Re: [9fans] EROS, Vapour

[David Gordon Hogan <dhog@plan9.bell-labs.com>]

[-- Attachment #1: Type: text/plain, Size: 22 bytes --]

Inferno does this.


[-- Attachment #2: Type: message/rfc822, Size: 2620 bytes --]

From: Eyal Lotem <eyal@hyperroll.com>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] EROS, Vapour
Date: Fri, 31 Aug 2001 08:45:21 GMT
Message-ID: <3b8ea46f@news.bezeqint.net>

Boyd Roberts wrote:

>> I think all the OS projects that use C++ figure that one out sooner or
>> later ...
> 
> i fear that C++ coder's brains are so damaged by the experience that
> they are _incapable_ of figuring it out.  snow crash, indeed.
Well, EROS is moving to C, and Vapour is as far from C\C++ as you can 
imagine.
The whole *point* behind Vapour is that it uses a safe language, rather 
than damage-prevention hardware protection, which allows a lot of the tests 
to move to compiletime, and allows interesting things, such as guaranteeing 
running code complies with object interfaces, without verifying this in 
runtime (Sometimes it could be unverifiable, or very expensive to verify).
A serious advantage of Vapour for example, is since it doesn't reuqire 
hardware protection, it can run many threads simulataniously, all in kernel 
mode, all synchronizing via cheap test&set processor instructions, whereas 
in Plan9/etc. an expensive switch to kernel mode is required, making 
heavily parallel things unpractical (the cost of mutexes becomes higher 
than the gain of parallelism).