Some notes on writing, learning, teaching, noticing, designing, and
threatening people with legal action, as well as technical language
and intellectual property, plus a recommendation and some URL's.

As a periodic reminder, you can find copies of most past RRE messages
in the Web-based archives that are maintained by a couple of private
companies.  Links to these archives can be found on the RRE Web page:

This URL is mentioned in the -=-=-=-=-='ed "tag" at the top of almost
every RRE message.  I do not keep copies of past RRE messages myself.

One of the hazards of running a list like this, it would seem,
is occasional threats of legal action by people who imagine that
they have some grievance against me.  The other day, for example,
I received the following message out of nowhere from someone who
had been subscribed to the list for several years.

    Get me the hell off your mailing list or I will press charges
    of harassment. I am a Information Technology lawyer and I will
    prosecute to the fullest. Get My Drift?

C'est quoi?  I don't know.  I've asked him, and I will be interested
to see what he says.  Given the peculiarities of mailer software, it
is conceivable that he has some legitimate complaint, even if he needs
to learn some a more constructive way of expressing it.  Not that
I can complain, given that I'm not always polite myself with people
who hassle me.  In any case, if I am informed that this guy has filed
criminal charges against me, you'll be the first to know.  The major
problem with threats like this is that, rationally or not, they cause
me to lose sleep.  I need my sleep, and on several occasions I have
given thought to abandoning the list altogether in order to make
sure I get enough.  Come to think of it, maybe I can improve my sleep
by starting my legal defense fund now.

This is the problem with the Internet: it provides an instantaneous
channel for all of the sweetness and light that we could possibly
wish to convey to one another, but it also provides an instantaneous
channel for all of the emotional poison we wish to dump on one another
as well.  Is it really worth it on balance?

Recommended: Richard Saul Wurman, Information Architects, edited by
Peter Bradford, Graphis Press, 1996.  Richard Saul Wurman is a graphic
designer known for his innovative reference works, most particularly
the outstanding Access Guides already recommended on this list, and
for his networking between the technology and design worlds.  This
volume, planned by Wurman and edited by another important information
designer, Peter Bradford, gathers images and texts from a couple dozen
others who pioneer the craft of using graphic design to communicate
information.  It's great, and is best (and easily) read from cover
to cover in one sitting.  The designer's creed of rethinking things
from scratch announces itself from the first moment you open the book:
almost all of it employs white type on black pages.  (My cynical self,
noting the book's unusually stern warning against reproduction, argued
that this was a deliberate choice to frustrate photocopying.  The fact
remains, though, that the black background works quite well to frame
the images.)

The contributors describe projects that include the weather page
of USA Today, museum displays, expository graphics in magazines,
and instruction manuals.  It was depressing how many of the projects
were cancelled because of their unconventionality, and surprising how
cheerful the designers presented themselves as being about it.  One
example, by Bradford, was a reinvented children's dictionary -- surely
a crying need -- that the publishers cancelled because they didn't
know what to call it and therefore didn't know how to advertise it.

But two things struck me most forcefully.  One was the overwhelming
sense that the people who used these materials have bodies: they walk
into a room, they open a box with a computer in it, they are on the
road in a strange town, or whatever, and the information artifact is
supposed to fit both physically and cognitively into that situation.
Being graphic designers, these folks know a great deal about human
eyes -- what it's like to look at things, and what colors are for, and
the value of a consistent set of conventions for everything, and the
fine line between naturalistic representation and abstract schematic.
They know something about hands, too, though they hardly mention
them.  They have a strong sense of the vastness of their design space,
and they take seriously the iterative process of exploring it.

The designers' approach contrasts with the emphasis on disembodiment
in the cyberspace ideology, which I think is exactly wrong.  Science
fiction disserves us here: William Gibson's heroes are basically
shamans, but we are not designing systems for shamans but for normal
people.  (Besides, real shamans have day jobs too.)  The ideology
of disembodiment in prevailing cultural constructions of cyberspace,
not to mention the tacit presupposition of disembodiment in computer
system design, is an artifact of the clunky interface devices that
are available to us today.  Once we are unchained from big, cumbersome
monitors, we'll be able to compute in the settings of our choice,
and we'll choose our settings because of the things we want to do
with our bodies: sit in the back yard, travel somewhere, meet someone
in person, fix a car, make dinner, or whatever.  Computer design
will then require a fuller sense of what people do with their bodies.
That's already happening with wearable computers, so-called augmented
reality systems, so-called smart meeting rooms, and so on.  I just
think we have a long way to go before we remember -- or learn for the
first time and integrate into our design practices -- how information
happens in places, how information is part of embodied activities,
what eyes and hands really are, and so on.

That said, I also sensed that most of the designers in "Information
Architects" did not convey a strong analysis of the institutional
settings in which people were using their artifacts.  This includes
their job categories, their structured relationships, the rules
they're working under, the constraints and goals and career paths that
provide the long-term horizon for their work, their diversity as to
culture and language and institutional location, their various states
of knowledge and expertise and how those attributes of a person are
themselves institutionally organized, how the individuals were members
or larger groups, how they fashioned their identities as members of
such groups, what other information resources they had available as
part of the settings in which they used the stuff that the designers
designed, and so on.  These are surely all important dimensions of the
user's experience, and designers should surely take them into account.

A next step, it seems to me, is to build on the terrific work that
the designers have done by expanding the scope of analysis that is
brought to bear on information design problems.  That was the goal of
my essay on "Designing genres for new media".  The title was perhaps
inaccurate, given that I never did explain how to design a genre,
figuring that designers already knew that.  My point, rather, was to
provide more analytical input to the design process.  Other analytical
methods exist besides mine, of course, and still others are needed.
My point is simply that it's important to start with the existing
disciplines of design and all they've learned, and move forward from
there into a new technical space and new analytical considerations.
No longer are we designing artifacts whose whole functionality can be
held in your hand.  New electronic artifacts can easily be "tethered"
back to an organization, by wires or wirelessly, or else they can
communicate instantly with similar artifacts all around the earth.
So it no longer makes sense simply to design artifacts.  Rather, we
have to design artifacts and institutions at the same time, and the
first step is to understand the institutional location in which the
artifacts are being used.

I talk with a lot of people about "new media" business plans, most of
which are remarkably naive.  My sense is that at least half of them
can be filtered out immediately based on one simple calculation: how
are you going to recover your production costs?  Information being
what it is, your revenue is at least potentially proportional to your
audience.  That may sound like good news, but in the real world it's
bad news for the simple reason that you are competing with businesses
who have already built massive infrastructures and marketing machines
that enable them to distribute their content to more people than you
can dream about.

Let's do the math.  If you spend $X producing some content, and you
manage to reach an audience of N people with it, then you need to
recover $X/N per audience member in some combination of fees and
advertiser support to break even.  And you will lose unless that ratio
is at least equal to that of your competitors.  If your competitor
is television, then you're dealing with an industry with a very large
N -- in many cases three orders of magnitude higher than you can hope
for, even in the medium term.  That means that you need production
costs that are three orders magnitude lower than television, or else
you must persuade consumers and/or advertisers to pay three orders
of magnitude more for your product than they do for television.  That
does happen, of course, with specialized industry newsletters and the
like.  The problem is that way too many new-media business plans are
written by people who believe that they are pioneering something sui
generis, and who therefore take no account of competition.  This is
the story of interactive television, which has much higher production
costs than regular television and many fewer households equipped to
receive it.

Bret Pettichord made an interesting observation in response to my
article about dissociative patterns of thinking in the field of AI.
I noted in that article, as in my book and some other articles, that
computer people frequently employ the same word to identify both a
thing in the world and a representation of that thing.  He provided
an example that hadn't occurred to me.  In his organization, he
says, the computer support people use the word "bug" to refer both
to a malfunction of the software and to a user's message complaining
about the malfunction.  In fact the latter usage predominates; to
"fix a bug" is, operationally speaking, to resolve a complaint, not
to repair a malfunction.  "Bugs" -- the users' messages -- go into
a formalized queue, are assigned an identifier and a priority, become
someone's responsibility, and so on.  They are institutional facts,
and they are more objective in that way than the error that may or
may not exist in the code.  This use of the word "bug" is hardly
unprecedented in the English language -- in rhetorical terms it's
a metonymy, and as such it is not inherently wrong.  Bret's concern,
which we both recognize is hard to evaluate, is that this ambiguous
use of language might lead the support people to perceive the users
themselves as the problem.  Even if no such effect can be proven,
it would nonetheless seem wise to adopt Bret's suggestion that users'
messages reporting possible software problems be called "gripes"
-- a term employed in some other fields.  That's still not the most
neutral of language, but at least it clearly differentiates the report
from the problem being reported.  Relations between technical support
people and their users are too often strained and defensive, and a
little linguistic reform might help.

This observation about the word "bug" is just one small sample of the
complicated fault lines between the technical and nontechnical worlds.
Almost everybody I know lives along these fault lines -- people who
are not just technical, and not just something else, but hybrids:
media artists, sociologists of technology, people who make software
models of social phenomena, interface designers, computer industry
reporters, cyberlibrarians, cyberlawyers, computer scientists who
have moved to nontechnical fields, people from nontechnical fields
who became gearheads on the side, and so on.  These folks combine
technical stuff and nontechnical stuff in a great diversity of ways.

I find these people terribly interesting.  Whereas all sociologists
(for example) live reasonably similar lives, following career paths
that have been well-mapped by others, these people who straddle the
technical/nontechnical fault lines are all unique.  Every one of them
has had to figure out their own identity and pioneer their own path.
They are, by necessity, good at explaining themselves to a variety
of audiences.  I'm not saying that plain-vanilla lawyers and plain-
vanilla anthropologists are any less interesting as people.  The
difference is that someone whose career moves entirely within an
established discipline can presuppose a great depth of tradition.
People whose careers span boundaries, by contrast, need to improvise,
cobble together, have a strategy, and reinvent themselves and the
world around them every few years.  This life is not for everybody,
and I wish we had better cognitive tools and institutions to help
support such people's careers.  Theirs is, after all, a pattern that
more and more careers will follow in the future.

Language provides a window on the social processes that cross the
technical/nontechnical boundary.  Let us consider two more examples,
the words "moderated" and "listserv".  Both words, obviously, arise in
the context of Internet mailing lists.  In vernacular usage, the word
"moderated" derives from a particular context: there's a public debate
over some issue, with speakers representing various sides, and then
there's someone called a "moderator".  The moderator is supposed to
follow certain norms: not expressing an opinion on the issue, applying
the rules of debate in a way that is fair and impartial, and so on.
The vernacular concept of "moderator", in other words, presupposes a
particular institutional setting and all of the roles and values that
go with it.

On the Internet, however, a "moderated" mailing list is a mailing list
with a designated individual, the "moderator", who has the technical
means to determine which contributed messages are distributed to
the list's subscribers.  It's a technical concept, defined entirely
in terms of the functionality of the mailing list server.  People who
know how a mail server works can use the term "moderator" correctly.
And the term applies equally well to any way in which the software
can be used.  So, for example, technical people routinely refer to RRE
as a "moderated" mailing list, even though I (the presumed moderator)
make no pretense of being fair, holding back my opinions, letting
people speak, encouraging discussion, etc.  That's not the purpose
of the list.  This situation creates misunderstandings.  I personally
cannot get accustomed to the technical version of the word, and so
I sometimes misunderstand system administrators who ask me questions
like, "is RRE a moderated list?".  Misunderstandings also arise when
nontechnical people, hearing that RRE is a moderated list, expect to
find discussion going on there.  They contribute their comments to
the discussion, find them automatically rejected by the mail server,
and try to read meanings into the rejection.

This pattern is quite common: technical people harvest words from
the vernacular language to identify technical concepts, and they
rarely (if ever) intend to retain every single connotation that a
word carries in vernacular discourse.  Instead, they draw on very
specific connotations of the word that map roughly onto the technical
functionality they want to name.  The technical concept is therefore
both narrower and broader than the vernacular concept: it is narrower
because certain elements of the vernacular meaning are not carried
over into the technical setting, and it is broader because it is
applied in every context that is consistent with the narrower subset
of connotations.  Problems then arise when nontechnical people do
not possess the mental models that they need to comprehend the full
scope of the word's use in technical settings.  The problems are even
worse in the many cases -- such as "moderated" -- where the technical
concept is meant to occupy and support the same institutional locus
as the vernacular concept.  Some mailing lists, that is, really are
"moderated" in the vernacular sense of the word, and in that context
the word retains its original meaning.  Users, especially nontechnical
ones, who become accustomed to the word in that sort of context may
understandably become confused when it is used differently in other

To explore this phenomenon more fully, let us turn to the second word,
"listserv".  The other day I had breakfast with a representative of
a national organization who was recruiting me to write and speak about
technological issues for his membership.  Although I am sympathetic
to this organization, I was unwilling to do what he asked because I
didn't feel that he was proposing to use my time effectively.  He had
an old-fashioned model of communication.  With the Internet, I (like
many others) easily and routinely distribute my thoughts to thousands
of people.  It's not worth my time to write for a newsletter that
only hundreds of people will read, unless those hundreds of people are
very targeted indeed.  I suggested to him that he establish a mailing
list, rather like RRE, to circulate relevant material to anyone among
his membership who wants to receive it.  Advertise high quality and
low volume, I promised, and stuff will get around.  Alas, he didn't
understand what I was talking about.  He was aware of the concept of a
"listserv" -- a term that I did not use.  But for him, a "listserv" is
a discussion list, and he said (reasonably) that he doesn't have time
to maintain such a list.  I had to explain to him that a "listserv" is
quite capable of one-way broadcast distribution of material, and while
that might not sound so democratic in the abstract, it can play a
perfectly constructive role in the context of an organization that is
democratic in larger and more important ways.

What accounts for this guy's misunderstanding of the word "listserv"?
Notice, first of all, that the word "listserv" was born technical;
although it was derived from vernacular elements -- "list" (as in
"mailing list") and "serv" (as in the technical term "server", itself
derived from a vernacular word) -- nobody could mistake it for a
vernacular word.  Notice, too, that he was using the word "listserv"
as a generic category, defined in terms of its functionality and
social role, without evincing any awareness that "Listserv" is the
proper name of a computer program.

The misunderstandings here -- a technical term assigned unwonted
nontechnical meanings -- run in the opposite direction from those
associated with "moderated".  In both cases, however, the connection
between the two meanings is the same: the vernacular meaning is
situated in institutions -- that is, in typified human relationships
-- whereas the technical meaning refers to the input-output behavior
of a software program.  In the case of "moderated", we can hypothesize
that the problem arises because technical people have mutilated a
vernacular term.  But that's not fair, or at least it's not complete,
and the case of "listserv" shows why.

With "listserv", the problem concerns the epistemic situation of the
beginner -- that is, the ways in which nontechnical people are able
to know things about technical systems.  Faced with a new situation,
whether computer-intensive or not, people will start to understand
it operationally -- that is, in terms of what they can see and
do.  They will learn how the situation behaves -- "when I did this,
that happened".  In particular, they will encounter the situation
holistically.  (On this topic see the "practice theories" of thinking
in Pierre Bourdieu, Outline of a Theory of Practice, translated by
Richard Nice, Cambridge University Press, 1977; Sylvia Scribner and
Michael Cole, The Psychology of Literacy, Harvard University Press,
1981; and Jean Lave, Cognition in Practice: Mind, Mathematics, and
Culture in Everyday Life, Cambridge University Press, 1988.)  An
expert will probably have a mental model that conceptually separates
the different elements of the situation and predicts how they will
interact.  A beginner, however, has no such model, or a wrong model
imported from somewhere else, or models that are fragmentary and
superficial when compared to those of experts.  And beginners form
their models inductively, generalizing from their experience.

A correct model will channel generalizations in the right directions.
Lacking correct models, beginners often overgeneralize.  Recall,
however, that a technical concept such as "moderated" and "listserv"
cannot be understood without a mental model of the software -- a
model that cleanly separates the behavior of the software from all
of the other elements, institutional and social and strategic and
meteorological etc, that the user's total situation might contain.
My friend from breakfast is familiar with "listserv's" from contexts
in which discussions are going on; he does not possess a mental model
that enables him to generalize to other potential uses of the same
software, and thus to other potential ways that the software might be
situated within an institution.

This phenomenon has numerous consequences.  I have attended numerous
meetings of regional computer user groups, for example, and the gulf
between beginners and experts is usually on painful display.  These
meetings will usually have a moment when everyone is invited to raise
questions, which everyone else is then invited to collaboratively
answer.  These shared-thinking sessions are frequently inspiring --
the experts usually have the very best intentions and the beginners
are usually grateful for their attempts to help.  But they are also
frequently frustrating -- the beginners cannot use the words properly,
but without using the words properly they cannot effectively ask
their questions.  You see them standing there, trying to use the
words, knowing that they are not using the words correctly, feeling
like idiots in a public meeting.  Then you see the experts, sitting
in their chairs, trying to figure out what the question is, asking
clarifying questions that involve more technical words whose meanings
the beginner does not command, until the whole thing either entirely
falls apart or someone magically figures out what the real question
was, having encountered the same question from some other beginner in
another context.  It's as if the experts and beginners are separated
by a glass wall.

Another consequence of the problem is more even insidious.  I have
written before about my frustration in teaching Web design to classes
of primarily nontechnical students.  Surveys in class suggest that
those students have almost all explored the Web, but on a superficial
level.  Some portion of them will have built simple Web pages, and
there are always a few who already make a living building Web sites,
but most of them have only the vaguest idea what is happening when
they use the Web.

(The notion that kids these days just magically know how computers
work is, so far as I can determine, simply false.  Kids are accustomed
to having computers around, and to doing very simple things on them,
but I am not persuaded that the proportion of young people who can
program a computer, or who have an accurate mental model of even one
technical concept, has become measurably significant.  Indeed, such
notions resemble all too closely the reassurances by many vendors'
sales people that organizations that purchase computer systems can go
ahead and spend their full budget on computers because the employees
will be able to train themselves.)

The problem with the students in my classes is not simply that they
don't know what the Web is -- I'm a teacher, after all, and I know
perfectly well how to solve that problem.  The problem, rather, is
that the students think they know what the Web is.  They have quite
definite ideas about what a "Web site" is -- so definite, in fact,
that many of them survive all ten weeks of an intensive course without
comprehending more than a tiny fraction of the design choices that
they have available to make.  For me, "the Web" is a technical concept
-- it's defined narrowly in terms of what HTML and HTTP and their
friends will do when employed in one way or another.  Many of the
students, on the other hand, understand a "Web site" in terms of the
particular genres of Web pages with which they are familiar.  A "Web
site", for these students, is a batch of Web pages that include a
bunch of information about topic X.  A "Web site" for an organization
is likewise a batch of Web pages that tell a bunch of stuff about that
organization.  Stuff is laid out in such-and-such ways, intended for
such-and-such audience, and so on.

I start many of my lectures and lab sessions by saying things like,
"in order to design something, you have to let go of your preconceived
ideas, start your analysis from the beginning, and see where your
analysis takes you".  But this bit of wisdom is useless by itself --
how do you know what your preconceived ideas are if you cannot imagine
alternatives?  Analytically demonstrating a diversity of existing
Web sites will eventually help, but only if we take weeks of class
time to develop and apply a vocabulary that gradually deconstructs
and displaces the unarticulated, inductively-arrived-at assumptions
that the students have gotten from their past encounters with the

Before anyone concludes that nontechnical people such as my students
are dumb, however, notice that the inductive method of understanding
that people naturally employ in a new situation corresponds neatly
to the kinds of meanings that vernacular words tend to have: meanings
that are situated in institutional contexts.  A word like "moderated"
in its vernacular meaning, or "listserv" in the meaning that normal
people infer from their experiences, usefully relates mechanisms to
typified settings, and typified settings to the institutional orders
that those settings embody.  Technical words are frequently misleading
precisely because they "wobble" with regard to the institutional
setting -- the Internet can be used in many different ways, and
the whole elaborate process of cultural learning around the Internet
consists in large measure of the development of multiple models for
integrating the Internet into the institutional practices around it.

The early days of the Internet were driven largely by abject fantasy
about the institutional world, and early hypesters did very well by
gathering these fantasies under visionary rubrics such as "cyberspace".
Now, however, we are becoming more sophisticated in our understanding
of the relationship between two perfectly valid levels of analysis
of these matters: the technical level, which pertains solely to the
input-output behavior of a computer program, and the institutional
level, which meshes that computer program into a set of social roles,
norms, categories, relationships, expectations, genre conventions,
business plans, public policies, legal rules, and so on.

Loel McPhee at CommerceNet was kind enough to send me one of their
books, "Search for Digital Excellence" (by James P. Ware, Amir
Hartman, Malu Roldan, and Judith Gebauer, McGraw-Hill, 1998), about
emerging best practices for the use of the Internet in companies,
and I was continually torn while reading it.  Its message is that
a company's Internet presence has to be consciously codesigned and
coevolved with everything else the company does.  Part of me kept
saying "duh", since that has been the central message of this list
(among others) for some time now.  On the other hand, I welcomed it
as a sign that the fever has broken, and that the Internet no longer
stands out in the imagination as some kind of super-technology that
defies the laws of gravity.  The Internet really is participating
in significant changes in business, and the business-to-business
electronic commerce story is, if anything, underhyped because it
is so complicated and technical and unglamorous.  But just for that
reason it's exactly the kind of story that we need to tell now.  My
larger point is that our language still makes it difficult to tell
such stories.  We are continually blind-sided by the tensions between
technical and nontechnical language, and by the culturally organized
hyperbole that frustrates our attempts to bring the technical and
nontechnical worlds into a productive relationship.  I don't know how
to solve this problem, but I like to think that it would be alleviated
if everyone understood it.

Here at UCLA we are teaching people how to write.  Of course, lots
of schools are teaching people how to write, but those who read this
list regularly are aware that I am unhappy with prevailing fashions
in writing instruction.  I am trying to find a happy alternative to
the destructive extremes of both conservative and liberal approaches
to writing.  I am also trying to teach people how to write without
spending half my life writing detailed comments on their draft essays.
To this end, I've spent a lot of time this quarter reinventing a
method that (I am told) is used by creative writing workshops.  We've
set up a pipeline.  Each week, one of our PhD students is assigned
to write a 1000-word essay about the presentation of a speaker in
our department's seminar series.  Then they rewrite their draft essay
four times, each time getting detailed comments from three other
students.  These comments are not random, but are directed by detailed
instructions that I made available on Web pages.

Each round of comments has a different purpose.  The purpose of the
first rewrite, and thus the first round of comments, is to squeeze out
as many words as possible.  The second draft having been compressed
enough that the precision and imprecision of words become evident, the
second rewrite then starts from scratch based on a fresh analysis of
the overall point, the audience, and the appropriateness of language.
These functional issues having been brought to the surface, the
third rewrite addresses deeper issues of the essay's relationship to
its subject and its readers.  The final draft then cleans up details.
If you want to see the system as I've prepared it on Web pages, go
to .  Everything
is still rough, not least my own writing in the assignments, but you
should get the idea.  Some essays are now emerging from the pipeline,
and we'll send a batch of them to RRE when they're ready to go.

In teaching this class, I feel like I'm rebuilding civilization.
You can't think clearly unless you can write clearly, and I am on a
mission to revive the art of clear writing.  If my Web pages inspire
you to compel your own students to learn how to write, then that would
be great.  As an intellectual matter, my ideas about writing derive
from several sources, but especially from the Russian psychologist
Lev Vygotsky.  Vygotsky was one of those complex figures who managed
to conduct legitimate research during the Soviet era, and one always
reads that stuff wondering how much of it fundamentally depends on
its ritualistic invocations of the supposed wisdom of Lenin.  In fact
another major source of my thinking was another such figure, Mikhail
Bakhtin, without a doubt the finest literary critic ever, and several
scholars observed that these theorists really have more in common with
American pragmatist philosophy than they do with anything from Lenin
or even Marx.  I think that's basically right.

What these guys had in common is the idea that thinking is not a
monologue but a dialogue, or a polylogue I suppose one should say,
a conversation of many voices.  Vygotsky believed that you learn
to think by internalizing your interactions with other people.
Now Freud believed this too, but Freud focused only on the negative
interactions, and only on the emotional aspects of them.  Vygotsky
extends the point to include all interactions and especially
interactions in which people are working together toward a shared
goal.  Bakhtin, for his part, demonstrated how people's voices,
particularly the voices of the authors and characters of literary
works, combine existing voices from the society.  Together these
theories provide a theory of learning that is quite different
from anything in the west (see James Wertsch, Voices of the Mind: A
Sociocultural Approach to Mediated Action, Harvard University Press,
1991).  I am not "applying" the theory in the sense of taking it
down off a shelf and mechanically cranking out its consequences.
I came to these ideas by reflecting on my own experience and looking
for concepts that could make sense of it.  That's how theories should
be used -- appropriated incrementally and dialectically through one's
own experiences of reflectively doing things, and not turned into a
scholastic code.

In any case, the idea is this: by creating a pipeline of drafts and
structured comments, I want to establish a stable, flexible pattern
of cooperative activity that students can internalize.  Vygotsky's
theory predicts that every classroom becomes part of your mind,
and so it's better to have a classroom that embodies positive values.
Many liberals, reacting against the rigid and factory-like classrooms
of authoritarian educators, have interpreted this idea as an argument
against structure.  But that's not right.  You can't learn without
structure, and the key is to provide a structure that adapts to each
individual's needs.  The structure should also provide what Winnicott
called "holding", or that others call a "container" -- a supportive
emotional framework that pushes people to do their best without
shaming mistakes.  This is another area where both liberals and
conservatives get confused.  Liberals value self-esteem, but their
actual methods do little to encourage self-esteem while causing much
damage in other ways.  Conservatives go the opposite direction and
seem to make a special point of shaming people as much as possible.
Let's get them both out of the classroom, and instead develop a
positive vocabulary for the emotional things that happen as people
learn together.

For example, when commenting on someone's draft, it's important to
distinguish between perceptions and judgements.  Perceptions aren't
right or wrong; they're just what you saw when you read the author's
essay, and the author can benefit from learning how others saw
it without anybody deciding whose fault any unhappy misperception
might be.  It's also useful to distinguish between perceptions
and suggestions.  Too many commenters suggest revisions to a draft
without explaining the reasons for them; if the author rejects
the suggestions, it's still important to know what perceptions
motivated them.  Norms of confidentiality are also helpful, as well
as occasional speeches about phenomena such as impostor syndrome
(the amazingly common delusion that one has been accepted to graduate
school by mistake) and the skill of reporting bad news about someone's
work in a supportive way.

What does not help anybody is false praise, like gold stars on
badly written papers.  Students need guidance about what to do next,
but this guidance can be framed in terms of goals for improvement and
not in terms of absolute gradings of "A" and "C+" and "F".  Provide
steady, cheerful pressure to do better, make precise what "better"
might consist in, and leave the grades for the end of the course.
Following this method, I find that I can lean very hard on students to
improve their work.  I just smile, commiserate with the difficulties
of learning, and say, "the next step might be ...".  The proof, of
course, is in the writing that the students actually do, and we'll
soon enough see what happens with that.  In the meantime, I hope that
others will benefit from what we're trying to do.

For some time I've been aware that many libertarians oppose the whole
idea of intellectual property.  This has always clashed with my sense
of libertarians as people who regard private property as the very
foundation of liberty.  And so in the midst of a conversation about
something else, I asked loyal RRE subscriber and libertarian Kragen
Sitaker about this.  He replied that intellectual property is just
another kind of monopoly resulting from government manipulation of
the economy.  I then asked him, "why is intellectual property any
more of a government-granted monopoly than any other kind of property?".
Here (reformatted for compactness) is his off-the-cuff response, the
likes of which I had never heard before, though Kragen assumes that he
did not invent any of it himself:

In four ways: naturalness, uniqueness, expense, and unenforceability.

  * Naturalness. Other kinds of property are, to some extent,

Land has been "owned" by groups of people for tens of millennia.
Individual possessions have been owned since the beginning of
agriculture. It is possible to defend one's right to a tangible piece
of property by physical force on an individual, or small-group, basis.

By contrast, intellectual property has only existed since perhaps
1600 (although I'm not really familiar with the history, so it might
be more recent than that) and has never been as widespread as it is

This doesn't explain how intellectual "property" is more monopolistic
than other property, but only how it is more government-granted.

  * Uniqueness. Ideas are unique. If my neighbor has a car, and I want
one, I can hire a mechanic to make one just like it, assuming I can
get a good enough look at his. End result: two separate cars with two
separate owners.

By contrast, if my neighbor has a technique that enhances her ability
to rake her yard, and I want to use an identical technique, I can
analogously hire an industrial engineer to design an identical
technique. But the end result is not two techniques, but one. Who owns

This is important in the context of monopoly. If Al builds a big,
successful car factory, and Henry thinks he could do a better job of
building cars with a car factory, Henry can build his own factory. He
can't use Al's, but he can build his own factory, and he can compete
with Al.

By contrast, if Imatec comes up with a big, successful technique to
calibrate displays and printers so their colors match, there may very
well be no possible alternative way for Apple to calibrate displays
and printers so their colors match. It's as if, once Al builds his
factory, anyone else who wants to build cars has to rent space from
Al. This obviously produces some problems.

  * Expense. It costs about $700 to register a patent establishing
"ownership" of a technique. It costs $0 to establish my ownership
of a bicycle, and about $30 to establish my ownership of a car. This
clearly favors folks backed up with big capital.

  * Unenforceability. Effectively enforcing intellectual-property
claims requires a totalitarian government that monitors every
communication between two people, and possibly everything everybody

If I have ten thousand objects in my house, I can be relatively
certain that only a small fraction of them will ever be stolen. If
I live in a modern country, it's likely that none of them will ever
be stolen, even if I invite people over and lend them books regularly.
By contrast, if I originate ten thousand ideas, nearly all of them
will be "stolen", unless I never implement or describe them.

Some people observed that my lists of "things that are are good" and
"things that piss me off" dwelt on matters pertaining to travel, and
they teased me for my seemingly glamorous lifestyle. I wish. What's
really going on is that things -- both good and bad -- are easier to
notice when you're traveling. In a familiar environment, something
can annoy you very mildly at 2:37pm every afternoon without you ever
focusing your attention on it. But when you're traveling and all of
your routines are disrupted, you can encounter annoyances for the
first time, before they fade into the woodwork.

This phenomenon was a central motive for my dissertation research on
computational models of routine activities. I strongly believe that
our everyday lives are much more complicated than, and completely and
qualitatively different from, our conscious understandings of them. We
selectively notice some aspects of our lives and not others. We notice
those few things in our lives that are new, or difficult, or broken,
or that fit into our cultural or ideological sense of what's important
to notice, but we are only dimly aware of the 99% of our lives that
doesn't fit into those categories.

This is important for many reasons. One reason is that the people who
design computers get their ideas about their users' lives largely from
their cultural and ideological understandings of their own lives. If
those cultural and ideological ideas are mostly wrong, as I believe
they are, it follows that computers will probably be accidents waiting
to happen. Because computer users are absurdly accommodating, however,
the problem can go undiagnosed for years on end.

In my dissertation I wrote about these things rather bluntly in the
context of AI and its ideas about human action. I then rewrote about
80% of my dissertation in more scholarly language, with footnotes and
everything, to produce my book, "Computation and Human Experience".
Most of my friends were opposed to this rewriting, but I didn't want
to go up for tenure with a book that was filled with blunt polemics.
I've already written about some of the techniques that my friends and
I evolved for becoming aware of the normally invisible details of our
own everyday lives:

It's hard to convey how strange an experience this was. You can get
some sense of it by moving to a new house: as you put all of your
material belongings into boxes, your everyday routines start to break,
one by one, and you confront the fine details of just how your life
had been organized. Then you move someplace different and every day
is filled with little problems that you face for the first time:
the first time you take a shower in your new bathroom, or turn a key
in your new lock, or look to see what time it is, and so on. Then
the second time you do each of those things, you start to settle into
a routine; your experience from doing it the first time comes back
to you, and for a small moment you're aware of choosing to do it this
way. Usually you have bigger things on your mind, however, and so
the moment slips away and you forget all about it. Detailed real-time
awareness of such things is what Buddhist meditation is about, although
our methods for becoming aware of them bore little relationship to
those of the Buddhists.

Technological change, with its inevitable disruptions, ought to be
an occasion for us to recover this kind of fresh experience of our
own lives. Mostly, though, it isn't. Instead it's a hassle, or it
gets drowned by emotional noise and hassle, or we go into it kicking
and screaming, or we go into it under the influence of millenarian
ideologies. The problem in each case is that we focus on what's new
about using a different technology, rather than on what is revealed
as more fundamentally the same. This concern was the motivation for
Terry Winograd and Fernando Flores' book "Understanding Computers
and Cognition: A New Foundation for Design" (Ablex, 1986), and while
I have some problems with their particular formalism, many of the
same underlying intuitions motivate my own ranting and pleading on
the subject of the tacit millennialism of computing. The fundamental
problem with this millennialism is not that it is bad religion,
though that's what it is, but rather that it pulls people away from
experiencing their own lives. If you're not aware of your own life
then you can be tricked into any old illusion.

On the subject of pens, Michael Dennis tells me that there's also a
Kinokuniya Bookstore in New York at Rockefeller Plaza (10 West 49th
Street), and Bruce Jones tells me that there's another one in LA that
I'll soon be looking up.

Also, Staedtler is on the Web at

Some URL's...

Deepening the Digital Divide: The War on Universal Service

12th International Bled Electronic Commerce Conference

Geoff Nunberg on the future of the library

Electronic Policy Network

article on university "science shops" that help the local community end