virus: Web, links, and substance

Alexander Chislenko (
Mon, 30 Oct 1995 21:26:31 -0800

A colleague of mine recently complained that
"this Web is all links to links to links, and no substance".

My immediate reply was - no, there is a lot of stuff there,
just follow the links, don't be lazy, the links are there to help
you find the STUFF a lot more conveniently than you ever could before.

Now I think there is a better answer to this.

One could claim that the Web is *nothing but links*, and my colleague's
strong statement is *literally* true.

Consider a hypertext journey:
linkpage1 -> linkpage2 -> ... -> linkpageN -> (aha, STUFF!) Text:

"Sasha and Masha are friends".

Finally, you found some radiant STUFF! at the end of the link tunnel!
Or... have you?

Now, let's examine the above little text:

- "and" is a *link* between "Sasha" and "Masha"
- "are" is a *link* between "Sasha and Masha" and "Friends"
- "Friends" is a relational link and surely not STUFF! by itself
- "Sasha" and "Masha" are still just words that consist of linked letters
and are, in turn, no more than links to the real Sasha and Masha who live
off-line (or, in this case, to imaginary ones who do not "exist" at all).

So then - all real STUFF is off-line, right?
- Not so soon. Aren't people "just" sets of linked atoms?
Aren't atoms just a bunch of linked quarks?
Now we reduced all "substance" to substrate quark powder, which sounds quite
absurd. Even if I agree to equate "substance" to "substrate", the *only*
observable - and interesting - feature of this "substance" seems its ability
to link into complex entities. The behavior of these entities is
determined by the nature of the links and their joint architecture.

If you change the links considerably, you not only add new ways (language
or HTTP) to refer to the old entities - you change the behavior of the system
as a whole.

So the links of the Web are its substance in about the same way as
conceptual references provided by the wordings of the linked texts, or
operational instructions contained in the machine-executable files.

Maybe, that's evident.

| Alexander Chislenko | | Cambridge, MA |
| Home page: |
| Mailing list: |