Between Theory and Practicetag:www.iskm.org,2010-02-04://12016-12-03T07:18:45ZMovable Type 5.2.10Light Censor @iPhone - Between Theory and Practice - EBlogtag:www.iskm.org,2013:/blog//6.292013-07-10T13:33:41Z2016-12-03T07:18:45ZiPhone has a light censor, and it's smar...Kenji Maruyamahttp://www.iskm.org/
// get brightness
[[UIScreen mainScreen] brightness]
// set brightness
[[UIScreen mainScreen] setBrightness: 0.2f]
]]>
iOS / Android Version Distribution Pie Chart using d3.js - Between Theory and Practice - EBlogtag:www.iskm.org,2013:/blog//6.282013-05-11T21:35:48Z2013-05-12T01:51:35Z iOS / Android Version Distribution 201...Kenji Maruyamahttp://www.iskm.org/
iOS / Android Version Distribution 2013/2 source: chitika
]]>
]]>
objectiverecord - Between Theory and Practice - EBlogtag:www.iskm.org,2012:/blog//6.242012-04-20T22:58:19Z2012-08-02T12:07:46ZI made a O/R wrapper for Objective-C and...Kenji Maruyamahttp://www.iskm.org/ // create table, only 2 line
EntityManager *entityManager = [[EntityManager alloc] initWithDataPath:@"/Users/KENJI/Temp/test.db"];
Hoge *hoge = (Hoge *)[entityManager create:@"Hoge"];
// insert and update, only 2 line hoge.count = 3;
[hoge save];
Don't you think it's very simple ? SourceCode is @ github. If you have an interesting in this library, check it out.
]]>
Twitter Picture Search - Between Theory and Practice - EBlogtag:www.iskm.org,2011:/blog//6.212011-01-19T02:01:55Z2011-01-25T17:08:28ZTwitter Picture Search is a simple mashu...Kenji Maruyamahttp://www.iskm.org/
Twitter Picture Search is a simple mashup service. I launched it for searching picture everyone tweet from everywhere. You can see the realtime search result on GoogleMap.
The following figure is a example when I input "snow" in Japanese, the service shows me some of pictures that somebody just tweet . ]]>
Twitter
Twitter Search
Google Map
twitpic
plixi
Yfrog
Ow.ly
movapic
Mobypicture
img.ly
Twitgoo
pic.im
Enjoy searching Twitter picture!
]]>
Analysis of your friends of friends on twitter - Between Theory and Practice - EBlogtag:iskm.org,2010:/blog//6.112010-02-20T20:10:13Z2011-01-05T02:00:47ZWe analysed friends of friends on twitte...Kenji Maruyamahttp://iskm.org
The result was classified by four categories :
Popular Person
Active Person
Talkative Person
Watcher
]]>
]]>
Science of motivation - Between Theory and Practice - EBlogtag:iskm.org,2010:/blog//6.62010-02-10T19:40:46Z2010-07-07T01:21:40ZKenji Maruyamahttp://iskm.org
In 21th century, we are asked for surprising and not obvious solution in
our work. ]]>
A feature of Social Search Engine - Between Theory and Practice - EBlogtag:iskm.org,2010:/blog//6.52010-02-03T19:37:56Z2010-02-17T14:42:14ZKenji Maruyamahttp://iskm.org
The Anatomy of a Large-Scale Social Search Engine, the paper has just been accepted to WWW2010 by Aardvark. First of all, what is Social Search Engine? The aim of Social Search Engine is to find the right person to satisfy a user's information need. In their paper, they compares to a traditional web search engine like Google:]]>
Google came out of the Stanford Digital 'Library' project, but
Aardvark has roots in another age-old model 'the village paradigm'. In
'village', people use natural language to ask question, and may be
replied without delay by community member.
Content is in peoples' heads. Not finding web
site, but just finding right person who has your answer.
In traditional search engine, trust is based
on authority. In Social Search Engine, trust is based on intimacy.
Anatomy Model
where
u is user, q is question, and t is topic.
Whether their model is inspired by aspect model and earlier work in
search in P2P Networks or not, they expresses that it allows real-time
routing because of evaluating only p(t|q) at query time, others
probability function are computed by User Indexer and Social Graph.
At
indexing people, some precondition is strange to me.
If a
user has expertise in a topic and most of his friends also have some
expertise in that topic.
Using Bayes' Law to compute for p(u|t)
These
inferences are subjective.
What does they know person who has
most suitable answer for questioner? It depends on 1) how many words
user profile contains question topics, 2) in social graph, a degree
between user A and user B, 3)and who is available to answer quickly.
I
have two question after reading this thesis. Why we hasn't used chat
application like Skype to question person? What is our motive for
answering person who doesn't know?