Hi all, I had a look at the code on the web page (and got it all to compile with some minor hacks :-) and it looks like the idl compiler generates c stubs (although I didn't get this working on my own idl - I couldn't find the right script to use). I was wondering if this was the intended overall approach? - I was thinking more along the lines of developing a dynamic marshalling interface to the orbit core and then having the idl compiler generate the skeleton code in python that talks to that (i.e. no C compiler involved). BTW, There was a discussion on the orbit list a while ago about this and if I remember right, the general conclusion was that writing a custom dynamic interface was a better idea than attempting to implement dynany in orbit, mainly because of its poor performance and difficulty of implementation. What's the plan folks? Cheers, Phil.