30 August 2006

The Yahoo zodiac on the Tree of Life

color in green
on the yahoo zodiac
all discussions
of activities you enjoy

and color red
those you don't enjoy

shrink the zodiac down
to fit inside the 'head'
of any/every human branch
on the tree of life

and where that branch
is doing activities it enjoys
light that zodiac-point green
(a green stripe for the duration)

but if there's a mix
of enjoyable and not
show this as a mix of
green and red stripes

consider babies' mix of stripes
and how the reds begin

they don't know how to fix
except by wailing

but as they're 'socialised'
other branches exert
more-or-less subtle pressures
to betray their own instincts
and admix artificial reds
with natural greens

until as adult
reds become their 'normal' color

until as post-adult
they recover their green

29 August 2006

No-Ajax is better than Ajax

rebooting my machine
to kill off some runaway
Firefox javascript

reminds me
that no-ajax is better than ajax

The self-conscious computer

unlike the undefinable 'consciousness'


is trivial
to implement in computers:

even the blinking lights
of 50s scifi cliche computers
represented a sort of self-consciousness
as they announced its inner state
to the world

and it wouldn't be hard
to add a layer of nano-probes
above the cpu chip
that monitored which subparts
were active or inactive

helping find bottlenecks
infinite loops

or at the OS level
monitoring which system calls
are used the most
which processes
are stealing all the cycles

and at the desktop level
which apps get used most
which command sequences

and when this becomes second nature

the principle of
most-often-used = easiest-to-access
will be the fundamental
organising principle

Your OS avatar

(this would have been a slashdot comment
but that thread is way too far along)

if your OS is ever going to adapt intelligently
to your ways of working
it will need a general model
of all the ways anyone works

and a specific model
of your special quirks

even to the level
of a map of your physical workspace
with an avatar reflecting your current posture

and whether you have the teevee or radio on
or when a phonecall comes in, distracting you

we should picture that animated avatar
mirroring the realworld phonecall
then readjusting posture
and resuming old tasks

so add to this OS-mini-sim
a pixellated model of the pc display
that understands
more or less
why you have different windows in front
for different tasks

and can run alternate animations
of all the different
workflow command sequences

(can you guess the 100 commands
you invoke most frequently?)

and all this avatar-modeling
can also be previewed in
seminatural language

a little novel about computer use in general
and the peculiarities of one user

(intermediate stage? an expansion pack
for The Sims
that gives a sim a simulated net connection

allowing socialising
accidental data loss

28 August 2006

Indra's rainbows

metaphorically (at least)
the mini-images of the zodiac
within each sign
can be viewed as rainbow/spectra

as Indra's spherical dewdrops
refract the full spectrum of white light

but also
a disturbance in any sign
should set up a mirror disturbance
in the mini-mirrors

if they're sensitive

27 August 2006

Calculating Yahoo-signs

we've established a
Yahoo-betical order
for the commonest million words

(but immediately loosened this a bit
by re-arranging the elastics)

so when we contemplate
a web community
we can 'light up'
every such word that's been
posted to that community

and light up, too
all the word-pairs, triples, etc

and by the time we get to
sentence-length sequences

most 'stars' we light
will be sentences that occur
nowhere else
than on that community's sites

and the elastics
linking words to phrases
to sentences
to paragraphs
to webpages

should draw all those
into a cluster
that defines that community's

A semantic Web of Indra

if a web community
'majors' in entertainment, say
and 'minors' in computers and the net:

we might expect that minor
to 'pull' the entertainment community
counterclockwise, towards computers

and since any group/topic
can have (2ndary) relations
with any other group/topic

there's a distinct 'pull' for each
clockwise, or counterclockwise
a short pull, or a long one

so each topic community might have
13 little subtopics
mirroring the 13 2ndary pulls

Indra's Net is like this

Communities on the 'Yahoo zodiac'

this new zodiac
differs from the old
'radial ontologies'
in that
the zodiac's 'stars'
are words and
strings of words

where the ontology postulated
ill-defined story-fragments

but they should map closely

now since every internet community
has its own favorite turns of phrase
they should each have a predominant 'sign'
on the zodiac

and someday no doubt
Google Trends
along with mapping places and times
of searches
will list the web communities
(slashdot, imdb, rigorous intuition)
that use each phrase
most often

26 August 2006

Yahoo-betical order

my intuition is content
that our elastic wordsphere
is more-or-less unique
(barring, say,
mirror reflections,
and crossing fingers that
it would
3D space
rather than aligning
down in 1- or 2-D)

and we can flatten it
along any chosen axis
so we choose whichever
offers the best separation
of the 14 Yahoo root-categories

and end up with a
canonical 'Yahoo zodiac'

as well as a very-fine-tuned
Yahoo-betical order
for all million words
based on their zero-to-360
radial angle

but this is just a temporary seed
a kludge to nudge our nodes
into a circle

so now we can cut that old elastic

momentarily locking words
at their Yahoo angle
but setting their distance
from the circlecenter
based on their frequency of

and now we add back in
word-pairs and threesomes, etc
attached by new elastics
to their component words and phrases

and with their orbits again
determined by the same metric
of frequencies of use

and one can hope these new elastics
when the strict Yahoobet is loosened
will enforce a pretty similar ordering

(words that shift radically
must have some
reason for doing so)

Math-quandary revisited

let's forget the circular orbits
for now

and start with the commonest word
(probably 'the')

and stretch a separate springy elastic
from 'the' to each other word

with the preferred length
determined by the observed average distance
between occurrences of these words
compared to the expected average distance
based on simple frequencies alone

so that pairs that tend to occur
closer together
will have shorter elastics
and pairs that occur farther apart
longer elastics

and we link every possible wordpair
by this metric
with the least likely pairs
linked by half-million-mile elastics
so the network as a whole
fills Moon's orbit

(not just the flat disk
but an orbit-sized sphere)

we can hope
the Yahoo clusters
will be well-separated in this space

and we can guess
the commonest words
will be pulled towards the center

and if we now
flatten the sphere

we ought
be able to
restore each word
to its frequency orbit
(most frequent closest
least frequent farthest)

without disrupting
the topical clustering...?

25 August 2006

The idealism crushers

the previews for 'bully'
raise my hopes, a little
because they suggest a battle
(in the setting of a school)
between idealists
and idealism-crushers

which is a classic literary theme that's
so far
problematic for computer games

the classic/generic modern political novel
shows an idealist learning the ropes
and suffering infinite compromises

or via charm/luck/authorial-bad-faith
turning some corner of the tide

and the same for novels about journalists
and the same for novels about soldiers
and the same for novels about scientists
and the same for novels about educators
and the same for novels about artists
and the same for novels about social workers
etc etc etc

history offers few inspiring success stories

and computer game authors
resort to the same bad faith
to give idealism room to win

(can the political reformer ever hope
to knock on enough doors
to counteract the corporadoes'
swiftboat campaign?)

but in the miniaturised society
of a boarding school
surely one can knock on every door
and enlist fair-minded allies

Radial-ontologies math quandary

we arranged the million most-common words
alphabetically along the Moon's orbit

and added every possible word-group

and then we tightened each word/group's orbit
according to how common it is

and linked groups to component words
with imaginary elastic bands

and freed each word/group to slide
within its own orbit

speculating that the elastic
would pull words into groups
approximating the Yahoo toplevel categories:

Arts&Humanities, Business&Economy, Computers&Internet,
Education, Entertainment, Government, Health, News&Media,
Recreation&Sports, Reference, Regional, Science,
Social Science, Society

arranged somehow around the disk of orbits

but mathematically
the elastics would really just
clump everything into a tarball

so a subtler metric is needed


take the frequencies of any two words
and calculate the expected average distance
between occurrences of these words

calculate the actual average distance
(smaller for related words,
longer for unrelated)
and make the elastic 'repel'
when the distance is longer

but even then
the repulsion can act
'clockwise' or 'counterclockwise'

and it's not obvious how or whether
the collective force would neatly
fill the circle/orbits

so i want to virtually 'jiggle'
the system, with elastics
so that it finds its minimal energy state

(but is this realistic?)

20 August 2006

Self-knowledge via Web2.0

"Computers could be a tremendous way for us to look at ourselves in a very spiritual way... could teach us more about ourselves than we've been able to look at, so far." --Kate Bush
anything you publish
on the internet

gives you the new opportunity
of standing back, and asking

how will others see this?
what does it show them about me?

whether it's your favorite bookmarks
your favorite mp3s
your favorite pictures
your favorite books

your political views
your personal life
your autobiography

what patterns do i favor
that i never realised?
which do i avoid?
how do i compare to others?
what can i learn from
their ways of sharing?
how do they react to mine?

are there patterns that embarrass me
in their naivete

what am i uneasy about sharing?
what happens if i share a tiny test bit?
(does it help me grow?)

what research/exercises
can i do to feel less uneasy?

how do my tastes vary
mood to mood
month to month?

shall i share different facets
on different pages
different sites?
under different names?

can i summarise all facets
on a single informative page?

can everyone
however different
re-use the format of that page?

18 August 2006

The Stalking Expert

it's easy to set up
search-engine rss-feeds
that watch for new material
on a given topic

but suppose your chosen topic
was a living human

and you wanted
even sinisterly
to watch for every tiny hint
of what they're up to
using ALL available searchengines

especially monitoring
their online presence(s)
if they have any

forums they post to
forums that gossip about them

nicknames, pseudonyms, avatars
titles of their creative works
lyrics and other quotes
family members

building a database
of facts-already-known
so that old material
can be downgraded

tapping into a continually updated
of searchable sites
(Stalker Central)

and imagine that this
becomes routine

this week only
stalk any ten celebs
for $19.99

14 August 2006

Finnegans Wake on the Tree of Life

if joyce was right
then every story on our
radial ontology
should be part of a yin-yang cycle
with a distinct natural rhythm

each healthy storyrhythm
jamming improvisatorially
with all the others

with the most extreme cycle
the saga of a hero's
trespass and fall

wrenching apart his city
his civilisation collapsing

eventually reborn
like daisies on a battlefield

and the hero's hesitancy
before this dreadful fate
producing a
in the great rhythms

a hesitant nation of meditators
tuning in
to the rhythm of their breathing

listening for the [catch]
in the improv jam

13 August 2006

The human sciences on Flickr

20th century social science
can be caricatured
as 10,000 pseudoscientists
in search of something to measure

for if you tried to reconstruct
the 20th century human psyche
or diagnose the 20thC human predicament
based solely on those ten-thousand's
your reconstruction would be shallow indeed
not even recognisably human

abstract statistics on human lab-rats
or at best stilted case-histories
feigning 'objectivity' by suppressing humanity

but imagine we could offer them
ten million family snapshots to measure

to classify by apparent emotional relationships

to trace the emergence of character traits

toddler dances unselfconsciously
kindergartener mugs self-consciously
girl tween begs the camera to flatter her

adolescent grows surly
open personality closes up
too cool to grin

first love imbalanced
one leads, one follows

intimations of disaster
even in the wedding pix

tough guy transformed by fatherhood

etc etc etc

the classic literary themes
now accessible as objective data records

ready for a new strategy of 'measurement'

09 August 2006

A radial ontology for webpageviews

to review:

we seek to ground
all thinking
in a precise common visualisation
of local spacetime

Earthglobe flattened Bucky-wise
time-layers stacked, most recent on top

colored lines tracing worldpaths

with any statistic
mappable to color

usually red for high values
and blue for low

and today we consider webpageviews

15 years of hosting

with the webhosts at any instant
in those 15 years
ranked for popularity

with the top one-fifth colored red
the bottom one-fifth blue

so we can trace
ever-increasing numbers of hosts
starting blue/slow

most never cracking green

but some few catching on fast
some few more catching on slow

popular hosts tending to cluster
in the hi-tech corridors

centers of red and orange
diluting outward to yellow and green
with the boonies normally blue

(exceptions being of special interest:
who's that, bucking the trend?)

popularity could alternately
be gauged cumulatively

so slow and steady tortoises
beat meme-of-the-hour hares

earlybirds earning regular colorboosts
just by enduring

now let's break this down by pages

pages color-ranked by popularity
instantaneous or cumulative

hosts adding pages
quickly or slow

the most popular sites
adding hot news hourly

with any individual page
quickly lapsing into obscurity

(which are the steady tortoise-pages,
the Net classics everyone must read eventually?)

now let's break it down by words

less read blue
more read red

headlines and first paragraphs
are read most
boring sections skimmed
only the best writers
making every word count

so now we count
for each second of passing time
how many times each English word
is being read

and we collect these stats for a day

and build a list
of the million commonest English words
red orange yellow green blue

now we alphabetise that list
and reduce each word to a dot
of its appropriate color

and we project that line of dots
into space
along the orbit of Moon
one million dots
across 1.5 million miles
(leaving us lots of mental room to expand
although the scale is finally irrelevant)

and into those 1.5 mile gaps
we insert alphabetical word-pairs
down to aardvark-zymergy

ten word-pairs per inch
each a colored dot
to represent that pair's popularity
using the same colorscale
as for single words

and now we divide each one-tenth inch
to hold a million threeword triplets
and again and again and again

until any popular 5000-word web-rant
has its own colored dot
based on how often it's been read


we move all the red-dot word-strings
into the closest Earth orbits

the orange dots almost as close
yellow and green also intermediate
blue dots still at Moon's length

and we link each wordpair/triple/etc
to its component words/pairs/etc
by elastic bands

(since a pairing will normally be
less common than either component
these bands will link inner-more components
to outer-more groupings)

and we'll let the elastics pull the dots
into mega-clusters

(red dots still constrained to red orbits
but freed from their original alphabetical positions
within those orbits)

one can hope
the resulting groupings
would match, eg,
the basic Yahoo categories

technical word-clusters
sports word-clusters
literary word-clusters
political word-clusters

and an individual's web history
ought to favor certain clusters

if we track any individual's surfing
we can visualise the stream of words
they're reading

shrinking our whole Moon model
down to a disk that fits
within the circumference of their brain

lighting up, each moment
the dot for the word they're reading
and for the most-recent pair
triplet etc

so when they finish reading
a popular page
they'll have incremented the 'counter'
for that page's particular word-sequence

maybe boosting it
from blue-orbit to green