Disqus - Latest Comments for jfarmer

Re: High-order map implementation in JavaScript

jfarmer — Mon, 12 Oct 2009 14:00:33 -0000

map is a specific case of inject and can, in fact, be implemented as a call to inject. inject can be used to implement other things, like Array#sum.

http://gist.github.com/208582

Re: Yahoo should invest in products, not advertising

jfarmer — Sat, 26 Sep 2009 18:02:40 -0000

Yahoo! Mail is still way larger than Gmail.

Did you know Yahoo! IM has more monthly active users than Facebook?

Yahoo! has tons of assets, it just doesn't know how to make use of them.

Re: Yahoo should invest in products, not advertising

jfarmer — Sat, 26 Sep 2009 15:18:31 -0000

That's really unfair to the engineers inside Yahoo!.

Re: Twitter is retaining more users than Nielsen thinks

jfarmer — Tue, 19 May 2009 19:00:16 -0000

No, it does not. Those users will show up as non-retained users. If they never tweeted, they won't show up in the 1+ tweets group, though.

Is there a reliable way of identifying these accounts? I can filter the data to exclude them.

Re: Twitter is retaining more users than Nielsen thinks

jfarmer — Tue, 19 May 2009 17:31:28 -0000

It is based on the source. Their second report just included more sources.

Re: Twitter is retaining more users than Nielsen thinks

jfarmer — Tue, 19 May 2009 15:12:23 -0000

Kyle,

Nielsen released an updated report with those third-party apps included, here: http://blog.nielsen.com/nielsenwire/online_mobi...

According to them the numbers remain the same.

My question is, what is their methodology that they have to whitelist all the sources of Twitter traffic? How would they know they aren't missing some key source of traffic?

You can just get the data directly from the Twitter API.

Of course, Nielsen has a vested interest in protecting their proprietary methodology, so nobody is ever going to be able to reproduce their results. We'll just have to take their word for it.

My data, OTOH, I'm going to release under a CC license once I find a place to host it for free. You can also conduct the experiment yourself by generating a random list of Twitter IDs and fetching the data using the Twitter API.

Re: Erlang: A Generalized TCP Server | 20bits

jfarmer — Sun, 17 May 2009 12:29:31 -0000

The stubs at the bottom are there because I didn't have time to implement them and I didn't think they'd improve the example.

But the gen_server behavior has a lot to offer, especially if you want to make your server more robust.

Re: Erlang: A Generalized TCP Server | 20bits

jfarmer — Sun, 17 May 2009 12:23:33 -0000

The loop is stored in the State#server_state record, which gen_server passes to each of its callbacks.

Re: What Verna Taught Me

jfarmer — Sat, 09 May 2009 00:00:47 -0000

That's a good story. You should write about it w.r.t. your work on OpenID!

Re: The Dangers of Genetic Optimization

jfarmer — Fri, 08 May 2009 18:49:51 -0000

My point was that you didn't learn anything about your customers that can be applied in a different tactical situation. You don't know anything about their motivations, their interests, or their state of mind.

So, although you get one result (increased revenue) you also lose something of value (verifiable knowledge about your customers). There's a tradeoff there, and perhaps you're willing to make it.

Re: What Verna Taught Me

jfarmer — Wed, 06 May 2009 15:16:56 -0000

Haha. Chopin was huge, so it was more about the process than any specific problem.

I changed how I approached it and spent more time talking with the front desk clerks to see what they needed before I started coding away. Since then that's what I've always done, because otherwise I run the risk of wasting my effort and my customers' time (and possibly money).

Re: Social Network Notification Strategies

jfarmer — Tue, 05 May 2009 23:08:29 -0000

Sagar,

Thanks for commenting. That does look interesting!

It looks like their project is mostly about instrumenting a network to detect cascades. Is that right?

Re: Social Network Notification Strategies

jfarmer — Tue, 05 May 2009 19:24:24 -0000

Hey Matt,

Glad you liked it and thanks so much for commenting.

A more general notion of influence is captured by eigenvalue centrality, which is similar to PageRank — every person's influence is proportional to the sum of their follower's influence. This sets up an eigenvalue problem, whose principal eigenvector contains the "influence" of each of the nodes.

The nice thing about Kempe's algorithm is that it sidesteps all these issues and just asks the question directly: if we influence this user, how many other people can we expect to be influenced as a result of the cascade? That seems like a pretty objective measure of influence, although it's computationally difficult.

The downside to both of these is that they require extensive knowledge about the underlying graph and diffusion processes, which we don't always have access to. I've sent an email to Kempe asking about this. I hope he responds!

Re: Close to the Problem

jfarmer — Sun, 03 May 2009 16:18:53 -0000

This is related to the principle of 現地現物 (Genchi Genbutsu), part of the Toyota Production System. Roughly it means "see for yourself" or "go to where it's made."

The idea is that you can't understand the problem until you experience the problem. Reports, memos, etc. are at beast approximations of what is happening in reality, and at worst distortions. When in doubt: go and see for yourself!

Re: Let The Students Teach

jfarmer — Mon, 27 Apr 2009 19:03:40 -0000

In mathematics there's something called the Moore Method: http://en.wikipedia.org/wiki/Moore_method

It's usually used to teach topology. You start with the basic axioms of topology and with a little direction let the students guide the course of the class.

A quote of Moore's: "That student is taught the best who is told the least."

Re: Two Models of Behavior Adoption in Social Networks

jfarmer — Sun, 26 Apr 2009 22:56:55 -0000

Interesting. I'm working on a viral growth visualization project and was describing it to people as a sea of neurons. Diffusion of information or behaviors in the social network follows a lot of the same principles, as I understand it, e.g., all-or-nothing.

Do you have any information about actual formal models of these phenomena? The more math the better!

And thanks so much for the comment — really awesome.

Re: Two Models of Behavior Adoption in Social Networks

jfarmer — Fri, 24 Apr 2009 19:26:43 -0000

Indeed, and thanks for taking the time to read and comment!

Celebrities and the like are an interesting phenomenon in social networks. They are always highly connected, moreso on directional SNs like Twitter and MySpace, but the extent to which they are influenced or cause influence are much more ambiguous. Their effect on behavioral and opinion dynamics is probably a research topic all in itself.

And I actually wrote a Ruby script to generate BA graphs just the other day!

Re: Two Models of Behavior Adoption in Social Networks

jfarmer — Fri, 24 Apr 2009 14:43:07 -0000

Interesting. I don't know anything about statistical mechanics. Can you elaborate?

Re: Two Models of Behavior Adoption in Social Networks

jfarmer — Fri, 24 Apr 2009 14:34:34 -0000

Cindy,

Good point.

The way I see most social networks (and social apps) growing is by bootstrapping through the cascade model, until they're dense enough in their parent network that they become self-sustaining and threshold psychology becomes the dominant factor.

The risk with being too aggressive with invites early on is that you never build up the density to become self-sustaining. By promoting early adopters and keeping them engaged, you can help foster the core density from which your social network can grow organically.

FWIW, although Facebook bootstrapped by spamming everyone in Harvard, they grew with the threshold model in mind from day one. As I understand it they never even considered signing up a new campus until there were at least 40 people from that campus requesting accounts. And they started with smaller campuses where density as easier to create and sustain.

Re: Hypothesis Testing: The Basics | 20bits

jfarmer — Wed, 22 Apr 2009 19:37:06 -0000

You're right, I equivocated in my use of "p". At first I use it to mean the value of p under the null hypothesis, H_0: p = 0.50, but then use it to mean the measured value elsewhere.

I'll fix it.

Re: Almost Viral: A Hybrid Acquisition Strategy

jfarmer — Thu, 16 Apr 2009 02:30:50 -0000

Jeremy,

Thanks for taking the time to read my article and double-check my math! I really appreciate your insight.

Your response gave me a few thoughts.

First, fair point about market clearing price and CPI. How likely is it that you can hit that rate profitably, though? The top casual MMOGs are making like a $2 ARPU, right?

Second, throttling messages isn't always possible. On Facebook, for example, user-to-user notifications, requests, and news feed items have to be sent as soon as the sender issues them. I'm thinking of the iLike story where they drove around SF picking up servers from everyone and their uncle.

Still, you're right that this is probably the better solution for most people if they're just trying to keep the servers from starting on fire.

Third, viral growth leads can lead to demographic problems, too, which throttling doesn't address. By definition a viral process selects the users who are best at propagating, not necessarily best at monetizing. Controlling demographics with viral growth is much harder than controlling demographics with paid acquisition.

In any case it's something that needs to be understood, modeled, and tested.

Even if you don't agree that intentionally being "almost viral" is a good way to grow — and I'm really just presenting it as a distinct strategy with its own set of tradeoffs — I think it's important to understand the tradeoff and how the viral coefficient interacts

I couldn't find the formula I derived here anywhere else, so hopefully it helps someone!

Re: Almost Viral: A Hybrid Acquisition Strategy

jfarmer — Wed, 15 Apr 2009 17:56:15 -0000

I originally published a version with a bad formula, so Google Reader probably cached that one.

Re: Almost Viral: A Hybrid Acquisition Strategy

jfarmer — Wed, 15 Apr 2009 14:52:08 -0000

Glad you liked it!

I guess that makes me a social media pornographer?

Re: http://blog.rubybestpractices.com/posts/rklemme/001-Using_blocks_for_Robustness.html

jfarmer — Tue, 14 Apr 2009 15:56:54 -0000

I like the idea of this blog and am looking forward to more content.

I'd recommend you install a syntax highlighting plugin, though, since there's sure to be lots of code.

Re: Easter Eggs in Facebook Chat | 20bits

jfarmer — Mon, 13 Apr 2009 15:22:26 -0000

"Easter egg" is the name people use for a hidden feature in software. In this case, there were several hidden smileys embedded in the Javascript.