friggeri

I've blogged an explanation of what it does and how it works over here : http://friggeri.net/blog/who-did-i-forget

I've added an explanation on the how/why on the front page. Thanks !

Thanks for your feedback !

1. I didn't want to add too much complexity to the app by going into details of advanced Social Network Analysis, but I'll think about adding details.

2. Good idea, I'll work on a screencast/walkthrough

3. I understand that, but I have no other solution: having access to the user's facebook friends is an absolute requirement for the app to run.

Nevertheless, thanks for your feedback.

Same thing here, I launched http://whodidiforget.com a few days ago, and after getting a little traffic from HN it fell flat.

Sorry for hijacking your post but my question is similar to yours: what to do when you're not on a first name basis with big blogs editor (and those blogs do not respond to email tips) and you don't yet have a large network ?

I blogged about the science behind the app here : http://friggeri.net/blog/who-did-i-forget

As I said on the other thread, we only use the underlying friends graph for speed reasons: if we had to weigh the edges between people, it would be very costly in terms of API queries, which take time.

I'll rephrase that: adding a weight to an edge is not hard, I actually having several heuristics on hand to do that. It's just really time consuming the first time you do it, because you have to load a lot of data in order to be accurate. Moreover it requires additional permissions, which I'd rather avoid.

Of course, that's be ideal, but it would take longer to compute the weights. The idea here was really to show that just by looking at friendships it's possible to find accurate results.

Unfortunately, the github repo is private because I can't open source the whole algorithm right now. But I'll definitely do a write up in the next few days.

The only information it uses is "who is friends with who". I'm not looking at messages/interactions/networks/existing groups/friends list. Just the underlying graph of friendships.

How about http://www.hnsearch.com/ ?

Given that my work as a PhD student focuses on social groups, I thought I'd launch a little website to showcase the applications of what I do. I'd appreciate any feedback you might have.

For those interested in the gory little details, all the analysis is done client side, in Javascript (as a matter of fact, the whole app is just a static website hosted on github pages). If some are interested, I might blog about the different aspects of the implementation.

The real trouble is that when you link to a hashified URL, you are actually embedding in your web page (an encoding of) the content of the page you are linking to. Think matryoshka.

I had pretty much the same dilemma six months ago, finally I went with the 13" Macbook Pro w/ 8Gb RAM and 512Gb SSD. It changed my day to day life (but I also need my machine to do some heavy data crunching while on the go, so the slower processor was a major no for me).

Got it down to 118 bytes ;)

    function tm(I,t,e,s,i,c,k){for(i=0;s!=e;i+=k.m,s=k.n){k=I[s][(c=t[i])||"B"];if(!k)return k;t.splice(i,1,k.w)}return t}

You could actually save one more byte by changing the name of the function, but that would be cheating, right ? ;-)

Edit, Refined it to get to 110:

    function tm(I,t,e,s,i,c,k){for(i=0;s!=e;t[i]=k.w,i+=k.m,s=k.n){k=I[s][(c=t[i])||"B"];if(!k)return k;}return t}

Related to this, I'm a PhD student from INRIA[1] and as part of my research on social communities I launched at the beginning of the year an experiment called Fellows[2] on Facebook doing more or less the same thing as Katango. Except that Fellows was not an iphone app but a web experiment which aim was validate a new quantification of social cohesion I have introduced (the results were presented in a peer reviewed conference, the final manuscript is on arxiv [3]).

I have been wondering about the difference of media treatment Katango and Fellows received. I tried everything imaginable at the time (pitching large blogs, sending emails, asking friends if they knew someone I could talk to, etc.) and wasn't able to get that much press, whereas when Katango launched they were much more covered.

I'm not complaining, we managed to have ~2500 participants which was enough to validate our model, but I'm still wondering why in one case the story got picked up and in the second case not. Is it because they're a startup and I'm a student ? Is it a question of USA vs. France ? Could it be because I lack the good connections to get introductions to media ?

Sorry for hijacking this thread, but I'd really appreciate insight on what they did right/I did wrong.

Thanks !

1: http://www.inria.fr/en/

2: http://fellows-exp.com/

3: http://arxiv.org/abs/1107.3231

Shameless plug: http://fellows-exp.com We're a team of scientists from INRIA working on automatic community detection and we launched this experiment at the beginning of the year, already seeing great results.

I guess I haven't looked into Objective-C that much, basically just stopped after seeing all the NS* addons, I'll definitely take a longer look, thanks.

Thanks, interesting read. Will certainly follow up if I find more interesting stuff.

I should clarify : I'm not looking for an OO language or something with all the bloat of C++. Just a minimal superset of C with lambdas.

The fun thing about transpilers is that you can chain them together. Waiting to see some C stuff ported to Lua using Emscripten via JS.

"good programmer != only able to glue libraries together" != "good programmer == not able to glue libraries together"

How about Just Getting a sort algorithm Done ?

I disagree. Beyond the idea of sorting algorithm, what this shows is the inability to think algorithmically. A good programmer is not someone who is only able to assemble pieces of libraries together.

I'm not the author, just thought the idea was pretty nice.

It's great !

I just have two small requests. First, it seems you're using libdiscount (or objc bindings to libdiscount) to compile markdown, could you give the ability to pass flags to the compiler ? (e.g. activate footnotes). Second but less critical to me, I'd love to be able to tweak the CSS of the generated page (less critical because I can just tweak the style.css file in ORCDiscount).

If Mou had those two (small) features I'd gladly donate/buy it when it comes out.

You're absolutely right, I should have included from StackOverflow above.

The other way of looking at this data is that Android developers need more help [edit: on StackOverflow] than iOS developers. Which does not imply anything on popularity but tell something very interesting on the quality of the SDKs or developers… (I'm neither an Android nor iOS developer)

Really addictive, well done !

Keybindings are way off on non-qwerty keyboards.