Nascif Abousalh-Neto writes: > Hi Paul, > I think that in the case of Transmogrify we didn't have much choice. > We wanted to interface with an engine and an interface that was already > defined, so we could not change the control flow to use queues and even > handlers. Maybe an adaptation layer could do the trick, I don't know. I
That's what the MATLAB-Emacs interface is. It's an adaptation layer between two systems that were developed completely independently of each other. > didn't have enough elisp knowledge to create and manipulate sockets from the > lisp side, so gnuclient sounded like a good compromise. > It's pretty simple. There's an example in the JDEbug code. > Do you think it would be possible to expose the mechanism you > implemented for MATLAB as an open, documented, JDE extension API? It sounds > like you have solved the communication portion of the problem already, while > Nick's JUCI takes care of hiding the elisp learning curve from Java > programmers. Combine both and you have an open plug-in interface for JDE! > Yes but it's too trivial to implement as something separate from Nick's JUCI. It should be included in it. - Paul
