[RC5] v3, etc.

Joe Sunday void at null.net
Sun Jan 18 22:48:32 EST 1998

Peter Bortas wrote:

> Joe Sunday <void at null.net> writes:
> <SNIP>

> core. Also, you'll need separate cores for each project. However,
> > writing the first core in Ansi-C (Or any other standard language
> > available on multiple platforms) would allow cores to be compiled
> > for all platforms from one source, with optimized assembly routines
> > for each platform following, as the C version likely won't be as
> > fast as a hand-tuned assembly version designed to take advantage of
> > certian features on the host machine.
> It would be nice if there was a bytecode-core so that more obscure
> platforms like Acorn, Amiga or whatever didn't have to wait for new
> cores. Also, a security manager wouldn't hurt, and thats much easier

If you write one core in C, you don't need to make a machine independent 
code, just compile the one file with a compiler on every platform. Should 
possible to have cores for everyone out within a day or so, at least 
ones. They'll be slower than hand tuned assebly, but you'll have them out 
away, and they'll still be faster than java or any other byte-code like
implementation where a translation is required at runtime on every 
Machines for which coders are available can have special tuned versions 
as time
goes on. As for a security manager, I assume that someone at d.net will 
go over
the core source, and we can limit the system calls to those in the 
standard C
library, which prevents someone from looking at your registry(Win95), 
etc. (In
fact, this would have to be anyway in order to ensure portability), and 
limit use
of file operators and such. A malicious core would be spotted before it 
made it
to us, I would hope. I sincerely hope d.net will not distribute cores for 
the source hasn't been reviewed.


To unsubcribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest

More information about the rc5 mailing list