Selected Newsgroup Message
From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Mentifex musings on the comp.ai vote
Date: 28 Apr 1999 00:00:00 GMT
Message-ID: <3727354c@news3.us.ibm.net>
References: <37261136.0@news.victoria.tc.ca>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 28 Apr 1999 16:20:28 GMT, 166.72.21.171
Organization: Intelliwise Research and Training
Newsgroups: comp.ai
Arthur T. Murray wrote in message <37261136.0@news.victoria.tc.ca>...
>[snip]
>In comp.ai various people are mentioning Arthur T. Murray by name
>as one of the posters they are thinking about as they deliberate
>whether or not to moderate the newsgroup. They are considering
>moderation, while I am considering a complete withdrawal from
>posting to Usenet. (Wow, I am so addicted to Internet posting
>that I even read back over what I have written thus far and I
>imagine whether or not it would make a good Usenet post.)
>[snip]
>Why can't people in comp.ai see the sincerity of my AI diagrams
>and of my lifelong effort to create
>http://www.geocities.com/Athens/Agora/7256/acm.html public domain
>AI?
Hello, Arthur,
I have been touched by the content of your last message.
If you can accept a humble opinion about your question, then read
on. First, let me tell you that I admire your persistence. I can
say that this is something that I believe you have in common with
the guys from the CYC project (which is no small thing to have!).
Unfortunately, this seems to be insufficient to succeed.
Your main concern appears to be the kind of reception that your
ideas historically have received in this and other fora. I'll
tell you that this kind of frustration is more common than you
may think. There's a lot of good texts being put forward by
respectable names out there which receive similar indifference,
even when published in respectable journals.
So, what's the problem?
The problem is that there are a lot of ideas (that's good!) but
only a few are really meaningful or promising. To judge which ones
are and which aren't, every reader (we!) establishes a set of
"internal criteria". One of these is the clarity of the idea. Other
is the support of that idea on knowingly good previous work. Other
is proposing a different and innovative solution. Another is how
good the text is in "luring" the reader. And even other is to
propose rock-solid thoughts and the way they're all tied together.
Let me analyze each one of these aspects.
Clarity.
It is tempting to use jargon or words with a different meaning.
In fact, the deeper we are, the more we're tempted to use special
wording. But doing so without explaining to the reader what they
mean only serves to the purpose of diminishing the consideration
one's idea is given. It is not a good method to "enrich" the text
with this kind of wording, although it is easy to find some
scientific papers that choose this way. My advice is to start any
text with a textual justification of the new words (which are they,
why they are being introduced instead of more common ones). If you
manage to write a text where the "meaning" of that new word becomes
clear and self-evident, then you've reached perfection.
Support.
We are lucky for being alive in these times. We have a delightful
and tremendously complex problem to solve: that of discovering
how to embed some intelligence in our stupid machines. But this
problem is very, very complex. It cannot be solved by a single
man's wisdom. No, I'm not saying that it can't be solved by a
single man, I'm saying that it can't be solved only by the
*ideas* of a single man. Advancing in AI *must* use the ideas
and techniques developed by previous thinkers. We all know
that, but if one text doesn't leave this *clear*, then the text
is not considered seriously. Note, however, that it is not
enough to say that we're using someone else's ideas. You must
refer to them explicitly, repeat its main conclusions and point
your reader to the relevance of the citation, almost as if you
were telling he/she for the first time. If you let clear to your
reader that you've grasped the essence of that cited author,
even if the reader already knew about him, you will be adding
credibility to your entire text.
Innovation
Repeating what has been done in the past is a good guarantee of
obtaining the same results. This is the essential principle to be
followed by any assembly line of any industrial company. But this
*is not* good for science or innovative research. Innovation means
sometimes going *against* what was proposed before. That means we
sometimes must *defy* what others taught us. But to do that, one
must be absolutely clear *why* we are doing that. Let your goal
stand up and show that previous ideas didn't achieve it. Show,
then, that your idea is an innovative option to get that result.
Luring Readers
I see mostly two ways of obtaining success with new ideas.
You use your name (which means, you're a established and
respected researcher, so anything you write will obtain
immediate attention) or you use a text that *lures* your
reader. Want an example? Say you're writing about the
eternal fight between the symbolicist and connectionist
guys. An article that starts with "Simple Recursive
Networks fail to account for rule learning" could generate
some interest, both in the ones which like SRNs as to
the ones who dislikes. But you've got to complete that
title with an article that really addresses your claims.
That leads us to the next item.
Rock-Solid
The more you solidify what you write, the better.
Often, what is necessary is writing the article as if you
were an *oppositor* of your idea. Write how "stupid" you
were in proposing that thing. Then, write what are the
thoughts that rebut *your own* opposition (showing that you
were right in your initial hypothesis). If you do that with
enough new (and solid) evidences or with the *support* from
the work of acclaimed authors, you'll be solidifying not only
your ideas, but also your name. Nobody will judge you by what
you are, but by your ideas and their "likeliness". If you've
pondered a bit as an oppositor, you pass the idea of
impartiality.
Tying all together
Even considering all previous suggestions, one idea may seem
difficult to swallow if it is composed of loose items. There's
nothing like "coercing" your reader to accept the next item
based on the previous paragraphs. Obviously, your final destiny
is to reach the conclusion you want. It is much like those
Agatha Christie stories: you're given snippets of information
that goes on assembling one narrative that compels the reader to
discover the "killer".
I hope that you accept these words as my way to express what
you could do to improve the exposition of your ideas.
You should not care being called a "weirdo". I know quite a
bunch of scientists that had a similar nickname in their times,
but that are today on the most important pages of the history of
science.
Regards,
Sergio Navega.
Back to Menu of Messages Sergio's Homepage
Any Comments? snavega@attglobal.net