[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Castor JDO



On Monday, September 1, 2003, at 11:50 AM, Michael Fortin wrote:

> Hibernate has a great reputation, really good docs, and a simple api.  
> We just started using for one of our projects.  You should know that 
> neither Castor (last I looked) or Hibernate fallow the JDO spec.  
> There is also OJB from apache (http://db.apache.org/ojb/), it's 
> development seemed a bit stalled when we first looked at it some 
> months ago but their site has been updated since then.
>
> Michael
>

FYI: OJB is very much active. It has been in feature freeze for a bit 
pending the 1.0 release (rc4 is out right now). Two new committers were 
in fact just added this week (full disclosure, I am one of them).

Hibernate is a very good product as well. I would *highly* recommend 
looking closely at OJB and Hibernate before using Castor. Both of them 
provide roughly an order of magnitude performance improvement on Castor 
in addition to being more flexible products.

In addition to the major architectural differences between OJB and 
Hibernate, from a user's perspective the biggest difference is the 
API's used to access the backing store.

Hibernate uses its own HQL query language, an SQL-like language derived 
from ODMG (and significantly more pleasant to use according to the 
majority of people who use it).

OJB uses ODMG and provides a lower level (though very easier to use) 
object-level (as compared to query-language level) API via the 
PersistenceBroker (http://fischer.skife.org/ojb/pb-tutorial.html is 
part of the new set of tutorials I am working on that are no on the 
main site yet) . A JDO API is available via a plugin to the JDO 
reference implementation but it is not very well documented yet (I have 
it on my todo list). A full JDO implementation is slated as the top 
priority after 1.0, and in fact the major pieces needed for it are 
already in CVS (org.apache.ojb.otm.*)

-Brian McCallister