
Hi Chimera developers, I was wondering if it would be possible to publish the Collaboratory IDL files? I would like to make a remote control application for chimera, but don't want to use the puppet extension. It seems like a nice way to do this would be to use the Collaboratory interface. Thanks, Dave

On Wed, 1 Feb 2006, David E. Konerding wrote:
I was wondering if it would be possible to publish the Collaboratory IDL files? I would like to make a remote control application for chimera, but don't want to use the puppet extension. It seems like a nice way to do this would be to use the Collaboratory interface.
That's an intriguing suggestion. We consider the Collaboratory IDL an private internal interface and highly subject to change (e.g., from release to release and we're thinking about switching to a non-CORBA mechanism to minimize firewall issues). So while it would be possible to control chimera via the CORBA interface, it would be a lot of work for something that is likely to break. The puppet extension is the prefered remote-control mechanism. It is publicly exposed via the --send command line argument. And you can send files to a running chimera, including Python .py files and chimera command .cmd files. If chimera isn't running, it will start up with the given files (although not in the background yet). So what is the real issue? Do you want a continuous remote interface to chimera? Is the overhead of the --send option a problem? Please tell us. There are probably alternative solutions that are better. - Greg

So what is the real issue? Do you want a continuous remote interface to chimera? Is the overhead of the --send option a problem? Please tell us. There are probably alternative solutions that are better.
Hi Greg, Yeah, the continuous remote interface, with explicit ability to access molecular objects, along with notifications of changes to the objects, are what I want. We have plenty of experience with building web services; I would suggest the following alternative (the problem being mainly a matter of increased developer time compared to the CORBA approach): convert the CORBA IDL files to XML schema, use our ZSI package to generate the Python bindings, then build a 'collaboratory hub grid service' (I say grid service instead of web service since the basic web services specs left out notification and messaging, but . All the concepts in CORBA map more or less directly to web services, so conversion of the collaboratory implementation code would be pretty straightforward. Dave

On Wed, 1 Feb 2006, David E. Konerding wrote:
Greg Couch wrote:
So what is the real issue? Do you want a continuous remote interface to chimera? Is the overhead of the --send option a problem? Please tell us. There are probably alternative solutions that are better.
Hi Greg,
Yeah, the continuous remote interface, with explicit ability to access molecular objects, along with notifications of changes to the objects, are what I want.
We have plenty of experience with building web services; I would suggest the following alternative (the problem being mainly a matter of increased developer time compared to the CORBA approach): convert the CORBA IDL files to XML schema, use our ZSI package to generate the Python bindings, then build a 'collaboratory hub grid service' (I say grid service instead of web service since the basic web services specs left out notification and messaging, but . All the concepts in CORBA map more or less directly to web services, so conversion of the collaboratory implementation code would be pretty straightforward.
So first of all, we would be happy to provide you with the CORBA IDL files so you could write the code yourself :-). However, our current preferred method would be for you to write an extension that supported the subset of the chimera functionality that your application needs. You, as the original collaboratory author, are uniquely qualified to do so. A UCSF maintained solution would be better, but we're not ready to standardize that part of chimera yet. - Greg
participants (2)
-
David E. Konerding
-
Greg Couch