[MUD-Dev] Unix vs NT and its relation to MUD environments.

[J C Lawrence]

If the following isn't entirely coherent I apologise.  I'm fat with a
very heavy meal and the cranium keeps chanting "Sleep!" as versus
"Type good message!"  

This message is also prompted by a study I took part in today on the
human usability and ergonomic concerns of the design and operation of
plugs on system cables for large systems (eg SCSI, cluster
inter-connect, etc).  It got me thinking about interface _function_
(as versus design) for MUDs.

  What is it that a MUD interface is actually trying to DO?

  How is it DOING that?

  Is it DOING that well?

  Could that <thing> be done better or differently?

In the case of the study I found that I liked several of the proposed
designs for very different and almost contradictory reasons.  Each
suggested different useful, functional and valuable concerns to me as
a person who has assembled many large systems and clusters, but most
interesting was that the concerns and the way the designs approached
them in two cases precluded even the existance let alone the handling
of other concerns (ie QQQ does XXX well, but YYY poorly, but RRR does
ZZZ excellantly and in a manner that XXX doesn't even exist in the
problem space and YYY is done fairly well.  ZZZ doesn't exist in the
problem space for QQQ.  (I hope that's understandable (think
horizontal versus vertical insertion, insertion versus removal
mechanics, cable guides and cable interference with plug
insertion/removal, directional sensitivity, packing density and
manipulation ease for horizontally and vertically parallel socket
arrays, rotational orientation awareness of plug, ease of twisting the
plug and attached cable to an orientation and manipulating it there
with accuracy against stress, etc)).

ISD magazine held a discussion with various EDA vendor reps and a
large audience of EDA users on the question of "Linux vs NT".  The
transcript is rather fun reading:

  URL:http://www.isdmag.com/dac/linuxtext.html

However one of the more interesting sections occurs near the end:

--<cut>--

  Dan Small (Microsoft): Yeah but the scripting is almost the failing
    of UNIX not a virtue. [talkover] (Audience laughter)

  ...deletia...

  Audience 12: The statement you made was very characteristically
    Microsoft arrogance. You said basically that you don?t need
    scripting, and scripting allows you, scripting says, "Okay, we
    don?t know everything that you?re going to do as a user, you can
    go off and put these things together."

  Dan Small (Microsoft): Scripting means the tools may be inflexible
    enough to adopt to every situation that they use them [talkover]

--<cut>--

Dan's last comment quoted above is both insightful and arrogant as
claimed.  It states both that the tool vendor knows (everything) that
the user needs, and that the user should/will never need more than a
good/well designed tool supplies.

The problem is that neither is true, not compleatly.

Scripting does do a lot to make up for the weakness and feature lack
of tools.  Tools will also never be the one-size-fits-all solution
that vendors would like to cream about.

But, to return this to the MUD world and the perrenial question of MUD 
user interfaces, the same question and debate applies to the questions 
of interface configuration, and to MUD world programmability.

There are, if you consider the extremes only, two base approaches:

  1) The world shall define everything that happens and how the user
can make it happen, and that shall be the limit of the world.

  2) The world shall define how everything happens, suggests how the
user can make it happen, the user gets to pick and choose as to which
suggestions he accepts and which he creates on his own, and the limits
of the world shall be bound by what the user can create within the
constraints of the world mechanics.

cf Aber or Shades vs MOO or MUSH (I'm deliberately ignoring IRC which
removes all central enforced definition of the world characteristics).

The happy medium is presumably somewhere in the middle with the
specific point in the middle defined by the answer to two questions:

  1) How much can the user customise his interface to the game?

  2) How much can the user define the game world or its mechanics?

I am a big fan of user programmable MUD worlds.  I'm also a big fan of
goal-oriented worlds.  To date these two world modelss have been
considered mutually contradictory due to security concerns.  I intend
to reverse that assumption, but getting back to the topic:

  In our brave new world of clueless AOL users and rampaging
anti-social GoP'ers, can we safely say that an interface not close to
the configurable end of #1 and #2 is going to be viable?

  Non-programmable/scriptable MUDs assume that the MUD offers
everything that the user can/will want (non-scripting).  To a certain
extent this is epitomised by the hard-coded MUD's very large list of
socials as versus the ultimate flexibility of MUSH'es emote commands.
One size fits nobody.

  The heavy usage of the scripting features of TinTin, TinyFugue etc
with their scripting abilities on hard-coded text MUD's suggests that
users both want and demand scripting (I presume ZMud offers something
in this area).  The fact that the same client configurations extend to
soft-coded servers is merely a large side-benefit (cf LP's alias
command and others).

  Then again, judging by Usenet traffic alone, outside of simple
triggers, the apparent use of client-side complex scripting is
becoming quite rare (it used to be almost assumable).  Most new MUD
users seem scared of complex scripting clients -- they like the idea,
but only use it at the level of using un-examined macros others have
written.

  UOL (and M59 I assume tho I haven't noticed direct mention of it on
M59 user pages) has an extensive "problem" with users scripting their
actions both with the limited scripting supports within the UOL
client, and via client-external automation tools (simulates KB and
mouse inputs).  A common phrasing of which taken from somewhere on the
UOL pages is that making an action have a very low probability of
success is equal to making its success rate 100% when a user can just
do a:

    > do 10000 times: pick lock

  Its either black or white -- no grays.

  These people are scripting.  They are automating.  They are reducing
their world to customised tokenised self-adapting function blocks
which can be invoked as singular units, or chained and assembled in
variously complex structures.  They are also doing extremely simple
simulus-response systems which as a whole regularly exhibit complex
behaviours.  (cf Lambert's comments and assertions on visual scripting
environments).

  Script configurations tend to be widely and heavily traded.  Scripts
change the world and the world view in very incisive and pervasive
manners.  Having to maintain a stock of food in your inventory and eat
when you get hungry can easy be a key and defining game mehcanic --
until you have a script that ensures that you always have a lot of
food, and that you are always fed.  Ditto for the rest of the world.

How different would UOL, M59, and all the rest *really* be if any
possibility of scripting were utterly and absolutely removed?  Would
they have achieved the level of commercial success they have were
scripting *NOT* possible?  Why (not)?

How different would they be if they offered extensive scripting, say
to the level of a full micro-language with flow control, variables,
and all the rest?  (Note the difference between demand and offer --
I'm not saying that users would be required to use it, just that it
would be available and that they could drop/plug-in scripts others had 
written).  Would this have affected their success?  In what areas
would it help?  Hinder?  Why?

In interface design, when is one-size-fits-all a Good Thing?  A Bad
Thing?

Okay, I'm outta here.  Gonna go find a dead lower life form, eat some
of its slowly rotting carcass (yeah, I'm hungry again), and head back
to the humble abode.

--
J C Lawrence                               Internet: claw at null.net
(Contractor)                               Internet: coder at ibm.net
---------(*)                     Internet: claw at under.engr.sgi.com
...Honourary Member of Clan McFud -- Teamer's Avenging Monolith...