Re: putting bonobo into an IDL

[Michael Meeks <michael ximian com>]

Hi Peter,

On Wed, 2003-04-16 at 14:02, Peter Van Osta wrote:
> With ORBit and Bonobo is be feasible to turn our high-throughput
> microscopy reader into a real distributed image acquisition and analysis
> system. I already ahd the CORBA part readyfor use, but I was missing
> Bonob to complete the system.

	Heh ;-) well, I don't know quite what you're so happy about; but there
are several things that may make libbonobo unsuitable for your needs,
let me list them just in case you've not met them yet.

	a) Not Threaded - all requests are processed in the same thread,
	   re-enterantly - not always a massive problem ;-)

	b) bonobo-activation provides no distributed support - so you get
	   to solve the 'first reference' problem yourself without help
	   ( in practice if you have a shared NFS server or somesuch,
	     just writing an IOR to a common place is usually sufficient)

> The system design makes it possible to run several automated microscope
> readers (with sample loading robot) attached to a high-throughput
> backbone or robotics system and in parallel analyse the images on either
> the same computer or on powerful servers.

	Yes; sounds like it would allow that.

> We have been doing high volume automated microscopy for more than ten
> years now (on Silicon Graphics workstations and servers), but now with
> ORBit and Bonobo the system will run smoothly and distributed on Linux
> systems. I hope to add the sytem to the succesful examples of
> "industrial scale" applications, based on ORBit and Bonobo.

	Heh ;-) please let me know immediately you have problems, I'm very
interested in helping make this work for you.

	Regards,

		Michael.

-- 
 mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot