[MUD-Dev] Re: MUD-Dev digest, Vol 1 #231 - 4 msgs

[J C Lawrence]

<<I've CC'ed this to Meta as some parts better pertain there>>

On Fri, 10 Nov 2000 12:25:27 -0600 (CST) 
Cat  <cat at realtime.net> wrote:

> Given that reviewing before posting doesn't scale...

I can attest to this on both sides.  On the one side it doesn't
scale across mutltiple venues.  Judging by my experience with
MUD-Dev and a couple other forums, the limit is about 3 forums per
human, with the concomittant problems of restraining burnout and
sustained continual interest (and before anybody gets worried, no,
I've not lost interest in MUDs or MUD-Dev.  I have however had,
"Aww, heck, I just don't want to deal with today" days, and I
suspect I'll have more over the next few score years) Next it
doesn't scale by volume let alone internal topic range.  There are
days as a moderator that I have a tough time keeping up with
traffic, let alone being able to contribute usefully to it. <sigh>.

That said, human scalability, and in particular the moderation
demands of *distributed* systems in particular are an area of
endless fascination.  Oddly enough I'm interviewing with a startup
on Monday that may indirectly relate to this area, tho more from a
highly tailored marketing/promotion aspect: think of advertising
tailored to target groups so finely defined their populations are
numbered in the low single digits, and the concomittant problems
with retaining a universally referencible advertiser-happy public
image) which delivering and maintaining such small target marketing
relationships.

Scary.

The automation problems in signal moderation are scary (let alone
the question of signal definition) and are not unique to
discussionary forums.  When you really get down to it, the every
popluar PK, stamp colletector, black rose incident et al problems on
this list are really just problems in contextual signal definition.
Peer review ala Advogato have no smaller peerage problems than
MUD-Dev has, and do suffer from glory mechanics.  

I've been putting some thought into this for the Wiki-nature of the
new Kanga.Nu site I keep promising.  I'm beginning to think that
Google might have the right idea, both at the Wiki level and the
mailing list levels (think replies and graph structure).

As a crude point, consider a graph of MUD-Dev posting patterns, with
nodes representing posts, and links representing reply
relationships.  A few obvious abstracts become interesting:

  What characterising posts that receive many replies?

  What characterises branches (threads) that receive many posts?

  What characterises both the parent node and threads of unusually
  large posts?

  What characterises the posts toward the end of threads, and are
  there patterns in the graphs which can be determined and which
  seem to predict thread terminations?

Now take that lot, and start comparing it to Wiki node
inter-relations (and you start needing a degree in topological
analysis) and I think we may have something.  No promises, but it
smells like the right direction.

If somone would like to do some of the above statistical and pattern
analysis...  Perhaps one of you chaps still in college?  

> I think the only practical approach is "post everything instantly,
> review content only when there's a complaint, and if it's
> determined to violate your policies, take it down".  Falling into
> the "review some material" category, this is clearly less costly
> than "review all material".  And if not one single person has
> bothered to complain about some particular page, how big a problem
> is it really?

While this model is effective in many areas, it does not work in
others (eg MUD-Dev).  cf Usenet, community expectations, character.
Of course, this is also ego bolstering.  I do need to keep myself in
a job.

> Mind, I think I could manage to make the world a lot more
> interesting with just a few million.  I know how to use money more
> efficiently than a lot of the maniacs out there.

Ye gads.  You really shouldn't give me straight lines like that.

> Allow users to post all content, and review only when there's a
> complaint.  That's a correct 21st century information age kind of
> approach.  

ie:

  Allow players to do anything they want, to define the game any way
  they want, and then only step in and question upon complaint, with
  the defintion of "rights" and "rules" being contextually defined
  by some abstraction of the playerbase active in that venue
  (largest population group, most vocal population, most articulate
  population, group most in allignment with reviewer's personal
  views, group most in allignment with original intent of the
  game/area, group that most piques someone's fancy/interest, other)

Not only do you get to build in the game, but you get to build and
define the base rules and mechanics of the game by your own
politiking.

> Refinements include mechanisms to allow users to have areas where
> they volunteer to vet content for each other, for those that
> prefer cleaner or "safer" content.  (What you're safer FROM, and
> in more danger of, depends on who is "sanitizing" the content and
> with what agenda.)  Further refinements to the refinements consist
> of allowing multiple users (anyone that wants to, in the extreme
> case) to layer on top of a given content area their own pre- or
> post- moderation info, with each user able to choose which
> moderators they want to "subscribe" to, or none at all for an
> unfiltered feed.  

There is a wonderful book by Fred Saberhagen that I've discussed
here prevously called "The Veils of Azlaroc".  In it there is a
rather odd planet which is periodically subjected to what is called
"veilfall".  Loosely, anything/everything subjected to veilfall
becomes partially removed from the surrounding non-veilfalled
universe, with the effects being cumulative with increasing
veilfalls.  The details are well drawn and interesting -- I
recommend the book unreservedly.  

The model also has some parallels for MUDs.  Consider a game where
each player has the opportunity to explicitly and implicitly state
their relationships to various other players, groups, and
behaviours.  As they do so, and their role within the game world
becomes increasingly defined their perceptions of the game is also
altered, and the mechanics by which the game (at a mechanical level)
are tailored to that individual player to produce a customised
worldview.

This could be as simple as slowly abstracting the PK'ers and
monty-haulers awawy from the socialisers.  Yes, they'll still
inhabit the same world, but increasingly as members of each side
continue to play the members of the other populations will be
presented in less detailed fashions by the game, perhaps becoming
ghostlike, or in fact never being mentioned at all unless
specifically looked for.

Consider:

  Bubba walks into a market square.  

  It is crowded with people.

  Were Bubba a merchant he would be inindated with details on the
  economic activity occuring in the square, the various other
  merchants, those he recognised, and details of their merchanting
  techniques.

  Were Bubba a soldier he'd spot every guard, bodyguard and soldier
  in the square, and have some concept of their capabilities.  The
  merchants would be mostly background chatter.  "There are many
  merchants here."

  Were Bubba a socialiser he'd spot the other socialisers like a
  flash, as they would him.  They are like creatures.

  Were Bubba a PK'er he'd spot the other PK'ers like a flash.  The
  socialisers however would be mostly if not entirely invisible
  unless he directly looked at/for one.
 
> I watch Verant commit virtual homicide upon one of their own
> customers for writing a story on someone else's turf, and I laugh
> like a mad hyena in the night.  

The sad thing is that complexity theory applies to cooporations just 
as well as it does problems and their solutions.  

--
J C Lawrence                                 Home: claw at kanga.nu
---------(*)                               Other: coder at kanga.nu
http://www.kanga.nu/~claw/        Keys etc: finger claw at kanga.nu
--=| A man is as sane as he is dangerous to his environment |=--


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