Lists Home |
Date Index |
>Do people write program to compose protocol messages in XML
>(probably DTD as well), send to the other end, and then the receiver
>uses an XML parser to get a DOM structure from the message, and
>access the message data through the DOM structure?
It can be easier.
Open a TCP stream between the two processes and attach SAX parsers to
them. Have the two notify their peer what kinds of messages they are
going to be sending:
Each message is simply an element:
And when you're done sending, close the "document":
Then close the stream. You'll get element events from the parser.
If you're happier with pull parsers, you can use them as well.
>This is kind of big overhead and people who want to use XML as their message
>encapsulation need to understand XML first.
They need to know very little about XML. Layers of structures and
tools and schemas make it overly complicated.