Re: Non-Blocking I/O

[korpela(at)albert.ssl.berkeley.edu (Eric J. Korpela)]

In article <7a8f40$qsj$1@nnrp1.dejanews.com>,
 <menakkis@my-dejanews.com> wrote:
>Ruediger Kupper <Ruediger.Kupper@Physik.Uni-Marburg.De> wrote:
>
><...>
>> Okay, so there seems to be no way around using some intermediary C-Routines
>> which handle reception.
>> IDL just doesn't support Inter Process Communication...
>
>Frankly I'm surprised that this interesting thread didn't tease out more
>responses.  I'm sure that there must be *dozens* of lurkers out there who
>have been down this road.

Well, having the thread occur on a weekend probably doesn't help the 
response volume.  I have been down this road before, and basically
have come to the conclusion that if IDL doesn't have what you want, adding
it, while not exactly trivial, is pretty damn easy.

Now no one is beating down the door to get to my web site and get VARRAY
(which will add some shared memory support to IDL).  No one has expressed
much interest in asking me to speed up publication of a multiprocessing 
library for IDL (see examples given on this group a couple months ago.)  
I've been using standard UNIX IPC mechanisms in IDL for a couple years now,
but haven't taken the effort to make my software available.  Nothing makes a
4 processor machine run like using all the processors.

Maybe the demand doesn't exist?  If the demand did exist someone like David
Fanning probably would have done it by now.

Eric
-- 
Eric Korpela                        |  An object at rest can never be
korpela@ssl.berkeley.edu            |  stopped.
<a href="http://sag-www.ssl.berkeley.edu/~korpela";>Click for home page.</a>