std.parallelism: Request for Resumption of Review

Russel Winder russel at russel.org.uk
Tue Apr 12 00:12:33 PDT 2011


On Mon, 2011-04-11 at 19:01 -0400, bearophile wrote:
> dsimcha Wrote:
> 
> > This message is to formally request a resumption of the review for 
> > std.parallelism,
> 
> I like your module. I suggest to put it into a std.experimental.parallelism Phobos module in the next DMD release.

Uuurrr... excuse my ignorance of previous process, but isn't this about
voting std.parallelism into Phobos?  This is a binary question, either
it goes in as std.parallelism or it doesn't.  Putting things in as one
package and then moving it seems like a sys admin and developer
nightmare.

Personally I see it as a "no contest" get it in and get it in now.  But
perhaps this will be taken as campaigning.

Moral from C++0x, Java/JSR166, Go:  get parallelism stuff into the
language early, at the risk of having to issue bug fix releases.  C++0x
is only just getting threads after 30 years.  Fortunately asynchronous
functional call made it is, as did futures, but sadly not closures.
Java has had threads since its arrival on the general scene 16 years
ago, but now people know this is infrastructure not stuff for
applications programmers.  JSR166 was intended to do something about
this, but some bits got into Java 5, some into Java 6, some will go into
Java 7 but only in Java 8 will the really good stuff get in.  Goodness
knows which version of closures will go in when.  This is turning into a
15 year program to get things into a language that should have been in
from 10 years after the language hit the streets.  Go put its model of
parallelism in from the outset.  OK there have been a few bugs that had
to be fixed, but they took the research work of 30 years in other
languages and put it straight in.  D has some concurrency tools and a
form of closure (albeit a bit awkward to create them), but it needs more
parallelism tools now before the ships of C++ and Go sail off with the
native code developer community.

C++0x with TBB and stuff that Anthony Williams is creating as part of
Just::Thread Pro to do actors and dataflow, also C++CSP2 if you want
CSP, could easily lead to a resurgence of C++ despite the mess that is
its templates stuff.

(People who still think C is a great applications development language
are beyond saving ;-) 

So to answer the question directly: putting this package in as
std.experimental.parallelism would be a way of halting progress of D
take up.

-- 
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:russel.winder at ekiga.net
41 Buckmaster Road    m: +44 7770 465 077   xmpp: russel at russel.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://lists.puremagic.com/pipermail/digitalmars-d/attachments/20110412/6bf68601/attachment.pgp>


More information about the Digitalmars-d mailing list