steven johnson
ants brain cities software – self-organising system … not necessarly having command and control system.. .collective inteeligence from simple elements interacting
cities and blogs
cities – small pieces loosely joined… massively parallel decision
james wolcott quote – business 2.0
“our emerging superbrain could use a little work”
emergence doesn’t just happen? push towards the point where this happens… there is some design involved….
tainted history of all digital communities allied to urban theor (too literal… apple eWorld… not a visual metahphor) more about flows and layout and interactions.
what cities are good at:
clusters – legibilty, collective memory (cf. samina quershi http://voice.aiga.org/transcripts/presentations/samina_quraeshi.pdf)
public spaces diversity
optimal density (just enough activity to keep a street lively – but not too overcrowded)
CLUSTERING
great cities are good at answering “search requests” – you can find the areas devoted to your goals/tasks – clustering…
there is a danger that the web becomes a city with know neighbourhoods… how is clustering created?
combination of top-down and bottom-up structure. washington sq. plan on some level to make space open… but people come and populate it, and infect it with their partcipation, and their agendas. busy, bustling… people playing “freebird” on gutair.
optimal denisty – how does this happen. (cf. jane jacobs, chris alexander)
look at the road patterns in chelsea area of NYC… foot traffic channel to the large avenues. large avenues feel vital, have lots of big chain stores. but overcrowded but at night the croos-streets feel pretty deserted.
compare to west village – diversity of routes that anyone can take through the space – flow of foot traffic is spread. every street has something going on at any given time, but no street has too much… optimal density… no chain stores!
feels intuitively human – human scale of size and acitivty.
how do cities do it
bottom-up interactions (every time it’s planned heavily it seems to fail – sucks out the self-organisation)
the strollers create as much of the structure as the shopkeeps (flaneur theory – cf. Lucy Kimball)
passive organisation
the swerve – when you walk from x -> y and stumble across Z…. serendiptous reactions. e ngineer (people who like this book like this one – amazonian)
to paraphrase Lennon: “cities are what happen when you buzy makin’ other plans”
this a problem in car centric cities – you can;t swerve at 50 miles a hour…
THE BLOG PREDICAMENT
problems forming higher-levl groups
– how to visualise neighbourhoods of blogs??
manangin overpopulation
the tyranny of time
little passive organisation
readers lack input
-in structuring the blogsphere
catastrophic success
-great explosion of tols, and a crticial mass… but need to hit next level (more is different)
great citites are not just great towns – jane jacobs.
“the pedestrian centric model”
bloggers come to you – surf engine model
search requerst based on current documents – paragraphs you;ve written
search limited on chosen bloggers
key words categories created passivley by usage
no need to ask users or creators to create keywords actively (cf. cory – metacrap)
grows more organised with se
suers can creat permanent searches
clusters of like-minded blogs form around shared keywords
Panel Discussion:
clay – brings the POWER LAW!!!! how do power law relationships relate to cities – the way you travel around the web is not the way you travel around the city.
question – stores are next to each other to maximse exposure to prime audience…. there is no economic cost to travel to disparate points on te web? so how will clustering emerge
clay – mental transaction cost – of keeping track paying attention to diff sites… that is what is creating clustering in the web – link-organisation (daypop, blogdex were curiousities, now daily necessitiy)
geoff cohen – go there (to the cluster) gets you exposed to so much more other info in the cluster: mental outlay – cost/benefit. (the swerve) get the incidental association, the useful link organisation, clustering around you in your cluster.
johnson – big sites resisting what the web was good at – trying to keep you on their sites. surfing happens cos you are interested, not cos you are bored. not like zapiing channels on the tv.
non-player chracters – real – lots of people don;t know tey can particapte yet… blogs make it easier for ordinary people to participate
Geoff Cohen
the problem of metaphor -a bridge is a good metaphor for a bridge.
technology better at extremes of performance – SR71s fly faster than balckbirds
technology moves quickly
nature evolves slowly
nature adapts well to complexity
a few simple rules
– embrace legacy (because you don’t have a choice)
– try everything (diversity works)
scan the large, multidimensional solution space (genetic alogrithms – don’t get infatuated with specific techniques – sometimes you just need to be smart about searching and make big leaps) some times if you ape biolgy you can ape the wqrong model, or we make the wrong biological observations.
– Pay attention
look to the outside world. software tofday is basically dead – the walking undead – zombies not demons… make software like a living organism – that know whatis going on around it. what is changing it’s circumstances. part of this is metrics… fitness (cf. google methodology of good click/bad click)
– Reflection.
self-examination, self-diagnosis. all most no systems have good reflection at this point.
necessary for software to see what it is doing and change. code needs to become more plastic.
– change
paying attention to what works, and amplify that… evolution is the process of amplifying success through feedback.
– be fruitful and multiply
distributed software seems to have a better chance of evolving??
group at MIT – amorphous computing research.
Don’t need the biological side… we should be able to aply biological princioes today to build better software.
clez virus… the bad guys are already doing this (viruses infecting viruses)
– stay loose
emit and accept (cf. POstel/tim’oreilly talk)
– work bottom up
termites build termite mounds… make termites poster-child of emergence, not ants!
– plan top-down.
don’t be bio-purist. humans do things that lots of nature can’t.
– rest in piece.
let things die. we can;t handle software death. how can we account for what happens when software should or does die. plan beyond a year?
10 principles – probably wnrg, evolve over time with feed back.
what is roadmap?
accept complexity. software is emergnet already – we call them bugs! it’s how complex things work.
reform the guild – major changes in how we teach software, may mean major changes in what “software” means. we inside the guild care, but real people don’t. “software artifact”? (cf. wombling, see further) need more people to be able to create sftware. to the extent that people who cannot read cannot read software… doesn;t mean everyone has to be a programmer. (turn of the centruy people assuming everyone had to behave like phone operators to use a phone… we just changed how phones were used) specifying goals, rather than speicifying procedure.
– abandon computATION
alan turing is the father of computing – maybe it’s time to move out of the house. how much further would we have got to if he was still alive (cf. morthogenesis)
– do we need a new language?
maybe not… (cf. benjamin l. whorf: whorf hypothesis – “the language you use changes the way that you think” – very true in computer science) address problems in the way your languiage pref makes you analysis and procedurally think. (cf. milton glaser)
– what we mst give up
“in all revolutions there must be terms of surrender” – things we hold close to our hearts…
-have to give up that performance is what we are aiming for
-that software is predictable and repeatable.
– will human readility go away (cf. cory) – how do you debug a natural system
– entering scary world of change – where computers say “maybe” get comfortable with “maybe”
– people are biological too… study our interactions…
cory doctorow
* ban the words “high quality content”
in the 50’s movie studios would let films be shown on tv. they were afraid of lsing their “revenue streams” walt disney broke the cartel by broadcasting his movies on NBC, to rasie money to build his thempark
al gore – information superhighway – hollywood invited to hearings they said “we are here to tell you there will be no-one there unless we put our content there”
10 years of innovation because hollwood didn;t want to put it’s movies online… people are the fundatmenal component of content on the ent, not movies…
digital tv take up in USA, FTC has been convinced by hollywood that high quality content is going to drive this.. . therefore no tamper-friendly devices (firewire, video cards, dvd burners) can be connected to US digital TV standards
* ban words “mission critical”
mission-critical mindset would have prevented email from being invented.
napster died yesterday – napster wasn’t reliable, wasn;t mission critical. 70 million subscribers in 18months… fasted adopted tech in the world. at ltigation point forced to go legal… get relaiable… reliable naspter sucked. no one wanted it. zombienapster.
the natural order of things as seen by business and legal is not the natural human order of things.
lots of great exmaples of grown-up “high quality content / mission critical”apporaches have been foiled by kids plaing with unrealiable stuff.
humna appraoch to catalogoing internet – if net keeps growing then every human would have to be indexing for 18 hours a day every day to supply metainfo.
google sidestpd this… distributed bottoom up, and as a result has killed all other searchengines and directories… even powers yahoo.
CNN vs. the distributed republic of Blogistan.
9/11 bloggers were sometimes faster and more reliable than trad. news orgs. (car bomb outside state dept.)
ban the word “optimal”
optimal for what? for who? usually for those who like reliability. optimsied. innovation is not optimal. the things that take us into the future are not optimal… by contraining the system to only foster that which is optimal, we hamstring our ability to go further.
turing machines are not optimal. universal machines not optimsed by defn. suboptimal, primordial goo tat we can make anything for.
legislative mindset based on optimisation means that Turing machines constrained.
TIVO is a great example…strapping a PVR and collab filter to TV was never envisioned by broadcasters… when you are reliable and optimised you close the door to innovation…
Petafloptimism 
