RE: [Objectbridge-developers] XML used in test.ojb.server, etc

[David F. <dw...@la...>]

At 10:29 AM 1/9/2001 +0100, Mahler Thomas wrote:
>>This got it running, but I'm interested in this example because it is the 
>>closest to what we are doing.
>>We have been running with poet and understand the ODMG approach.  I 
>>really need to have some kind
>>of querying ability with approximate matches with wild-card supported.
>>
>>Thanks,
>>
>>Dave
>>
>Hi Dave,
>
>We aim at providing a full ODMG implementation, thus we will have a full 
>OQL implementation.
>Until now we not startet working on OQL. currently David Dixon-Peugh is 
>working on Object Locking and Transaction management.
>I'm currently working on reworking the ojb.examples.server.App as a JUnit 
>test.
>One of my next goals will be the OQL implementation. I'm planning to reuse 
>the OQL implementation from CASTOR (castor.exolab.org).
>
>AFAIK Castor does not implement ODMG but JDO. I don't know if you are 
>familiar with JDO: it's the upcoming Standard API for transparent 
>persistence in Java. When you look at the samples on the castor site you 
>will notice that it has a lot in common with ODMG.

I have acquaintance with Castor and JDO, but Castor's JDO is not the same 
as Sun's proposed JDO.  Actually what I need (want?) is not necessarily 
full ODMG, but simply the ability to create a query from a Java object.  We 
are using that capability in storedObjects already with great success.  It 
could be converted to OQL at any time, but probably is a lot more efficient 
than OQL because one is dealing directly with the objects, rather than 
having to parse them.  In a sense it is a little lower than OQL.   I 
believe it is a lot easier to implement this than a full OQL parser 
handler.  It might be the first step in the process.  I think others have 
taken this approach, too.

In our application on sourceforge (OpenEMed), we have a persistent 
abstraction layer that is implemented over a number of different kinds of 
persistent store (ODBMS, and RDBMS).  The RDBMS part uses Software Tree's 
JDX product that works very well.  We would like to use an open source 
product, if at all possible, however.  Castor is a likely candidate, but 
still is evolving rapidly.

Thanks,

Dave

>
>regards,
>
>Thomas
>
>

David W. Forslund                               dw...@la...
Computer and Computational Sciences             http://www.acl.lanl.gov/~dwf
Los Alamos National Laboratory          Los Alamos, NM 87545
505-665-1907                                    FAX: 505-665-4939