MUD-Dev
mailing list archive
[ Other Periods
| Other mailing lists
| Search
]
Date:
[ Previous
| Next
]
Thread:
[ Previous
| Next
]
Index:
[ Author
| Date
| Thread
]
Re: [MUD-Dev] newbie
On Tue, 4 Nov 1997, Sauron wrote:
> [snip]
> > More generally, what is the distinction between softcode and hardcode?
> [snip]
>
> abilities. The real tradeoff is w/ hardcode you increase your cpu usage,
> and softcode it chews up your ram. Most good mu*s tend to have a balance
> of both types of programmed systems.
Does this tradeoff stem from the nature of softcode/hardcode? I've
noted that the two most widely used examples (LP for softcode and DIKU for
hardcode) show this to be the case, but I had always assumed that it was
(flaws in the???) the core server designs that caused this phenomenon, and
that they were accomplished in hard/soft code to be purely coincidental.
For example, a quick glance of DIKU-clone source shows that 80-90% of the
CPU drain comes from the pulse/heartbeat/tic system where everything is
polled for continous update. (disclaimer: that isn't meant as a flame of
the source...I actually started with a gutted Circle many moons ago, but
the first thing I did was strip that system and put in an event-driven
one, which seemed to speed things up conciderably) On the other side of
the coin, I could see where an eternal language could increase the RAM
overhead (caching the language itself, as well as code-bloat from having
multiple people generating generally unoptimized code), but a well managed
system would minimize this and the bloat would be negligable.
Any thoughts?
Gunther
- Thread context:
- Re: [MUD-Dev] newbie, (continued)
- Re: [MUD-Dev] newbie,
Miroslav Silovic silovic#zesoi,fer.hr, Tue 04 Nov 1997, 10:47 GMT
- Re: [MUD-Dev] newbie,
coder coder#ibm,net, Tue 04 Nov 1997, 17:44 GMT
- Re: [MUD-Dev] newbie,
Sauron dlove#kusd,kusd.edu, Tue 04 Nov 1997, 21:47 GMT
- Re: [MUD-Dev] Re: Idea: Hive-mind monster,
Brandon J. Rickman ashes#pc4,zennet.com, Sun 02 Nov 1997, 01:47 GMT
- Re: DESIGN: ECONOMICS: Money Supply,
coder coder#ibm,net, Sat 01 Nov 1997, 23:26 GMT
[ Other Periods
| Other mailing lists
| Search
]