Re: INFO: Hypertext

Steve Witham (
Thu, 6 Mar 1997 23:08:54 -0500

Max Rasmussen <> says:
>The Byterange scheme won't work in a hypermedia enviroment because the
>media is so easily edited. There got to be some kind of TAG in the actual
>text for any "include" system to work.

Byte ranges work with editing if old versions remain available.
That way, you refer to a range within a version. There's an interesting
technical challenge in following a byte-range link from an old version
to the corresponding place(s) in a new version.

>From: Hal Finney <>
>I'm not sure I see the necessity of preventing someone from changing
>his posted information. If he presents a weak argument, and then
>withdraws it, haven't those who disagree accomplished their purpose?
>How much more is gained by continuing to throw his earlier errors in
>his face? If the goal is to approach the truth, then there should be
>nothing wrong with people evolving their publications in the face of

If there is a mechanism for the author to publish subsequent versions,
and a way to see that the old version is superceded, and how the new
version differs, then we can keep the old version *and* let the author
retract/revise/ammend/clarify... it.

I can't think of a please-don't-store-and-retransmit protocol that couldn't
easily be gotten around by crypto and scrambling routers, even if *most*
people wanted to honor such requests. Even if you say that only the
original signer of a message can retract it. If the system is built to
prevent censorship and traces by info police or spies, then (I think) it
can easily be used to reveal retracted writings. And if there's
a demand for that, then (seems to me) there will be people in the business.

My instinct says that a system built around the truth works better than a
system built around an idealization of the truth. In this case the truth
is that things have been said in the past, *and* sometimes people misstate
things or change their minds, and sometimes people *really don't want* you
to read certain information. The system should represent all of these.
It would be idealization to think that everyone is trying to make the most
comfortable environment for debate, that no on wants to unnecessarily
embarrass anyone, that we can control information.

Funny finding myself writing about "the system" meaning the overall
effect that the diverse things going on on the net ought to have and
be designed toward having. I believe that design is one of the
most important parts of politics...but I'm also an anarchist. So I
believe in design for *and by* distributed systems. But still it makes
sense to say "the" system "should" work a certain way.

Also funny to find myself taking this somewhat fatalistic attitude toward
information control. I said before I don't like the gung-ho, "live with
it" style of futurism. But (I did also say) you gotta pick your battles.

I wrote:

>Fine grain links, bidirectional links, versioning, permanent
>storage, redundant storage are all nice features.

I forgot royalty, especially fine-grained royalty, and inline quote
links. These are all features of the original (~1981 spec) Xanadu


--           Steve Witham
      ___       ___       ___       ___
| | | ___ | | | ___ | | | ___ | | | ___   --pattern in the heater grills
| | | ___ | | | ___ | | | ___ | | | ___     of Boston Red Line subway cars