[MUD-Dev] Re: Introduction System

[birgit.schulte at philips.com]

--<cut>--
Note: This message was written via the list web archives.  There is
no guarantee that the claimed author is actually the author.
--<cut>--
Original message: http://www.kanga.nu/archives/MUD-Dev-L/2000Q3/msg00191.php

On Mon, 17 Jul 2000 10:24:55 -0700
J C Lawrence <claw at kanga.nu> wrote:

> > --<cut>-- 
> > Note: This message was written via the list web archives.  There
> > is no guarantee that the claimed author is actually the author.
> > --<cut>-- 
> 
> Yipee!

My pleasure ;)
I really like this feature, only I feel a bit discriminated -
I mean, having stated above ones message that it might be faked
leaves a strange feeling. So what about having the possibility
of identifying oneself via the password which is needed for the
membership anyway?


> This could get interesting:
> 
>   Instead of using a 1-100 range, use a much larger value set, or
>   even a collection of value sets.   Say 4 bits per value set (16
>   values) with 8 value sets in a two-word integer.  If you want more
>   types or larger sets, just increase the widths.  Heck, just use an
>   8bit char per value set -- makes the arithmetic faster and
>   simpler.
> 
>   Now use those value sets to map basic characteristics.  Say, all
>   the basic body types/species are enumerated in one value set, all
>   the hair colours in another, the skin tones in a third, the ages
>   in a fourth, etc.  The idea is not to make an exact mapping, but
>   arrive at integer values which loosely *represents* that
>   character's current appearance.
> 
>   Now, upon meeting an arbitrary character, scan the recorded
>   character value sets and compare them to the value set before the
>   player. If there's a match within N%, ouput a message ala:
> 
>     There's an oddly familiar dwarf here.
> 
>   You can then plug in various nearness percentages, history/age
>   factors (long time since you've seen Bubba == harder to recognise
>   him for even identical value sets) etc.

Yep, and I'd put in a "confidence"-value as well. Like, if someone else
described Boffo to me, whom I've never seen before, I'd get an impression
of him with a confidence-value according to the quality of the description.
(Might even work when looking at a painting - wanted-list, here I come!)

> 
> This also gives an initial handle on approaching disguises in a
> semi-intelligent manner as it allows easy support of both
> intentional and unintentional disguise factors.
> 
>   Bubba has a terrible fright and his hair goes white.  He is now
>   commonly mistaken for Boffo whose hair was always white.

I really like this approach, as it's close to what I was brooding about
for an intro-system so far :) Only problem I see is, that it would 
require one comparison for every value-set, and since lookup in a 
namespace would be quite frequent, this could become costly.

So I was pondering if those value-sets couldn't be put together
(let's assume the two word integer) in a way that those characteristics
which are hardest to change (e.g. race, gender) make for the most significant
bits, with the others following in order. Then I'd build a nested mapping
(talking lpc here), with those most significant bits building the keys
for the first level, and the LSBs at the next level. This would narrow the
range I need to compare to from the beginning.

On the other hand, this would pre-determine the range in which the matching
of the value has to occur (somewhere on the LSBs). Dunno yet if this is a good 
or a bad thing...

Birgit



_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev