels that a BSD style license like Apache's is actually a better model > and has no problem with changing the license if the donation of OJB is > accepted by the Jakarta PMC. > > This is really a one-of-a-kind project, and is definitely one of the > cases where an OSS implementation is close, if not better than its > commercial counterpart. The developer community is keen, there are great > number of users and we think that OJB would be a fabulous addition to > the set of projects that are currently housed at Jakarta. > > -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
