THIS IS HISTORY! Check fabien.benetou.fr for news.

Seedea, scalable creativity

Xye, consultancy for serious creators

Information

RSS

(Updates)

SandIdeabox

Here ideas are simply gathered without any specifically required form or quality, simply unrefined ideas.

Eventually top selected idea will move into The OIMP repository.

New idea unique title

(template to follow)


Cognitive advantage of visualization

Fully distributed p2p

Cross protocol P2P sharing

Cognitive leverage

Invariants

Functionalities emergence detector

Cloud computing market auto selectors

e-resources (intangible economy resources) pool

Fundamental Concept Tree Balancing

Decision Tree Maximization

Geography of the intangible

Visualization based on geographical-mapping principle of the UniNetVerse theory

Internet Shortcut Layer

Seamless communication

Open Access Science

MiniNet

Visual Food Weight

Wiki as a learning tool

Social reading

Modern Biliography

Helping Sheldon

Internet guide

Serendipity RSS

Twitter Trends RSS

Beat Mining

sheclicked.it

Food feed

Game of the Mechanisms of the internet

Augmented reality heuristics system

Shift Space Wiki

Smart Panic

Natural screen ambiance

Information freshness

Text 2 Age

PR Detector

Mapping Tomorrow

Feeling Complexity

Learning "tip of the day"

Free visuals

Generated Entrepreunariat

Automated Exit

Wiki Logs: The Musical

Domotic video broadcast

iPhone context music finder

StartUp Analysis

Referrer suggestions

HugtoberFest

Smart answering machine

Dance code repository

NetworkedPulse

Weekly Startup Session


To move to Seedea.Todo

pmWiki based prototype

economy-based seedea based on pmWiki

IMified wiki

Live notes from AI94 in Palo Alto

Method

quality of an idea

Improvements for Seedea

Watermarking

Automatic Visual Moduling

To move to Seedea.Optmodules

(moved)


Utopiah's analysis articles

Moved in the dedicated's page : Utopiahanalysis.Utopiahanalysis


Build RSS flux for trend watching

for top research labs per area (add journals ?)

add wikipedia list of important publy RSS ? ex. : CS algorithmic_information_theory key papers


pmWiki based prototype

walking through/experiencing your own cognitive network explicitely thus re-inforcing your memory about it, allowing you to improve both

do I have all the required data within the current wiki for the visu I want ? do I need more ? let's try with just this !

! Idea to eventually merge with Seedea


economy-based seedea based on pmWiki

! to merge with seedea.internalmotivation


Cognitive advantage of visualization

my hypothesis is that our brain today is highly visual, we usual it for visual tasks intensively thus by it's very plastic nature it improved a lot in this area and it EXTREMELY powerful with visual manipulation. thus complex sets of information are more easily "workable" visually. (no magic, just because we are used to it, our brain remapped itself to take advantage of this ability) thus a cognitive scaffolding HAS to be visual. thus an element of it, a cognitive tool, has to be to. (Thus your NoI map should be visual even if it requires a bit more efforts initially.)

Example : military plane fighter visual interfaces (should look for research papers based on that the Visualization page of Seedea thank's to XPlane founder)

! Idea to eventually merge with Seedea.Visualization


Fully distributed p2p

Goal

Being able to rebuild networks after major failures (all node disconnected) in an automated fashion

Principle

avoid any fixed point

Structure

n = node
N = network of nodes
S() = shared strategy for all n in N, ability evolve
from result of previous N rebuild based on last used S
MP = list of meeting point computed by S() outside of N
MP() = list of trusted node in N 
including the node who created MP() on the support
(using encrypted steganography as a protocol, parasitic behavior
eventually injecting data in the most requested documents
thus having a positive behavior on targeted servers to parasite)

based on larger previous card(N)
S could determine a smaller set of results (MP)
or a bigger (MP) set to distribute the load

as long as (MP) is not empty it should function, 
eventually by falling back to less and less requiring versions of S

the used strategy would be shared only amongst peer of your trusted network (and evolve over time based on network topology evolution, alias add/removal of peers)

Critics

FreeNet guy :

"meeting points" are atm the seednodes but opennet and seednodes/meeting points are only a workaround as many people dont have friends running freenet to make it easier for them to connect

too much hassle for very little cases People do that outside of the network (online or offline) + possibility to use OpenNet to re-escalate to DarkNet. + if there would be an automatic way, there would be another point where attackers could compromise the network (compromising the connections between friends is much harder)

But he said it could be faster than man made version

Example

it's basically what guys do when they look for females :P (MP) = local bars

Could also be used for botnet communication or any material on which the author does not want to be traced back. By hiding information into famous material that several people will download, it will be impossible to find who download it for the normal viewing purpose or for the hidden data.

Links to exlore


Cross protocol P2P sharing

Goal

bootstrap new networks and facilitate transition from other network (eventually enable fallback in case of new network problems)

Principle

use abstraction layers and peers as connectors

Critics

BT guy : such things are frowned upon as most multi-network clients are a jack of all trades and a master of none

i.e. they're not acting as a good implementation on any of the networks they're connecting to

they have... issues. but since they only seed (i.e. contribute bandwidth) they can't pose much harm


Watermarking

"watermark" EVERY personnal idea with a link to the Seedea project (meme, visual, url, ...) for Co-Evo.net (or any other reachable reference that include stats) BEFORE SHARING !


Cognitive leverage

15/08/08

example of cognitive leverage :

require : you live in a cognitive demanding environment (cognitive abilities are driven by your environment)

by becoming a center you reconfigure the network around you Using CL to maximize the chances of being a "center node"

network center node

 exists even in purely distributed infrastructure
  it emerges naturally as it is part of the optimization process

constant "oscillation" between purely distributed and central nodes

the topology of the network evolves based on usage

 the network constantly self-optimize as it is part of a environment with resources pressure

(schema) emergence of center nodes re-distrubition of the load (alternatives appearing)

	mathematical equation of treshold that precedes center node emergence, 
	simulation of the "balancing"/"oscillating" state and determine frequency based on env.
	(using sin/cos?)

why do center nodes emerge?

 to facilitate growth in complexity 
  as center nodes embeded simpler nodes functionnalities (!)

An entity that can use cognitive leverage to be a center node

 can use it's position to gain an advantage by studying the network topology
  and thus gather information to favorize and anticipate the evolution
   and reposition itself as a center node again later

(schema of dependency on center nodes curves and their superposition over time)

The ability of a center node to transition to a better position

 depends entirely on its ability to read and understand ithe networks it relies on
  (the network that provide it's required resources)

Ex :

	you need to understand the map and to have a ticket to move where you need to go
	who is the competitions and what is the size of the targetted population
	what are the barriers to entry ,...

Copies will emergence over time in order to gain central position too

 alternative will stabilize the network with fallback nodes 

the dependancy level over time will is probably (based on what ?!) a bell shaped curve


Invariants

15/08/08

(ordered by lifetime?!) Strategies

 medium
  resources

PS : strategies are resources too but they "transcend" the rest (weird ... gotta reformulate better)

Example :


Functionalities emergence detector

15/08/08

Definition of a functionality

(inspired by biological evolutionary science + computer science) usable function provided by an organism

Potential solutions

Example

Emergence

Evolution of the size of the population P using a functionality F over time (supposedly growing)

Detector

Produces a visualization that embedded suggestions based on good practices.

Crawl

Based on identified sources

Compute statistics

Based on equations

Detect trends

Based on compared functionalities emergences (use other trends for calibration)

Suggest Based on good practices

(have to add the written text)

To explore

PS : SensorWeb 2.0: Service-Oriented Middleware for Heterogeneous Sensor Networks seems to be limited to hardware based sensor, not directly for "software sensors" as in discovering new WebServices, API, SaaS, ...

Inspiration


e-resources (intangible economy resources) pool

Computation cycles

NameEC2MS Live Cloud
Specialization  

Data storage (distributed physical hosting)

NameS3FreeNetGNUNet
Specialization   

Message transmission

NameXMLrequestemailFreeNetTor
Specialization    

Cognition

NameFreelancers outsourcingAmazonTurkSASSOnline APIs from vendors
Specialization    

Physical hosting

See Seedea.HostingMatrix

NameGandi.netCyberBunkerSeaLand
Specialization   

To link with


Cloud computing market auto selectors

Moved to CloudArbitrage


IMified wiki

Principle

enable "write through IM"

Technical mean

Call IMified API from a pmWiki PHP cookbook

(to integrate to Seedea high level I/O)


Fundamental Concept Tree Balancing

result of a nice "conversation" = mutation of your FCT to

...thus the more mind-blowing an experience is, the closer to the root the C situation happens.

A>C>B regarding effectiveness but it's more costly ! (especially since we think about info creation vs deletion cost based on the reference from the guy who made the conference on ... arg... the french guy... black PPT... somewhere named in the wiki, had to check this out)


Decision Tree Maximization

the objective of a strategy is to access a network of resources

a strategy can be represented by a decision tree

one tactic can be to maximize the "largeur" of your tree in order to have maximum freedom of move and minimize the decision tree of entities competing for the same network of resources.


Live notes from AI94 in Palo Alto

concatenate the whole wiki in a one page (done) and check the length (done) organize it refine it and find the core of ti

live alimentation/integration to my wiki/tool (that's what I did back then in AI meetup)

I can , based on statistics, determine the best environment for me to get ideas based on the curves over time that I produce based on my data

Add GPS data after on each interaction

measure the number of click to integrate the idea to the system

measure it in order to have data to optimize it

measure the number of clicks to get someone else to ACTUALLY get it.

judge and optimize it

The luck factor (study on randomize self define people who think they are lucky) -> added to Shelfari, also read the booklet version

Critic from Don Steiny (from ISNAE, Institute for Social Network Analysis of the Economy)

idea funnel

have bet on the ideas not only with people who BUILD them but with external feedback from REAL users

Method

  1. Do it = produce the content
  2. Automate the process
  3. Abstract the process

(NOT the opposite)


quality of an idea

number of time the idea has been succesfuly used by someone who intended to use it (even if it was initially intented for another use !)

thus user feedback is required
if some solution already exist check

(exemple with Agglom recently)

conclude if it was a good idea or not
eventualy decide to do it anyway if
  • market can get a newcomer (other idea non totally scalable)
  • realization is a bit different (functionnalities are actually different)
  • consider the barrier to entry

Improvements for Seedea

move those improvement to their own individual pages (use a wikilist)

offline version

1) select the data (/wiki/) 2) add the mini-server for the destination system (server recipe for windows) 3) package all this (tar.bz2 or zip with PHP) 4) make a Seedea online page to obtain this (php on the server) 5) make a synchronization script (use the 2 different Site.AllRecentChanges)

Visualization

1) add the answers from the interviews (in the notebook) 2) decide what information should use the most visual place 3) define a visual language that is understandable for the targeted audience 4) make visual for very few ideas and try them on the audience

Lifeline of an idea

Use Sparklines

Graphic are generated for each idea as pmWiki page and can thus be included. Priority in frequency to the most popular ideas.

  • lifeline graphics navigation could be inspired by Google Finance management of the mouse wheel to zoom in and out in time

Idea = existing structure

(an existing structure to harmonize and formalize) += author += keywords += lifeline

Every data is automatically generated but with human supervision for possible correction. If possible, the corrections SHOULD be learned by the system (long-term AI goal).

Inspiration

Going further


Geography of the intangible

Atlas of Cyberspace by by Rob Kitchin and Martin Dodge (ISBN 978-0201745757)

Site Web issus de Internet : géographie d'un réseau par Gabriel Dupuy (ISBN 2-7298-1099-4)

urls perso

To share with


Visualization based on geographical/mapping principle of the UniNetVerse theory

Moved to Utopiahanalysis.UniNetVerseVisu.


Internet Shortcut Layer

Principle

Managing the different layers of shortcut on the Internet

Objective

Managing the different layers of shortcut on the Internet detect shortcuts that website like Gmail provide and be able to detect any conflit with Vimperator bindings and also discover shortcuts and new website I browse

Example

Method

automatically detect which shortcuts does a website trap (without analyzing code with events keypress as it could be too heavy?)

To explore


Open Access Science

Principle

Offer access to quality databases to researchers who are not inside labs who do not have large amount of resources

Solution

  1. Offer directly usable solutions to promote
    1. Science Commons
    2. GetCITED
    3. Publish or Perish
    4. Science Gateway
    5. Zotero > 2collab from Elsevier, free to use but open? (to check)
  2. List the open alternatives
    1. Directory of open access journals : covers free, full text, quality controlled scientific and scholarly journals
    2. Ranking Web of World Repositories by the Cybermetrics Lab CSIC
    3. PLoS
    4. FirstMonday
    5. Open Access by Clyde Smith
    6. Academic Commons (dedicated to Liberal Arts and education in general)
    7. ScientificCommons.org aims to provide the most comprehensive and freely available access to scientific knowledge on the internet
    8. IN-TECH providing OA content on Artificial Intelligence, Robotics, Manufacturing and Operations Researchsince since 2004
    9. arXiv,e-print archive, Cornell University Library
  3. Movements
    1. Right to Research promoting Open Access
  4. List the closed sources
    1. propose access by any legal mean possible
    2. propose way to circumvent classical accesses
      1. legally if they exist
      2. illegally if not other solution is available
        1. shared accounts (cookie forging, pre-configured vpn, ...)
        2. pre-selected off-line material (regular batch download from the databases, ossibily broadcatching)
      3. be conscious that illegal websites already exist
        1. http://bbs.techyou.org/
        2. http://www.passfans.com/
        3. http://www.abcdfree.com/

Interface

  1. enter the URL of the domain
  2. get login information with percentage of successes
  3. provide feedback on your try

To study

Analysis/Critics


Seamless communication

Principle

Automatically connect yourself and stay connected to all your communication channel (IM/IRC/...).

Solution

  1. have a remote shell for screen/irssi/clients-with-chatbot(IMified with a Bloglines notifier using API for update, listsubs and getitems)/RSS(On Facebook news)/Twitter/... eventually with some basic regexp and action on events
    1. Use a tool like WikiMatrix but for shells in order to quickly compare prices/performances(uptime/connectivity/...)/features(languages/datastore)/security
  2. smooth non-stop transition from a laptop/PC to an internet tablet like the Nokia N810 with Android/Maemo (as in auto switching when my tablet connect to the home/work wifi)

Progresses so far

To do

Inspiration


Principle

Have a visual geo-strategical summary of the tendencies in AI. Eventually based on our AImatrix.

Expected result

Visual map of the host who publish the most cited article recently

Solution

Needs

Process

  1. determine key journal with http://in-cites.com/research/2002/july_8_2002-1.html (or the more up to date http://sciencewatch.com/sciencewatch/ but beware of Thomson-only bias!)
  2. extract for every lab the set of information
    1. name
    2. location
    3. date of last publication
    4. date of last update for each project
    5. ...
  3. plot each lab on the map
    1. add their information
    2. link labs based on cross publication
    3. fundings

Inspiration


(cognitive proxy)Seedea information overload (!)

http://article.gmane.org/gmane.comp.web.wiki.pmwiki.user/51646 + add pmwiki vim syntax file with vim online edition (pmwe in Python)


wiki as a learning tool

Final result

Alternative solution

Import with DERI Pipes to a learning system

data from

Cramarama seems to also have a Wikipedia analyzing bot that transform (not so accurately) content to flashcard based on the same principle.

Study time of last edition with Forgetting curve with SuperMemo, a tool that implement Wozniak's algorithm

or also The Read / Review / Review Graph

See also


MiniNet

Principle

You limit the access of Internet to a growing amount of websites based on your learning goals.

Solution

  1. you use a tool (like OpenDNS) to limit access only to website that relate to the course
    • A child can easily disable OpenDNS by reverting to default DNS nameservers (or by allowing computer to detect DNS servers automatically)
  2. when a child encounter a limitation and want to dwelve deeper you open access to new resources
  3. you study learning paths and discover effective patterns (based on the efficiency of concept aquisition) and you facilitate those paths for other children

Inspiration

To share with


Structuring chatbot

Principle

It's a presence on IM/IRC that propose you to use good practices to define a problem by discussions.

See also StigmergyLive in The Lab

Solution

  1. how do we get there first
  2. then from here to there
  3. etc...

Inspiration

13:01 < UtopiahGHML> explaining to someone else is very often an extremely good way to progress. You have to sumarize, clarify, modelize, formalize and overall you end up with an updated model of the problem even if the other person does not provide "knowledgeable" feedback. 13:02 < brilanon> yes! 13:07 < UtopiahGHML> brilanon: you should write a chat bot to help you to improve this soft by discussing with you (questions on new terms, checking possible logic incoherence, etc...) and that you could code-update live with simple commands 13:09 < brilanon> UtopiahGHML, you are right, a bot that only reframed issues would be good, lol 13:09 < UtopiahGHML> notice me when it's ready :)


Visual Food Weight

Principle

It displays aliment (processed or not) with size proportionnated to the calories it provides (eventually within the aliment it's proportion of glucides, lipides and proteins by slices of different colors).

Solution

  1. have a DB of aliments with their constitutions and calories
  2. have a DB of visual representation (scalable thus probably vector based) for each aliment
  3. offer a way to naviguate through this newly generated viual database and offer tools to compare or to generate menus

Inspiration


Automatic Visual Moduling

Principle

Automatically configure the shared display based on the interaction situation

Solution

  1. prepare predefinied visual modules
    1. 1 person >O
    2. 2 persons face to face >O< or >O O|O O< or...
    3. 2 persons remotely
    4. N persons face to face
    5. N persons remotely
  2. try to detect the configuration based on geolocalisation and number of simultaneously connected users (and also the number of devices!)
  3. lively propose to change from one visual module to the next

Inspiration


Social reading

Principle

cognitive overload recovered by social means

Solution

using some social feed rinsing tool? you and trusted friends rate item within your feed client (online or standalone) then you choose to read only well rated feeds (or they would just appear bigger)

  1. mashup with bloglines+greasemonkey+website
  2. then from here to there
  3. etc...

Inspiration

Going through feeds where writters just post

Alternatives


Modern Biliography

Moved to Oimp.ModernBibliography as a full OIMP project!


Helping Sheldon

Principle

Yahoo! Answer themed on The Big Band Theory

Solution

  1. provide a space to share help
  2. add problem from the last episode
  3. earn money from specialized advertisement
    1. including Think Geek and other for-profit organization that target our audience
    2. but first generalist like GoogleAds especially since it requires no "official structure) limiting cost
  4. vote for
    1. the problem of the week (since it's a weekly serie)
    2. the next new feature (offer donations of cash and work)
  5. use previous work from
    1. snap2objects
    2. Seedea
    3. StigmergyLive
    4. ...and reference them explicitly
  6. eventually some ideas could be fed back to the author of the original show, using it as a way to remunerate us and the users and also gather a bigger crowd (creating events)

Copyright

Using the Sheldon persona is part of the questions to clarify but as long as we don't post long part of the show, based on Fair use, it should be ok.

Eventually being bought by CBC would be a quick easy solution

Inspiration

Discussion with Paola regarding the serie and its audience based on the idea of Marcel Mauss : The Gift based and the Xmas episode

Similar projects


Internet guide

Moved to Oimp.InternetGuide.


Serendipity RSS

Principle

an serendipity through RSS system to avoid to circle around in the websites you already know

Solution

  1. you subscribe to an RSS feed that will suggest daily a certain amount of new websites
    1. totally random
    2. based on your preferences
      1. explicit tags (input)
      2. OPML file
      3. url of your personal blog/website
      4. Facebook groups you've subscribed
      5. digging your delicious account for tags

Inspiration

Outils Froids on Serendipity

To explore


Twitter Trends RSS

Principle

Having a blog (with an RSS feed) covering the Twitter trends.

Example

having a 3/4 lines post on #spectrial to explain that it's on TPB

Running prototypes

WhatTheTrend by matthewmayer and also now Tagalus a covered by RWW (March 09)

Solution

eventually using Yahoo!Pipes to generate an RSS feed from

Alternatives

PS : this had been suggested to the official Twitter website (Ticket #32202)


Beats Mining

Principle

Find the most trendy songs based on your own taste according by mining the most recent sources.

Solution

  1. collect key places
    1. homepage of your favorite DJs with their playlist
  2. compute key indicators
    1. most famous this week (or other time frame)
    2. newest (fastest emerging)
    3. rarest
  3. collect suggestions from other key places
    1. labels
    2. online shops
      1. beatport.com
      2. nuloop.com
      3. junorecords.co.uk
    3. recommendation/suggestion engines
      1. pandora
  4. generate a playlist
    1. and for each song you have visual indicators (as a curve of profession) and keyboard shortcuts to browse extremely fast to
      1. rate it
      2. bag it/drop it
      3. comment it
      4. place cue points (might not work well with free extract)
      5. suggest it to friends (eventually making a network)
      6. export the playlist
  5. you come back to step 1 for a new cycle
    1. but this time with your history so that you have ONLY news

Inspiration


sheclicked.it

Principle

Know when somebody clicked a link you shared

Solution

  1. send the link to the website
  2. the website stores it
    1. it initiates the visits to 0
  3. the website gives you an 2 URLs back
    1. 1 to share
    2. 1 to track the information with an RSS feed
  4. you share the new URL to a friend through an email/IRC/...
  5. she clicks it
    1. the website that stores 1 visit
      1. the IP
      2. the time
  6. extra services to apply
    1. aggregating several feedback RSS
    2. statistics
    3. visualization
    4. geolocalization
    5. make link shorter and hide the AppEngine origin
      1. is.gd
      2. tinyurl.com
      3. ...eventually automatize and randomize this

Alternatives

Since then a lot of shortening services (including ow.ly from HootSuite provide close tool. Yet they don't give notification. Eventually a mashup with notify.me/notifixio.us could solve this.

Inspiration

I shared a link on IM to a friend who should open it only the following day. I found linkblip.com that provides a similar service (actually with "To view the exact time they clicked on this link, as well as their geographic location when they clicked it, please go here: LINK", it seems to even provide the extra, maybe just using Bloglines email generating option it could be enough)


Food feed

Principle

Let "edge otakus" enjoy their passion by directly delivering food but also automatically handling the command.

Solution

  1. list places supporting command through APIs
    1. IM+NLP could do the trick
  2. use the recommendation engine to pick a meal
    1. filtering out based on the history of commands
    2. filtering based on physician diagnostic (ability to monitor it if required)

Inspiration


Game of the Mechanisms of the internet

Principle

Guess for each website what is its core algorithm/mathematical equation.

Realisation

Overlay (eventually using ShiftSpace) on each website you visit to propose you N choices regarde the website function (its core added value, the "what you use it for"). The higher your score, the closer the potential choices thus the harder. Sets (association of website/functions) could be edited through a wiki as it's very simple to use even for a large mass of people. The content of the game/question would then be translated to XML/JSON to be queried by the client.

Example

I visit Netflix and the overlay propose me :

After picking I have the correct answer as recommandation engine a link to developp the topic a bit more. If I get it correctly, next time I visit it, the propositions will be potential equations that could be used by Netflix.

Notes

Our related projects

Inspiration

Thinking about supervized learning and the recently visited page on recommandation engine night held in early 2009 in Netherlands.


Augmented reality heuristics system

Principle

Display heuristics on top of daily life objects. The goal being to evolve more efficiently in one's own environment.

Example

Display the ideal soap/water ratio for the washing machine.

Solution

  1. identify your tools
  2. gather heuristics about your tools thanks to a global database
  3. display the heuristics
  4. provide feedback on your own results (worked/didn't work)
    1. for each positive feedback the system provide more "pro" heuristics

Notes

Complementary rather than replacing smart appliance

The advantage over localized embodied programming ("smart appliance" i.e. LG washing machine that can select the correct program based on weight) is that

They (AHRS and smart appliance) stay complementary.

Technical implementation

If glasses or overlay system stay too expensive, RFID + mobile phone could be a cheaper alternative.

Inspiration

Washing my clothes in a new machine ;)

See also

Our related projects


Shift Space Wiki

Principle

Automatically know if I already referred to the page I am currently visiting.

Solution

  1. parse my wiki where I annotate a lot of pages (videos, books, etc...) to generate a list of links pointing to the wikipage where it appeared (eventually with an anchor) + 1 line before/after
  2. use this list so that everytime I visit a page I already referenced on my wiki, a button appear on the page to point me to it

See also

Inspiration

Thinking about my notes on documentary I watched and how ShiftSpace could make me benefit from it more easilly.


Smart Panic

Principle

Running against the swine flu!

Solution

  1. track the notifications with your GPS (eventually crossing multiple sources)
    1. also use '09 Swine flu outbreak dashboard from theAirDB or related mashups
  2. use pandemic spreanding algorithm with for each update calculate the closet safest point based on your current location
    1. have heuristics based on the transmission vector and the flux
    2. EpiSimdemics: an efficient algorithm for simulating the spread of infectious disease over large realistic social networks, 2008

Resources

Inspiration

Chatting about H1N1 debacle (+finally make the SemanticWeb useful! :-# )


Natural screen ambiance

Principle

using the light up/down to simulate dawn/dusk on my laptop in order to keep a "natural" rythm

Solution

  1. call the light up/down button from the command line
  2. get the local time of dawn and dusk from an online web service
  3. apply the user configuration
    1. how fast should the transition go
    2. how many time can the user "counter it"

Inspiration

Documentary from the BBC on the "bodyclock"


Information freshness

Principle

Be able to quickly and simply visualize how fresh a piece of infromation is in a context you are not familiar with

Solution

Browser script (Greasemonkey) that displays dates with a colored background relative to how OLD the date is

  1. stick some CSS into the page
    1. today = css red, 1 week = css orange, 1 month = css yellow, 1 year and older = brown
  2. find the relevant nodes
    1. list of regexes to find dates either on userscripts.org
  3. having a mapping of date -> CSS class
  4. add said class

Inspiration

Reading yet another blog and having difficulty to assess how "fresh" the article was.


Text 2 Age

Principle

Approximate the age/generation of a writer based on the frequency of trends of rare words.

Solution

Build a linguistic analysis tool :

  1. gather trends of words per period
  2. parse the input text (or corpus)
  3. associate a weight/probability to each word
  4. order by weight/probability then deduce based on coherence of the most indicative (rarest with highest certainty and smallest period covered) words

Potential solutions

Inspiration

Reading aphorisms and finding "hitherto" to intuitively guess the age of the writer based on that information.


PR Detector

Principle

Confirm an intuition about a potential link between a group and media relay.

Solution

  1. detect a trend in media coverage (fast growth/reduction of coverage of a specific situation)
  2. hypothesize on what interest group (based on the content of the news) have motives to get this information public
  3. measure the news reference volume of the news and the group
  4. see if there is a potential correlation

Inspiration

Reading Tous les Colombiens ne s’appellent pas Ingrid, Paola Ramírez Orozco, Le Monde diplomatique, 2009 and tried to use Google Trend (attemps 1 2 3 4) to quickly confirm/infirm the hypothesis of the author.


Mapping Tomorrow

Principle

Map of utopias/dystopias in science-fiction literature knowing that it is very often (always?) inspired by learning the state of the art in science and engineering then imaginating further..

Solution

  1. gather they key idea of a period
  2. organize them in a phylogenic fashion (?)
  3. display a tree of the possible futures

Inspiration

Discovering Blood Music in H+ mag article on possible futures and thinking about the schemas from The Future of Subjectivity Edward Miller

Our related projects


Feeling Complexity

Principle

pedagogical value of a little device to make one "feel" complexity

Solution

  1. input the student algorithm
  2. estimate its complexity (using Big O notation or sth similar)
  3. change resistance accordingly
    1. see Haptic technology / Force feedback
  4. ask the student to use the handle (a la Babbage engine) to make the computation
  5. offer potential changes and go back to step 1 to allow comparison

Inspiration

Babagge engine, cycling and thinking, note on sqrt(n) Grover's algorithm during La Revolution Quantique. Eventually André Malraux's quote "On ne s’appuie que sur ce qui résiste."


Learning "tip of the day"

Principle

Display next to the search box a learning box related to the last search saying "did you know that you could also do X like this Y and add the option O?". Then provide buttons to "try it now", "I knew it already" and "add to my tips basket"

Solution

  1. use a web overlay system
    1. GreaseMonkey
    2. ShiftSpace
  2. import a learning list
    1. urls per site
      1. list of pattern of action
      2. associated with text to display
  3. generate a user property
    1. empty
  4. display

Inspiration

Searching in Google in a probably sub-optimal manner


Free visuals

Principle

Use visual search engine that find similar pictures to find copyleft, creative commons, ... equivalent

Solution

  1. find repositories of free picture (ask Mao if the one he knows)
  2. feed it to a comparison of pictures
  3. make an interface be able to send or link to the picture one has and want to find a free equivalent to
  4. link to the result

Inspiration

Thinking about (finally) correcting my mouseless-autoscroll GreaseMonkey script


Generated Entrepreunariat

Principle

Automatically generated a company based on a classical model but dynamically.

Solution

  1. lists costs over time (use market valuation)
    1. workforce by ability, price and places
    2. material
    3. energy
  2. A is the fastest growing price of material
  3. R is the fastest dimishing price of material
  4. B is the lowest workforce cost
  5. C is the lowest energy costs
  6. R = P(A*B*C)
  7. determine P and optimize it

Inspiration


Automated Exit

Principle

Anticipate risky situation.

Solution

  1. PIM GPS system
    1. provide local coordinates
  2. generate a database of simulated disasters
    1. room
    2. building
    3. city
    4. region
  3. automatically cache highly successful escape plans
    1. regarding the area we are in
    2. ordered by the most likely disaster
  4. update risks based on life feeds
    1. trusted tweeter accounts
  5. get more precise routes based on the update information

Related resources

Inspiration


Wiki Logs: The Musical

Principle

generate a "symphony" from the wiki activity

Solution

  1. wiki activity
    1. AllRecentChanges
    2. History
  2. generate a "symphony"
    1. MIDI
    2. Vocaloid
  3. convert to audio format

sed wiki_activity "s/action/note/" > music_format | convert_to_ogg > singing_wiki.ogg

Inspiration

Remembering that MIDI is a text-based audio format and that previous experiment have been made on the "sounds" of neurones.


Domotic video broadcast

Principle

cheaper domotic video system relying on existing home video infrastructure (using multiple TV sets in house for broadcasting useful things instead of TV shows for examples recipes in the kitchen, morning weather in the bedroom, etc...)

Solution

What is required

Steps

broadcasting part of the controlling PC screen on the coaxial cable and controlling that PC over wifi with an iPhone I think it could make such a system really cheap

Inspiration

Having to install the TNT box in Langrolay and knowing that I would actually not use the TV for classical programs that are broadcasted.


iPhoneMobile phone context music finder

Principle

Finding songs that are relevant to your current geograpical and time context.

Solution

  1. locate lyrics databases
    1. scap with 80legs
    2. list words about context information with DBpedia
      1. places
      2. time of the day
  2. index the scarapped songs per context information
    1. consider geoparsing Web services
  3. obtaining context information for the iPhone (GPS + timestamp)
  4. querying the freshly generated database to pull out
    1. list of songs from the more precise to the more general
    2. iterate over that list to play the first unplayed song
  5. a song you know is not present?
    1. suggest it and as soon as it gets validated, earn a period of free usage

Technology

"default" music player per platform

call it to play a song providing its title? the local filename? sth else? return code if the song can not be played?

See also

Inspiration

Passing by Paris during a very early morning trip to Bretagne and thinking about "Paris s'eveille", a song about Paris in the morning. Sylvain's description of the running app.


StartUp Analysis

Principle

Display on top of start-up websites a short summary of what they are actually doing as seen by others (bypassing buzz marketing to focus on internal mechanisms).

Solution

Inspired by


Referrer suggestions

Principle

redirect on different pages based on the Referrer information

Solution

suggestions and patterns are stored as wiki pages, making it very practical

Example

Inspired by

Editing yet another wiki profile pointing to my personnal website and wondering how 2 totally different persons could still find it interesting without feeling lost.


HugtoberFest

Principle

A fest for hugs.

Solution

  1. propose it to social entrepreneurs and related initiatives

Inspiration

German Oktoberfest and Free hugs campaign


Smart answering machine

Principle

Leverage the set of tools I use for myself in order to improve the classical static and not so informative "Sorry, I am currently away."

Solution

Tests

Remarks

Inspiration

Going to the pool and imaginating my answers based on patterns


Dance code repository

Principle

Provide a repository of organized dances in an algorithmic form. A dance is, after all, just an algorithm : step 1 move foot here; step 2 move other foot there; repeat X times; etc... A step by step process is by definition an algorithm.

Solution

  1. define a format
    1. Dance Notation Markup Language (DNML)
    2. XML-encoded STEP (XSTEP)
  2. import new dances
    1. curated manually
    2. generated through BHV files
    3. tag each by
      1. estimated date of apparition
      2. inspirational dance (inherance)
      3. geographical origin
    4. provide visualization using for example MikuMikuDance
  3. find existing similar community-driven website

Proposed at http://en.wikipedia.org/wiki/User_talk:Ohka-#Dance_code_repository

See also

Inspiration

Wondering about the importance of new activities and that the enjoyable ones are usually those in which you learn and progress, including dance classes.


Networked Pulse

Principle

Share your pulse with somebody far away from you.

Solution

  1. pulse watch (usually for pro runner) measures your heart beats
  2. send the signals with bluetooth to your mobile phone
  3. mobile phone send this as data packets over the Internet
  4. other end wear a bracelet with a little round picot (like the one use for computer screens for blind person)
    1. when it read signal it pushes up

See also

Inspiration

Reading about social predictions.


Weekly Startup Session

Principle

Increasingly allocate each week 1h of a day to create a startup pitched to a friend.

Solution

  1. have an idea solving a problem
  2. find a friend you would appreciate working with and has the competences to push the idea forward
  3. take 1h the next week and discuss about it
  4. at the end of the session you have 5min to decide together if you
    1. add 1h because you both want to allocate more resource to it
    2. remove 1h because at least one of you is not convinced and think there is still time to be slow
    3. maintain the current session length because at least one of you is not convinced and think there is still time to be slow
  5. if participants are not sure about what to do with current session follow a default plan and using templates (Blue Ocean Strategy, Business Model Generation, ...)
    1. principle, values, goal
    2. roadmap
    3. targets
    4. financing
    5. prototype
    6. business model
    7. ...

Inspiration

Wanting to make Augmented Reality Heuristic System (ARHS) work partnered with a busy friend. Also considering work on Genifer with its author.


New idea unique title

Principle

What does it do and why it matters

Solution

  1. how do we get there first
  2. then from here to there
  3. etc...

Inspiration

How did you get the idea?


Page last modified on July 05, 2010, at 11:46 AM