Hey Michael,

this is an aside:

I used to work with EOF some time ago when WebObjects first came out (we did a 
project for USPS).  EOF was pretty primitive by todays standards but at the 
time it was an insanely advanced concept.   It was my assumption that Cayenne 
was started as a "next generation" EOF (I believe that Cayenne has far 
surpassed EOF now though).  Is this a fair characterization?

Joe



On Aug 8, 2011, at 10:19 PM, Michael Gentry wrote:

> Hi Joe,
> 
> In the past I used Cayenne with a legacy database that had 96-bit
> binary primary keys.  I had to create my own Cayenne DB adapter and PK
> generator and specify using it in Cayenne Modeler.  It wasn't too hard
> to do and worked perfectly with the existing system (which was
> actually WO/EOF-based).  No conflicts since they both used the same
> key source.
> 
> That's probably the direction you'll want to head ...
> 
> mrg
> 
> 
> On Mon, Aug 8, 2011 at 6:45 PM, Joseph Senecal <[email protected]> wrote:
>> Thanks to the help provided from this mailing list, my prototype was done on 
>> time and shows that the concept can work.
>> 
>> The next problem to solve is how to get a Cayenne program to peacefully 
>> coexist with legacy WebObjects programs that will be inserting records into 
>> the same table. I can restrict the conflict to one table, but that one table 
>> is central to all others and new records could be created by any process 
>> that loads any of the related fact tables.
>> 
>> The problem is that Cayenne and WebObjects use differently named sequences 
>> to allocate primary keys. These sequences are also formatted differently 
>> because WebObjects allocates primary keys one at a time where Cayenne 
>> allocates primary keys 20 at the time. The obvious solution is to configure 
>> the sequences for different primary key ranges. The problem is that these 
>> programs will be running at over a dozen different sites, which makes 
>> monitoring for exhausting of an assigned range problematic. My boss would 
>> find a different solution, if one is possible.
>> 
>> I'm sure that combining multiple ORM's has happened before. Does anyone have 
>> advice as to the best way to make them play nicely together?
>> 
>> Joe
>> 

Reply via email to