> > what does it mean " to create a message-oriented system"? > I have the same problem or the same idea, to use my own xml shema whitin > apache soap. > What I understand Apache SOAP is that it offers 2 approaches, one is RPC-based, RPC stands for Remote Procedure Call; the other is message-oriented.( If I did not copy and paste from the documentations wrongly :-) )
Followed the philosophy of technology, one is easy to learn and use, but loses fully control, the high-level API, that is RPC-based; the other needs you spend time to come down close to *dirty* low-level API and provides the functions you can control anything it allows. That is message-oriented. Data on the way via SOAP is in XML format, obviously. The problem is who is mainly in-charge to create and parse the XML format, as well as define the schema. Developers or Apache SOAP? Using RPC-based, you do not need to worry about how to create and parse XML and even you do not need to know what envelop is. Apache SOAP has implemented a set of serializers and deserializers. You just need to use them directly or create your own seriliazers and deserialziers for your classes transferred. It is black box on how Apache SOAP help you assembly from your class to XML and disassembly from XML to your class, unless you read the source code of Apache SOAP. :-) If choosing to use message-oriented services, you are responsible to fill in SOAP envelop or body yourself. Yes, I think you can use your XML schema freely there, but beware to add your own namespace :-) In my case, we are trying to make Perl talk to Java in bio-informatics field. We have specific XML schema for bio- data. I try to find the solution to transfer huge data, you know the DNA right? ;) , via SOAP. However, when I tried , using RPC-based, to transfer a java.util.Vector with 20,000 elements, I encountered java.lang.InternalException. The object is so big for JVM/HotSpot. For RPC-based, there is another *shortcoming*, that it sends an object after the object is generated completely. Message-oriented is more like a stream. Do you have any suggestion on my project? Thanks in advance. Please correct me, if you found wrong. Thanks. Juguang ----------------------------------------------- Xiao Juguang Fugu Informatics Group (Fugu B) Institute of Molecular and Cell Biology National University of Singapore 30 Medical Drive Singapore 117609 DID: +65 874 1413 [EMAIL PROTECTED] [EMAIL PROTECTED] http://www.imcb.nus.edu.sg http://www.fugu-sg.org -----------------------------------------------