Robot Wisdom auxiliary

a semantic-web-savvy pc
will know the web's deweydecimalsystem
and the consensus 'best' web resources
for each topic-area

you will customise this
to your personal preferences

rating streams of new resources
in favorite areas

catching up on resources
recommended by
topic-experts you trust

searching for experts/streams
with similar tastes

archiving/mirroring/monitoring
your web2.0 presence

 
 
 
 
.

i think my complaint
about blogging's dark age
can be boiled down to a handful
of formulae and statistics

eg for flickr (simplest case)
c4000 new pix per minute
maybe 0.1% of which (4/min) i'd fave
if i had a way to find them

the full database of a billion pix
waiting to be filtered
a checklist that should always remember
which you've already viewed

i subscribe to about 250 photographers' uploads
and 50-more's favorites
plus a single tag ('nude')
all via rss, which lets me quickly scan
mediumsized copies
of hundreds of pix per day

maybe 10% of which i fave
creating a veryhighquality stream
that should be subscribed by thousands
(current subscriptions unknown/zero?)

but missing at least 1000 potential faves/day
because i can't find them

simple strategy:
ruby/python/perl script
builds mirror of personal faves
and for each of these faves
the list of all flickrers who've faved it

and then calculates which flickrers
have faves that overlap mine most
(without excessive burden of
boring/offensive faves)

(flickr needs a stat
hinting how many pix each favoriter winnows
on average)

when you discover promising new fave-sets
you need to queue/ration
the process of sifting them
(the best ones might need better
to be added en masse
with exceptions deleted)

flickr should recommend photographers and faves
based on your known preferences

should always offer a stream
of likeliest-to-fave's
it knows you haven't seen

algorithmically tweaking its choices
as you surf

 
 
 
 
.

when i started my blog in 1997
keeping up with content, worldwide
wasn't yet a fulltime job




more or less everybody is
more or less potentially a
more or less fulltime
photographerphilosopherdiaristbloggercritic

generating barely-countable content-items
on mostly predictable hostsites

and trying, more or less
to discover and share
the best content created by others
commenting, critiquing and responding

without stooping to popular-mediocrity contests

struggling against ever-worsening odds




the browser for this future
needs to track all content authors
across all sites

remembering what's been seen
what's been recommended

streamlining the interface mechanics
so that a bottomless best-first queue
offers a customised range of content categories
with duplicates filtered/combined

new content from favorite authors
new recommendations from favorite critics
unplumbed favorites' archives

(inbox zero is passe)

 
 
 
.

flickr: pix, tags, sets, groups, comments, rss (small), flash-slideshow (w/#), faves (no sort, no tags, no sets, no ratings), faves-rss, faves-slideshow (w/#); rss/slideshows for artist w/tag, for set, for group; no aggregate feeds; privacy settings?; 3rdparty (flickriver)

twitter: tweets, subscriptions (no privacy re who), one aggregate feed per user, faves (no sort, no tags, no ratings)

imeem:

last.fm:

postsecret:

overheard:

blogger:

tumblr:

why don't all browsers share
a common core set of
smart
folders for bookmarks
mapping onto the universals of human life
via a startup questionnaire
about where and how you live

eg: each gizmo you own
mapping onto a smart folder
offering the best sites for support
localised for local service
any online communities for that gizmo
easily customised to your preferences
with a social feed (?)
that discovers new resources as they arise

a folder for each of your vehicles

for each of your pets
your hobbies
your clubs
your periodical subscriptions

for your music bookmarks
your photo/image bookmarks

bookmarks
are the primary currency/coinage of the Web

so if we want to
nail the bookmark entitytype
into its usual-story-historical framework
for semanticweb purposes

we need to sketch
the usual url-bookmark story

the creator substory:
content conceived
content created
content posted
url shared

the consumer substory:
surfer has motivation
surfer has methodology
methodology ranks url-sources
surfer offered url by url-source
surfer notes offer mentally
surfer notes offer physically ('to-read' list)
surfer visits url
surfer partially consumes content
surfer rejects content, erases notes
surfer extends notes for later
surfer wholly consumes content
surfer passes-on url with comments
surfer acknowledges url-source(s)

the derivative-work substory:
surfer conceives related new content
surfer acknowledges url's influence

the historian substory:
broader path of url traced/reconstructed
traced-path posted




of course
there are almost infinite variations
on this skeleton

but a true semantic Web
will only become possible
when every stage
of every bookmark's history
can be tagged with its appropriate
story-context




Real post date: 15 Oct 11:20am
(Blogger no longer lets me
finesse my linkblog to the top
via postdating
so my nonlinkblog posts
will now be 'pre'dated)

[draft]

i'm preparing to do some experiments
in google's new app engine

the 1st very simple: the 'portaliser'
(portaliZer was taken)
will massage a set of favorite links
into a simple page-layout

but the 2nd is much more ambitious
trying to add time-sequencing
to IMDb movie keywords/tags

eg 'Zoolander'

IMDb: main, summary

keywords/tags: Secret Agent, 2000s, Catch Phrase, Multiple Cameos, Spy, Spy Spoof, Frat Pack, Car Chase, Assassination, Brainwashing, Cameo Appearance, Conspiracy, Fame, Fashion Model, Fashion Show, Fashion, Orgy, Stupidity, Exploding Gas Station, Exploitation, Obesity, Sequel, Vanity, Saturday Night Live, Dwarf, Dwarves, Based On Sketch Comedy, Child Labor, Blackface, Bulimia, Catwalk, Erection, Femme Fatale, Malaysia, Masseuse, Reporter, Time Magazine, Directed By Star, Break Dance, Male Model, Character Name In Title


IMDb ontology experiment: [2005] (Modalities, Dramatic themes, Family themes, Sports, Animals, Music and comedy, Historical period, Crime-related, War-related, Horror and scifi, Sex, Animation)

not just
in
every grain of sand

but in every field
of every database

(even the simplest:

name address age gender title
etc)

we
must
see
the whole potential universe

ie
all those simple fieldobjects'
real-or-potential
relationships
with all other fields/objects

naturally ordered
most-important first (a

literary

criterion
you must then
(now) begin
to spend the whole rest
of your life
studying)

as if
every field has an extra
asterisk p.s.

where anomalies are noted

(how certain is the value?
why or why not?
when and why did it last change?
etc)

 
 
 
 
.

the remaining gap
(to be bridged)
between the concept-system of mathematics

and the full semantics of literature

might be smaller than it appears



partial-vs-complete sets
(many human stories hinge on
sets with missing eleements
eg the overlooked opportunity
the missed clue)

mappings between sets
(literary allegories
microcosms
graphic representations
mental models)

orderings
(chronological timelines
euclidean axiomatisations
most-important-first prioritisations (cf holograms)

cluster-sorts
object hierarchies

fractal thickets of relationships between
objects in the same hierarchy
(cf indra's web)

branching timelines
storycycles

timelines of changing priorities

clustering(s) of top-level priorities

 
 
 
 
 
 
.

strict reverse-chronological order
for blogs
is a semantically braindead ordering-scheme

so i used to put a lot of effort
into exploring other possibilities

especially
picking any promising topic
and winnowing search results
picking the most promising pages

and sorting those links by
type
of content

so the resulting index-pages
offered an implicit content-typology
simply via the clusters
of similar content-types
embodied in the resulting page layouts

eg for persons:
short bios
individual works
influences
gossip
etc

(if there's to be a microformat
it would have to be
a person-pages-content-types microformat)

 
 
 
 
 
 
.

at any moment
in any person's
mental model

some objects are
more important
than others

ie
mental objects can be ranked
by importance

rankings changing in time

and different rankings
can be averaged
narrowly or broadly
by various algorithms

locally important details
demoted
as averages are broadened
(other times
other people
other places
other species)

today's average
year-to-date average
family average
community average
national average
species average

these ?universally? important objects
are named in english
with words like

sex death ego money
love family brains body
beauty art communication
celeb gossip sports politics



novelists projecting their private priorities
onto realistic characters
with individual priorities

 
 
 
 
 
.

what can xml do with
For sale: baby shoes, never worn?

start with a short-story microformat
on top of a want-ad microformat

the 'characters' in the story
being the parents who placed the ad
and the lost baby
(and the newspaper that took the ad
and the readers the parents hope to reach)

the want-ad microformat
entirely ignoring the critical facts
(the baby
and how it was lost)

mode: selling
topic: thing-clothes-shoes
size: baby
condition: 9
price: unknown



we'd like the story microformat
to include a field for 'theme'
but the only
(ignored)
possible approach
is something like my solace

eg: solace/losing#baby

so we need experts in
art philosophy religion
politics economics history
fiction poetry humor
psychology philology anthropology sociobiology
criticism design logic math

to compile comparable storycycles
on all their areas of expertise...

 
 
 
 
 
.

[this is the first
and hopefully the
least
articulate and
most
controversial
of a projected series]

any semantic web first needs
a mental model of all possible semantic webs
(and this mental model must include
a mental model of all possible
mental models)

say

mentally model each 'object'
(including mental objects)
as a named point
of known or unknown duration
(some objects
more important than others
from the perspective of any given mind)

with lines between points
modeling relationships between objects

relationships changing in time

relationships mentally modeled
by related mental objects

relations affecting
and affected by
(more or less strongly)
related relationships with other objects

(all mental models of relationships
necessarily related to the mental models
of all related objects
and of their relationships)


and mental models differ in depth
in wisdom
some shallow some deep

deep or shallow
in many different ways

different wisdoms
characteristic of the deep or shallow
people who hold them

but mental models of semantic webs
and mental models of mental models
must be deep by every definition

constructed atop a distillation
of all wisdoms
(art philosophy religion
politics economics history
fiction poetry humor
psychology philology anthropology sociobiology
criticism design logic math)
only by the wisest of the wise

.

[i think hresume would work better with links to wikipedia articles on the schools and businesses, where available. this is a design experiment to see how much biographical content can be boiled down to annotations on wikipedia links.]

the format is: [link] [relationship] [dates] [keywords]

1953 birthyear

Springfield, Ohio birthplace 1953 mercyhospital

Yellow Springs, Ohio hometown 1953-1966

[Vernay Laboratories] job 1958?-1966 father

Antioch College job 1955?-1964? mother

Antioch College school 1958?-1962 labschool

folk dancing activity 1959?-1970? family

Cincinnati Reds fan 1960?-1965

Frank Robinson fan 1960?-present

Ken (Barbie) 1961 toy

Fred Gwynne creditedwith 1962? play

Midsummer Night's Dream 1962? actor

[Mills Lawn Elementary School] school 1962-1964?

Swingle Singers concert 1963?

ice skating sport 1963-1966

Malvina Reynolds met 1964?

[John Bryan Junior High] school 1964?-1966

Heathkits hobby 1964?-1968?

Minivac 601 computer 1964?

[del.icio.us home] [html]

[about]


[Wikipedia biography] [own template] [more]

education:
elementary
highschool
early reading
college1
college2
college3
college4
later reading

occupations:
The Farm
AI programmer


works:
neologisms
AntiMath
Solace: a textbook of romantic psychology

internet autobiography

website1 [sitemap1]

blog1 [blogmap1]

linkblog1
linkblog2

faves:
books [more]
music [more]
movies
videos

the semantic web movement is 8 years old
and has delivered little more than
a microformat for business cards [Wiki]

and nothing to rank with
roget or lenat

so i'd like to propose
using del.icio.us for prototyping
much more ambitious
microformat-like linksets

starting eg with
prototype person-profiles
[Jorn Barger]

replacing pagetitles with microformat fieldnames

using Wikipedia placeholders
if the fieldvalue isn't already a url

using del.icio.us tags
to sort fieldtypes



other possible prototypes:
website-profile microformat
webpage-profile microformat
blog-profile microformat
video-profile microformat
song-profile microformat
porno-profile microformat

(each of these should be
summarizable
via heraldic barcodes)

metasynthesis (4)

polti (12)

semanticweb (20k) semweb (7k) semantic_web (3000)

web20 (24k) web30 (7k) web_2.0 (6000) web2.0 (5000) web2 (4000) web3.0 (3000) web3 (500)
web2theory (35)

folksonomy (14k) ontology (11k) taxonomy (8k) folksonomies (3k) ontologies (2k) taxonomies (1k)

simulation (17k) sims (5k) sims2 (4k) simulations (2k) thesims (400) simstheory (0)

tags (33k) tagging (31k) tag (4000) metatags (750) metatag (400) tagtheory (1)

del.icio.us (40k) delicious (24k)

twocultures (50) 2cultures (2)

metadata (21k) meta* (18k) about (9k) meta (5000)

microformats (7k) microformat (2k)

patterns (48k) patternlanguage (200) patternlanguages (50) patterntheory (11)

informationarchitecture (4k) informationscience (900) informationtheory (750) infosci (500) infoscience (200) infotheory (150)

thesaurus (3k) thesauri (900) thesauruses (22)

conjecture (200) conjectures (20)

suggestions (1.6k) suggestion (1.4k)

anomaly (1k) anomalies (300)

cute (29k)
rant (6k)
brilliant (4k)
creepy (3.6k)
heh (3k)
lame (1.8k)
snark (1.6k)
snarky (500)
doh (500)
classy (400)
dorky (300)
tempting (20)
dontmiss (20)

document* (19k) documents (11k)
bestpractices (8k)
data (78k)
file* (27k) files (16k)
archive* (61k) archives (18k)
toread (370k)
database* (117k) databases (18k)
post* (30k) posts (7k)
message* (7.6k) messages (2.6k)
etext* (2.4k) etexts (1.2k)
ebooks (45k) ebook* (35k)
posting* (2.6k) postings (630)
thread* (7.5k) threads (6400)
conversation (5k) conversations (1.5k)
story* (31k) stories (26k)
biography (12k) biographies (2k)
books (354k) book* (183k)
email* (101k) emails (2k)
comments (17k) comment (13k)
feedback (5k)
dialog* (1.8k) dialogs (250)
discussion* (21k) discussions (1.9k)
debate* (11k) debates (1k)
poem* (8k) poems (7k)
lyrics (32k) lyric* (2k)
songlyrics (115) songlyric* (2)
songs (18k) song* (16k)
shortstory (1284) shortstories (1k)
()
()

()
()
()
()
()


tag bundlers: galfridus glass

smartipants: timbray

most web2.0 content
(eg blogs, flickr, librarything)
is explicitly autobiographical

documenting your life and tastes
and arranged according to
your personal calendar

so it's not hard to imagine
some web2.0 host
(eg facebook)
supporting a personal
autobiographical chronology
for each registered subscriber

that unifies/collates
all these separate
content autobiographies

with anecdotes photographs maps
lists of books read
movies seen
songs heard or sung
clothes worn
food eaten
objects owned

but all these categories of content
and most of the content itself

had its own separate existence
before it first crossed your personal path

(often before you were born)

and wikipedia should offer
a universal
chronological
index/ table of contents

that your (eg) facebook autobiography
should automatically recognise
as a secondary chronological sorting

"in 1973 i read a book
written in 1799"

"in 1987 i saw a supernova
that had occurred 168,000 years earlier"

"in 2001 i witnessed a melodrama
that's been playing and replaying
with infinite variations
since the evolutionary origin
of territorial aggression"

with every human concept
traced ultimately
to its evolutionary source

often billions of years in the past

while this deep
chronological thesaurus-sorting
(or etymogeny)

is barely dreamt of

 
 
 
 
 
 
.

starting
from the worldtree

and its
individual lifelines

isolate any
line segment
and flatten its spatial dimensions
down to just x
(with time as y)



paint this segment
using any convenient mapping
(eg the crayolacube
red-evil blue-sad
yellow-sick or -enslaved)



classify your segment by
start-color
and end-color



ignoring intermediate values
(here greyed)

notice the formal analogy
with the i ching's 64 hexgrams
permuting eight trigrams



each colorchange
within any segment
must have
causes

some expected
some unexpected



now
isolate any pair of segments
that start and end
at the same times

and classify this pair
by its two start colors
and its two end colors
(8^4 = 4096, too slow to render)



#1 of 64



sometimes the colorchange
in one segment
will help cause the colorchange
in the other segment

via
a strong or weak
linkage

strong

weak

 
 
 
 
 
.