On Fri, Aug 09, 2002 at 10:59:46AM +1000, Martin Jericho wrote: > I'm probably going to make a few enemies here... > > Our company is actually considering moving away from Torque in favour of a > JDO compliant persistence layer. There are several to choose from, and > having a standard interface has very important advantages which I think > we're all aware of. > > We have been using standalone torque for several months without too many > problems. The only major issue is memory problems with large datasets, but > I believe that will be addressed in the next beta. If I started a new > project however, I would not use torque. BTW, Castor is also not JDO > compliant, despite the use of the acronym JDO. > > Have you evaluated any other persistence layers? I'm not aware of any > open-source JDO compliant tools, but the costs of the commercial products is > very small compared to the price of choosing the wrong product. The OJB > jakarta project looks good, although according to the website is moving > towards JDO compliance but is not quite there yet.
Thanks for the hint. I have not evaluated JDO yet, only a few Open Source Persisentce layers - and Torque was the only one with a simple central Schema - exactly what I had in mind for augmenting with SOAP info. But I will also check out JDO. > Back to your original question, I think it sounds like a good concept, but > axis is not even at the stage yet of interfacing directly into any existing > classes, let alone torque classes. The persisted classes specified by JDO > only have to implement a certain interface instead of having to extend a > class as required by torque, which would make it much easier to implement. Well, the idea would be to write my own Serializers and not use WSDL at all - and thus not WSDL2Java, Java2WSDL or even wsdl.exe. I think that should be more or less doable. Or do you know a certain show stopper? regards, Andreas -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
