PDD 25 "Concurrency" - first review

[Allison Randal]

I've finished a first pass through PDD 25 on threading/concurrency. It's
largely a collection of prior thinking on the subject. Before I start
kicking it into a more structured form, I'd like to do an initial round
of discussion. This is your chance to mention anything you hoped or
expected from Parrot's concurrency models. How do you plan to use
concurrency, and in what contexts? What's your favorite concurrency
model and why should we consider using it? How integral a role should
the new STM play in Parrot's concurrency? Etc.

I've changed the name of the PDD from "Threads" to "Concurrency"
because: a) the notion of "threads" seems to have taken on mythical
proportions, so this is a symbolic step toward practicality, and b)
Parrot has more than one concurrency model, and this PDD will be the
overview of all them, plus the ways the various models interact.

Allison

[Joshua Gatcomb]

On 10/30/06, Allison Randal <all...@perl.org> wrote:
>
> Before I start
> kicking it into a more structured form, I'd like to do an initial round
> of discussion. This is your chance to mention anything you hoped or
> expected from Parrot's concurrency models. How do you plan to use
> concurrency, and in what contexts? What's your favorite concurrency
> model and why should we consider using it? How integral a role should
> the new STM play in Parrot's concurrency? Etc.

There is a recent thread over at PerlMonks on this very topic:

"Parrot, threads & fears for the future"
http://perlmonks.org/?node_id=580004

The thread is long and goes on tangents at points. Not every one has kept
an
objective perspective and it has its share of finger pointing doom and
gloom.

All negativity aside, there are a great deal of valuable points made in the
thread
which make it worth reading. While it has lost a bit of momentum - it is
still
being updated with new posts almost daily.

Allison

Cheers,
Joshua Gatcomb
a.k.a. Limbic~Region

[Allison Randal]

Joshua Gatcomb wrote:
>
> All negativity aside, there are a great deal of valuable points made in the
> thread
> which make it worth reading. While it has lost a bit of momentum - it is
> still
> being updated with new posts almost daily.

Excellent. Thanks for the pointer, Joshua. Definitely some good ideas
mixed in there.

Hmmm... high on my wishlist: sandboxing for non-concurrent code running
inside concurrent code.

Oh, the Io language, which I've been interested in lately, also makes
use of the concept of "futures" for concurrency. It's got a degree of
appeal to it.

Allison

[Larry Wall]

On Mon, Oct 30, 2006 at 10:28:51PM -0800, Allison Randal wrote:
: Oh, the Io language, which I've been interested in lately, also makes

: use of the concept of "futures" for concurrency. It's got a degree of
: appeal to it.

Perl 6's feeds (lazy lists) can also be viewed as a form of futures,
insofar as they tell the compiler that you don't care much about the
ordering of side effects between the producer and the consumer. They
just set up the return pipeline without necessarily demanding a return
value immediately.

Larry