Users may find themselves overwhelmed when they have to choose between attachments, properties and named objects when deciding where to put the payload. However, the decision can be simplified:
The developer defines the contract that the clients will use in order to interact with their service. As part of that contract, both functional and non-functional aspects of the service will be specified; for example, that it is an airline reservation service (functional) and that it is transactional in nature (non-functional).
The developer will also define the operations (messages) that the service can understand. The format (such as Java Serialized Message or XML) is defined as part of the message definition. (In the example case, they will be the transaction context, seat number, customer name and so forth.) When you define the content, you can specify where in the message the service can find the payload. (This can be in the form of attachments or specific named objects, or even the default named object if one so wishes.) It is entirely up to the service developer to determine. The only restriction is that objects and attachments must have a globally-unique name, otherwise one service or action may inadvertently pick up a partial payload meant for another (if the same message body is being forwarded along on multiple "hops").
Users can obtain the service's contract definition (either through either the UDDI registry or via an out-of-band communication) which defines where in the message the payload must be placed. Information put in other locations will almost certainly be ignored, resulting in the incorrect operation of the service.