Re: [-empyre-] archiving + preservation

[ben syverson <ben@bensyverson.com>]

On May 6, 2004, at 7:40 PM, Jaka Zeleznikar wrote:

> I find this a bit suspicious:
> - dragon (at least in my culture) has noting to do with computers (so 
> you can link anything to 'computer')
> - 'foxglove' can be popuped to 'fox' and 'glove' (but does it make 
> sense?)
> - 'rob' can be linked to 'rob' (first 'rob' in English, second in 
> Slovene (meaning 'border') ...
> - dragon: 'drag' 'on' ...
> - 'd' 'r' 'a' 'g' ...

True, although liken does not support links at the atomic level -- that 
is, we don't allow individual characters to become links. The cool 
thing about Chinese/Kanji is that each pictogram represents one or more 
concepts(although its true as Ken pointed out that they also describe a 
sound), whereas in the Roman alphabet, "c" has no particular meaning. 
While the sub-concepts of a complex word (ie, "dragon" in computer) may 
have little or nothing to do with the complex word, that is not 
necessarily a drawback in the context of liken, and may actually be a 
benefit. To address the "rob" issue, liken was really intended to be 
used with one language and its sub-dialects at a time; once you 
introduce more than one language, things quickly get too wiggy. 
Although the idea of a pictogram-native and pictogram-only wiki/liki 
continues to intrigue me.

Liken was designed with the knowledge that its suggested links may 
sometimes be non-sensical [and/or] irrelevant -- to criticalartware, 
this was an interesting feature of the system. Since liken's decision 
to proactively link two things is never actually random, but rather 
based on word frequency, it could link two documents that have nothing 
to do with each other, but that use similar language. A system of paths 
that was at least partially based on a common lexicon or snippets of 
shared language was exciting in the extreme.

If you take this further to the autolinks ("liki" links), "foxglove" 
can break down to "fox" and "glove," it's true. But "fox" and "glove" 
are not likely to be extant nodes in liken, since its goal is not to be 
a general purpose knowledgebase. However, "strong" will display a link 
for popular motion picture "TRON." Does that link make sense? Maybe 
not. But does the [practice/procedure] itself make sense? I think so. 
You don't wan to limit "TRON" to linking to just the word "TRON," 
because what if you get this: "TronCycle"? You definitely want that to 
link up. We'd have to have millions of nodes before over-atomization 
became a real problem.

I've come to love these features of liken, since they underscore the 
design goals of connecting disparate chunks of information. To navigate 
liken is to walk into a world of serendipity, ultrahyperlinking, 
complete relevance and utter nonsense.


> + if I understood correctly links are made more visible by being more 
> visited.
> I understood Vanevar B. and some other people that they wish something 
> similar to
> the way we think - and we think in emotions also (so some terms in way 
> of
> links might be more visible relating to our emotional state and not 
> only conceptual,
> so one day strongest link from rock might be music and the other day 
> geology)?

Most definitely! However, liken is a bit different than a singular 
mind, because all of our users help shape and alter the structure of 
liken. We like to point out that liken is like a neural network, but 
the local computation is not being done by small computing processes or 
neurons, but rather, by human computers.

So in that way, liken reflects how the collective consciousness of our 
users thinks. So one day, the strongest link from rock might be music, 
and the next day it might be geology, but it won't be because of one 
person, but rather the distributed neural net that you're already 
participating in by posting to empyre!


> + any way I find the project extremely interesting

Thanks! We continue to be inspired + excited by our users, and we 
constantly [re]examine our methodology...

left-click,

- ben