Re: [tlug] STM (was: Re: work times & accommodation @tokyo)

[Curt Sampson <cjs@example.com>]

On 2008-07-28 17:49 +0100 (Mon), Josh Glover wrote:

> Unless I am very badly mistaken, Java 7's f/j framework does the
> locking for you.

There may be a wee bit of locking as threads wait for subthreads,
but fork/join does not deal with any but the most trivial of locking
problems. It doesn't come anywhere dealing with efficient access to data
structures shared amongst multiple threads.

(The same is true of "par", of course; this is why this forking for
parallelism and and the locking issues are two completely separate
things.)

By the way, thinking about the issue more, I'm going to go out on a
(apparently rather strong) limb and posit that Java simply cannot have
an STM system like Haskells, where you combine separate operations and
those combinations are checked for locking errors at compile time. I
just don't see any way that the language could support it. I'm willing
to be proven wrong, of course.

cjs
-- 
Curt Sampson       <cjs@example.com>        +81 90 7737 2974   
Mobile sites and software consulting: http://www.starling-software.com