[alicebot-archcomm] Re: [alicebot-general] Why do committees exist?

[Conan Callen]

If a thing does not have any defined limits then it is in danger of
becomming a blob.

This is an interesting thread that brought up several questions in my mind
(marketing kinds of questions) Its pretty rough, as I typed up my thoughts
it turned out looking more like a business plan development exercise or a
marketing brainstorm session ;)

In a typical corporation the definition comes from the
marketing department who is constantly probing the mind of their target
consumer in an effort to develope a product that will beat the
competition. If a feature will not help to sell more widgets or is too
expensive it typically gets whacked. Richard has done a great job to keep
the features focused, but sometimes I cannot make out who he is trying to
target. Have you ever worked up a description like this Richard, is this
vision something that you would care to share? Are there any initives
from the financial backers? (initives are real popular in the press
lately)?

Because this is still a rather new concept / market there is not a
whole lot of previous experience to go on. Alice has a strong goal of
being minimal & simple, but a lot of people are asking for more flexible /
advanced features. What are the metrics to help determine which features
make it and which get cut?

What is the purpose of Alice - The start of an
industry? A commercial product? A spec upon which commercial products can be created? To be a learning tool to
bring programmers up to speed? Provide a starting point for more advanced
tools? Who is the target market?

Jon mentioned that he has presented many business plans involving
Alice / AIML, why didnt they bite? What do you think they felt was
missing or they were looking for? Undoubtably business people would be
looking for an opertunity to experience a return on their investments.
Were you able to present business opertunities with supporting financials
showing a real return and they still passed?

Conan

> Our committee process has been criticized for everything from holding
> back progress on AIML
> to being thought police and enforcing mind-control policies on the
> AIML community.  These criticisms
> are particularly painful to me personally, because I can hear myself
> making the same case years
> ago when I worked as a professor.  Those damn committees were holding me
> back, stopping me
> from doing my research, and conspiring against me at every turn.  Now
> I find myself on the other
> side of the equation.
>
> The technology all around us is very new, but human beings have been the
> same for a long time.
> Over centuries, people built up a system of rules of order ("Robert's
> Rules of Order", or parliamentary
> procedure) to manage conflict.  You can choose to follow those rules or
> ignore them.  But if you
> ignore them, conflict is almost inevitable.  Conflict usually manifests
> itself to us as "flame wars"
> on our mailing lists.  It could mean physical violence in the real
> world.
>
> No one is more of a "rugged individualist" than me.  No one likes to
> "color outside the lines" or
> "think outside the box" than I do.  I always say, people don't go into
> computer science because they
> enjoy working with the public.  I would be happy to sit behind a
> computer screen all day and insist
> that no one bother me.   Only very reluctantly did I come to the
> conclusion that a project like ALICE
> would require the cooperation of many others.   If you decide to work
> with others, you really have
> only one choice of managing conflict, and that is by committee.  If you
> know any other way, please let us know.
>
> Part of working in a group is realizing that others may have radically
> different points of view.
> The committee procedure allows people to voice those diverse opinions,
> in a structured way that
> limits the chance of "flame wars".  If you are convinced that "my way is
> the right way", and you
> can't persuade the majority of your peers that it is, then it probably
> is not.
>
> Finally, on the subject of learning, Kim is quite correct that there is
> no easy solution.  There is
> always a need for a teacher or supervisor to provide "quality control".
> Even if you created a bot
> automatically by scanning human chat logs, you would still need to edit
> them for quality.  Targeting
> is the most effecient method yet devised for supervised learning in chat
> bots.
>
> Rich
>
> _______________________________________________
> alicebot-general mailing list
> alicebot-general@list.alicebot.org
> UNSUBSCRIBE FROM HERE: http://list.alicebot.org/mailman/listinfo.cgi/alicebot-general
>