Hacker News Reader

Comment by mayank | Hacker News Reader

> This is the first (global?) distributed wireless network that evolved on its own.

Not really. It's emergent behavior in a complex system. You have a lot of entities (in this case, unpatched Windows XP SP2 and below systems), and a set of very simple local rules (if (winxp and no wifi) then create last adhoc network, if (human and see "free public wifi") then click link).

The fascinating behavior that emerges from the interplay of these local rules is the global "virus" if you will. It's a fascinating area of research, and even aside from the inflated egos of some physicists in the area, I'm glad I'm studying the field.

Read Duncan Watts' book "Six degrees" for a decent popsci introduction to the field (you might have to skip some self-aggrandizing of physics as the Savior of Intellectual Thought though).

Edit: cellular automata are a more recognizable CS example of this.

View on HN

That's not contradicting me? Evolution is a special case of emergent behavior in a complex system.

The rules of the buggy system mean that any ad-hoc network can be replicated if a user connects to it. That creates selective pressure on the name of ad-hoc networks in that convincing more users to click on them causes further replication.

So we have a (somewhat) self-replicating entity, with variation, and a selective process applied to that variation. Sounds like evolution to me.

(Of course the names are originally made by people; but if we assume that ad-hoc networks aren't named with the intention of creating a self-replicating pattern, they're effectively random from the perspective of the ecosystem. The success of "Free Public WiFi" is a product of selection, not human design.)

Evolution requires mutation. If the name occasionally got garbled then it would be evolution.

Evolution requires survival of the fittest. If the name got more enticing over time then it would be evolution.

Technically evolution does not require survival of the fittest or mutation. Evolution is merely the change in allelic frequency in a population over time.

Without selective pressure (i.e., "survival of the fittest") you can still have evolution via drift, the founder effect, etc. Drift is a very important source of evolution in a lot of situations (during my PhD I studied how a source of seeming drift in the coding regions of genes was actually caused by selective pressure in some cases).

Without mutation a population can still evolve, too - for example, if a disease struck and there was no mutation, an allele that protected against the disease would quickly dominate the population. This is a much more pedantic thing to point out (on my part), though, since every stretch of every organism's DNA mutates. The different mutation rates of viruses, bacteria, and humans due to replication errors (and the resulting effect on evolution) is a pretty fascinating topic, though.

Yeah I went back and read your comment again and see how you meant "evolution", which is perfectly valid (i.e., special case of emergent behavior). However, I know this is going to get increasingly pedantic, but evolution usually implies some sort of survival-of-the-fittest rule, which seems to stretch the definition here:

> That creates selective pressure on the name of ad-hoc networks in that convincing more users to click on them causes further replication.

I was thinking in a more formal mathematical sense. You can write a fairly short agent-based model (for example) to simulate the spread of "Free Public Wifi" and look for conditions where it remains suspended in the population. The analogy is closer to a virus than evolution.

But hey, I didn't mean to imply that what you were saying was wrong or --gasp-- a bad analogy in anyway, and I warned you that this was going to get pedantic. :)

I wonder if this virus has been able to effectively spread to Madagascar.

To put it another way: I wonder if "Free Public Wifi" was the surviving ad-hoc phantom in countries where the primary language isn't English.

Wow... that's a cool thought.

Even cooler is if you did manage to find a Madagascar that hadn't developed any widespread variant of its own either, you'd have the chance to design and introduce it. It'd be a neat social hack.

That's an interesting algorithmic problem. Assuming that you knew the structure of ad-hoc networks in terms of wifi range in a smallish region (for example, by wardriving in a dense urban area), you could pick geographical regions to target for introducing your local "adhoc network meme", which would just involve standing in strategically placed locations broadcasting that ad-hoc network SSID. That would vary with the geographical distribution of WinXP computers in the region. I think the whole social hack could be doable with detailed wifi maps of a region and a laptop. Wonder how much work it would be to implement.

Reply on news.ycombinator.com