Wanted: Speakers for London.pm Technical Meeting
benjamin.john.evans at gmail.com
Tue Oct 16 18:14:09 BST 2012
On Tue, Oct 16, 2012 at 5:27 PM, Dominic Thoreau
<dominic at thoreau-online.net> wrote:
> On 16 October 2012 17:06, Pedro Figueiredo <me at pedrofigueiredo.org> wrote:
>> and another one I've been mulling and considering writing to submit to LPW, but I'm afraid it wouldn't warrant me many friends in the L.pm community, it's titled "The problem with Perl" and is basically me ranting about the nice things in Java and Python that we (as in, myself until I started doing serious Java and Python, and some other L.pm'ers I've talked with) just tend to shrug off or don't even think about, or think they're bad but they're actually not bad at all.
> After the course I took at the OU last year (
> https://msds.open.ac.uk/students/study/undergraduate/course/m362.htm )
> If someone can bring the synchronised keyword into perl, that would be
> good. And make it work across platforms (like, load balanced servers)
> that would be good too.
You do not want synchronized and "visibility semantics pretty close to
the JMM" as your primary concurrency mechanism visible at the
high-level language level, or where anyone short of a Damian-level
hacker can easily get at it.
Sure, that may well be what your virtual machine (or equivalent
hardware abstraction layer) wants, but don't, for the love of Mike,
expose it in the HLL. That's pretty much the biggest lesson learned in
the last 15 years of implementation decisions in the JVM.
More information about the london.pm