INFO: Hypertext

Dan Fabulich (
Sun, 02 Mar 1997 23:27:45 -0800

I've recently taken interest in reading about hypertext, a decentralized
system for fine-grained bi-directional linking in writing. For a good
description, read Drexler's "Hypertext Publishing and the Evolution of
Knowledge," which makes an effective argument for hypertext.

I began to think about how such a system could be implemented over the
existing Internet. As I thought about it, I realized that through some
small changes in the protocol, Usenet already provides an ideal system
for implementing hypertext. It already allows for coarse-grained
bi-directional linking through message identification in the header
file. No article can be retracted from Usenet once it has been
submitted; your article has already covered the globe many times over by
the time anyone could have thought to suppress it. Hypertext could be
implemented simply by tweaking the NNTP protocol to specify not only the
message id, but an exact byte range within the body, and by slightly
altering the "news:" URL to read individual articles within those byte
ranges. From then on, instead of quoting an article, you'd add a line
in your post which referenced to "news:<messageid>/243-257." It seems
almost too easy.

Yet there is a problem which I can't resolve. I'm posting this here to
see if some Extropians can help me to think of an answer.

One of the most important inabilities in a hypertext system, according
to Drexler, is the inability to retract articles. This allows for a new
and powerful mode of criticism which was impossible before hypertext.
Did you notice someone contradict something they said long before? No
problem: simply reference their original article along with their most
recent post in your criticism. Now, whenever someone reads the
offending post, they will ALSO see a link to your criticism and the
article stating the opposite. Permanent back-linking is also important
to historians and other social scientists who are interested in studying
a particular point in time, be it last week or five decades ago.

Unfortunately, as far as I can see, "permanent" and "decentralized"
don't seem to fit together. In Usenet, there's no guarantee that you'll
find the post which you read last week; say nothing of decades ago.
Nonetheless, the permanent archiving of at least SOME of this
information is useful and important. How, then, can this issue be
resolved? Who would pay for a system to archive Usenet forever?

Drexler and other hypertext proponents suggest a royalty system, where
each time someone reads the text, the reader would pay the author and
archiver. Unfortunately, implementing a system like this seems
unfeasible and even undesirable, as it is antithetical to the idea that
"information wants to be free." Royalties try to restrict the
transmission of information and assign legally binding (read: government
enforced) monopolies on content. Royalties will not and should not
stand up in the face of the Internet, IMHO.

Individual authors cannot pay for their data forever. Eventually (if
not immediately) it would become unprofitable to pay for their
unretractable articles. Besides, what happens when authors go bankrupt
or die? Advertisers might pay for their data to be distributed, but
eventually they will want to stop paying for their advertisement. What
happens to their content then?

How about at least charging one-time micropayments or subscription
charges for information storage and retrieval? That would allow the
free market to take over. Information traders could pay a one-time fee
for information and set competitive prices to distribute it. People
could make a profit on their storage. Surely a system like that could
last indefinitely, as there is no end of information and no end of

But what about the permanent public archive? Most of USENET would be
profitless to store. If you're just a guy out in California who noticed
that the president has flip-flopped on health care again, would anyone
pay you to write about that? For that matter, would archivers even HAVE
a copy of the president's five-year-old health care press conference on
file? Even if traders charged a great deal to store arcane materials,
there would be very few buyers interested in most of that stored data; a
cost-conscious information trader would wisely pitch the least used
material and keep only the most profitable.

The hypertext system must somehow remain public and permanent to allow
for this new and obviously desirable form of criticism. Unfortunately,
no one will pay for such a permanent public archive, except perhaps the
government. (I barely even want to consider this suggestion. Like most
Extropians, I don't like government spending.)

At this point, I'm despairing that a permanent decentralized hypertext
system can work in a free society. If it's permanent, then it all must
be stored somewhere, and it will never be cost-effective to do so. If
it's decentralized, the system will drop unprofitable information and
may refuse to buy from most of the public.

Any ideas?

-He who laughs last thinks slowest-