cg at ami-cg.GraySage.Edmonton.AB.CA
Mon Apr 7 21:23:22 New Zealand Standard Time 1997
:>I believe that as much as possible of Chris L's MUD code will execute
:Nope. I bytecode with conditional runtime on-demand compilation. The
:bytecode is then timestamped (as is the source) and stored with the parent
:object to short-circuit future compilations (unless the source or
:inheritacen tree changes).
Ok, terminology differences. I think of bytecodes as interpreted. Perhaps
'simulated' is a better word. 'Emulated'?
Perhaps we can come up with a gradient (I'm sure this has been done before!),
and we can all point to where we are on it:
1. native machine code
2. threaded code
4. parse tree traversal
5. pre-tokenized interpretation
6. straight text interpretation
You're at 3? I'm at 4 in AmigaMUD, 6 in ToyMud. However, differences
like the need for run-time symbol lookup, or type conversions, can easily
move something up a notch in execution overhead. LP is at 3 now isn't
it? Are MOO and ColdX at 3 as well? Let's see, REXX is 5? Tcl is 6.
Shell scripts are 6+, Perl is 5? This is just *so* relevant to MUDs! :-/
Chris Gray cg at ami-cg.GraySage.Edmonton.AB.CA
More information about the MUD-Dev