[Home] [Groups] - Message: [Prev in Group] [Next in Group]

nu.kanga.list.mud-dev

7822: [MUD-Dev] Re: [Java] multithreading: update and a question

[Full Header] [Plain Text]
From: Vadim Tkachenko <vt@freehold.crocodile.org>
Newsgroups: nu.kanga.list.mud-dev
Date: Wed, 15 Jul 1998 23:39:41 -0500
References: [1]
Organization: Kanga.Nu
Chris Gray wrote:
> 
> [Vadim Tkachenko:]

[skipped]
 
> If I'm not remembering wrong, Java does allow you to find out the machine
> and OS you are running on.

System.getProperty( X ), where X is os.name, os.arch, os.version

> Is that how you plan to do things in your wrappers? That should work out
> OK, but as you say, it sure isn't the portable Java programming that we
> are supposed to have!

Unfortunately, it's not portable as a lot of us want it to. On the
bright side, this particular issue has something to do with performance
optimization, not with the portability, although I admit it may
influence the very ability to operate under given conditions.

Yes, this is how I'd implement it, and I guess I don't even have to
reinvent the wheel - the concepts of HAL (aka Hardware Abstraction
Layer), microkernel etc. existed for a long time.

And again, in this particular case, the many/one is hopeless, one/one
may be casted into many/many with approximately the same mechanism as
many/many uses at the kernel level. So the resulting performance
degradation wouldn't be so bad.

Or, just get the right OS from the very start, but that's religious.

> Chris Gray     cg@ami-cg.GraySage.Edmonton.AB.CA

--
Still alive and smile stays on,
Vadim Tkachenko <vt@freehold.crocodile.org>
--
UNIX _is_ user friendly, he's just very picky about who his friends are