[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Intel: timely article



By coincidence, and without noticing it was the same article, I read Edward A Lee's article on "The Problem with Threads" yesterday evening, and add my commendation to Kevin's and Barry's. It's a little short notice, but can Lee be invited to give a paper this year?

Also in my reading pile was an article citing Herman Roebbers in Electronics Systems and Software from the IET (UK IEE as was). The article was "A new face for FPGA design" by Chris Edwards, and Herman is quoted in respect to his work at Philips using Handel-C. Herman gets in some good quotes:
"We have been able to complete projects that were thought nearly impossible using the normal way of doing things."
""The automatic generation of parallelism is never going to work. You have to design the right way."
ÂÂÂÂÂÂÂÂ and to close the article:
"Electronics engineers do think in parallel and that is what is missing in many curriculums --- how to think in parallel."
Also, a few months ago, Jonathan Bowen had an article on myths of formal methods in IEEE Computer. If I remember rightly his overall message was that "formal" formal methods had not got very far, but used as a philosophy or way of working, they can have great benefits.

And there is still an opportunity to find out more about Microsoft's HPC initiative that was acknowledged to have been based on an original paper from Denis Nicole and other members of the WoTUG community.

So the WoTUG/CPA message is getting out. None of these examples is taking occam, CSP, or whatever as the only possible solution. But they are all using the principles and applying it to things where people really do have a problem, and in a way that does not seem impossibly revolutionary. Surely the way forward is to talk to these people, to find common ground, and to build on the common experience and common ground. And, as Kuhn suggests, find even a tiny number of niches where you give real value and are not rejected, so that you can then build out from the acceptance in those niches.

Of course it won't be easy. Everyone will have their own agendas and constraints. Everyone will have to bend a little and yet not abandon the key issues. And everyone will be rejected many times before finding willing ears. But there is now real evidence that there is a problem to be solved, there is real evidence that this community has tools and understanding that can *help* to solve the problem, there is real evidence that good papers on the subject can be published in RAE ranking publications, and there is beginning to be real evidence that a business case can be built for commercial exploitation (although Barry and I and others who have tried something similar know just how hard that can be).

Paul Walker
4Links

Barry Cook wrote:
Kevin,
Â
Thanks for advertising this.
I read this article this morning - I recommend it to everyone.
Â
It's author is saying the sort of things we've been hearing at CPA for many many years. It may be time to dust off some old papers and re-work them for publication now it seems the world is more receptive to listening to them (I'm especially thinking of a couple of presentations from Peter Welch on 1. Deadlock freedom by design and 2. How program segments need to be understandable within themselves and not require understanding of other segments they interact with). IEEEÂComputer is as good a forum as any, its readership includes the people that we should be reaching and it is, I hope,Âacceptable for RAE.
Â
ÂÂÂÂÂ Barry.
Â
Dr Barry M. Cook, BSc, PhD, CEng, MBCS, CITP, MIEEE
CTO,
4Links Limited,
The Mansion,
Bletchley Park,
MK3 6ZP,
UK.
----- Original Message -----
Sent: Tuesday, June 27, 2006 12:11 PM
Subject: RE: Intel: timely article

Just to add another article to the argument, anyone with access to IEEE Computer should take a look at âThe Problem with Threadsâ in the May 2006 issue (pages 33-42). Here is someone outside the community (as far as I can tell) developing a framework that uses âCSP-like concurrencyâ. They even mention the occam word, but have no references to anything done at CPA! I would say that we are doing what the industry wants; just no one knows weâre doing it. Getting under an IEEE or ACM umbrella may not be appealing to all, but getting into the major search databases is.

Â

The name of the framework designed is Ptolemy II and can be found at http://ptolemy.berkeley.edu/ptolemyII/

Â

Kevin Chalmers

Research Student

Napier University

Edinburgh