From: Robert Feldt <feldt@...>
Date: 2000-10-28T01:01:57+09:00
Subject: [ruby-talk:5922] Re: [RFC] Towards a new synchronisation primitive

On Sat, 28 Oct 2000, Conrad Schneiker/Austin/Contr/IBM wrote:

> Are there any good solutions in existing languages that can be copied 
> (with suitable Ruby adjustments)? 
> 
Very probably... ;-)

> Does anyone know if the approaches that Python, Perl, Tcl, Lisp/CLOS, 
> etc., take (whatever they are) are any good for our purposes?
> 
Might be but mind you that combining concurrency and OOP is still a pretty
hot/unsolved research issue (see for example a recent thesis giving a good
overview of the topic and proposes a solution 
http://citeseer.nj.nec.com/papers/cs/16468/http:%23@S@%23%23@S@%23www.mri.mq.edu.au%23@S@%23people%23@S@%23.%23@S@%23dholmes%23@S@%23thesis-submission.pdf
(The solution proposed might be the way to go to get the most flexible but
also most expensive Ruby object synch model: syncronixation rings). So I'd
guess that how to combine/include concurrency with a dynamic, pure,
interpreted OOP language is not common knowledge. ;-)

Conclusion from this is probably that we should supply one simple solution
for now and then supply more complex/flexible/costly solutions in
extensions.

If anyone knows what Python, Perl et al does on
concurrency/synchronization then please write a short description and give
some pointers.

Regards,

Robert