|start | find | index | login or register | edit|
by earl, 6814 days agosektion: 3 minuten moderne literatur - vielleicht neuzugruenden, hier aber zuallererst einfach eine leseempfehlung
woohoo! they finally got it ;) [und ich schreibe endlich mein xml-r/xml-m piece wie in 2001-03-18 versprochen]
So if we do asynchronous RPC, we're doing "messaging"?
this is one thing that was always so darn obvious to me ;) that's how jabber relates to XML-RPC; ... blah. but i could not have given a better answer than dave does in his question.
And I suspect that at a coding level, unless there's tons of replicated code, that there's a procedural interface to their message passing. If so, the distinction is non-existent, it's RPC, just adding another step to the marshalling process (which on today's machines probably is completely insignifcant performance wise).
quite a bit wrong; maybe it doesn't matter if the procedure call ist handled locally, which then of course would be LPC - but even in that situation the machine could use the time between call and response for doing something else.
but as soon as the procedure call is marshalled over a network - and thats the point of the R in RPC - there's a huge difference. instant messaging would get in huge troubles if the sender would have to wait until the message is delivered - IM doesn't wait; that's why it's called messaging - it's asynchronous.
David McCusker explains this in even more detail - and technically more precise
And now we'll begin implementing some of jabber's functionality with xmlrpc's marshalling format. but i have to tell you, i really like this.
|earl.strain.at • esa3 • online for 6834 days • c'est un vanilla site|