The Social Networking Appliance

March 19, 2010 4:40 PM

Plug in a wall-wart to delete your Facebook profile! Eben Moglen, General Counsel of the Free Software Foundation, solves the problem of proprietary social networks.
posted by tybstar (52 comments total)

I think we'll have to wait for social network interactions to become a little more meaningful than Farmville and glorified e-mail.
posted by shii at 4:46 PM on March 19, 2010 [1 favorite]

Free software has won...But it is also in danger of losing, because those same services now represent a huge threat to our freedom as a result of the vast stores of information they hold about us, and the in-depth surveillance that implies.

Free software hates us for our freedom.
posted by KokuRyu at 4:49 PM on March 19, 2010 [1 favorite]

I think we'll have to wait for social network interactions to become a little more meaningful than Farmville and glorified e-mail.

And snarky status updates...
posted by Alexandra Kitty at 4:50 PM on March 19, 2010

I appreciate the work the FSF has done, but they're not doing themselves any favours with such dense prose in their interviews. 'Pessimal solution?' 'Pareto-superior outcome?' The message gets lost somewhat when you have to dictionary-check. And phrases like 'not actually a deal we can't do better than' are also hard to parse.

I do like the encrypted email and the 'VPN in a box' ideas; the more turnkey they make this the more use it'll get. And with the proper encryption setup, it's definitely more comforting security-wise than the idea of having it on a phone.
posted by Hardcore Poser at 4:59 PM on March 19, 2010 [3 favorites]

A cheap, wall-wart sized server for VPN and light web serving is a machine that I could very easily get plugged into.
posted by fantabulous timewaster at 5:03 PM on March 19, 2010

This must be some new definition of 'solved' with which I am unfamiliar. His "solution" is that everyone buy a $99 device, which needs a continuous source of power and a continuous internet connection that allows you to run a web server. Oh, and said web server needs to run a non-trivial, interactive web site or web service that opens it up to numerous potential security issues. Oh and the software for this doesn't exist yet and he offers no plan for its production except that, I suppose, it will materialize from inside the community.

The SheevaPlug uses 19W, so it's a non-trivial power consumer. It would cost perhaps $10-20 per year to run, depending on electricity costs. That, on top of server maintenance, is a significant burden on a typical household.

This plan is mind-bogglingly unlikely to work even ignoring the massive first mover and network effects advantages that Facebook and other social networking sites have. He tries to address the network effects by suggesting that social networking portability will become common. This is unlikely to happen because a) Facebook has almost unstoppable marketshare and b) social networking site users are 'sticky' and aren't actually interested in moving to another social networking system.

I could go on, but in summary he's basically wrong about every part of his premise and the conclusion doesn't necessarily even follow from his flawed premises.

Also, the article describes him as the former general counsel of the FSF, so a mod might want to correct the post.
posted by jedicus at 5:05 PM on March 19, 2010 [6 favorites]

I appreciate the work the FSF has done, but they're not doing themselves any favours with such dense prose in their interviews. 'Pessimal solution?' 'Pareto-superior outcome?' The message gets lost somewhat when you have to dictionary-check. And phrases like 'not actually a deal we can't do better than' are also hard to parse.

I started to say that, in his defense, Moglen is a lawyer and law professor, but actually a lawyer and professor should know all the more the importance of writing for your audience and using clear language.
posted by jedicus at 5:09 PM on March 19, 2010 [2 favorites]

I apologize for not being snarky and dismissive, but I like that he's fishing for an interesting solution, even if this doesn't seem immediately workable, the concept is really interesting. I hate the fact that I've put all kinds of (admittedly trivial) data into Facebook and don't have an easy way to extract it.

If anyone finds a decent, open social networking app that respects your privacy let me know and I'll try to convince my 300 friends to move there with me.
posted by mecran01 at 5:14 PM on March 19, 2010 [3 favorites]

Why did my brain keep insisting on reading "wall-wart" as "wall-mart"?
posted by oneswellfoop at 5:16 PM on March 19, 2010 [7 favorites]

And so I don't come off as a useless critic, here's my solution to the same problem: form a company that runs a social networking site that actually provides its customers with real privacy guarantees backed by a warranty. Not only is your data guaranteed to be private, but if your information is leaked because of our intentional acts, we'll pay you $x. If it happens because of our negligence, we'll pay you $y. We will spend up to $z on attorneys to resist government attempts to access your data through the legal system. People could pay for extra coverage.

The company would keep no server logs and offer no targeted ads. All information would be portable and deletable.

Basically there's no reason why a centralized website can't offer virtually all the benefits of the system he proposes while having the major advantages of a much lower price, convenience, and day to day security.
posted by jedicus at 5:19 PM on March 19, 2010 [3 favorites]

It may be smarter to think of this as a cheap, widely deployed network of TOR nodes, with a fractional horsepower HTTP server on it, rather than a drop in replacement for facebook.
posted by jenkinsEar at 5:19 PM on March 19, 2010

Well, I should say, no reason except that there's no market for such a site because people don't actually give a damn about privacy or free software (or even security until it affects them directly).
posted by jedicus at 5:21 PM on March 19, 2010 [2 favorites]

form a company that runs a social networking site that actually provides its customers with real privacy guarantees backed by a warranty.

Why bother when they can get as many users as are reasonably required without one? The market has spoken - real privacy isn't a feature enough people care about.
posted by GuyZero at 5:21 PM on March 19, 2010

His "solution" is that everyone buy a $99 device, which needs a continuous source of power and a continuous internet connection that allows you to run a web server.

Ooh, you mean like my Linksys WRT54G? Awesome.
posted by five fresh fish at 5:29 PM on March 19, 2010

Ooh, you mean like my Linksys WRT54G? Awesome.

No, I don't. None of the models in the WRT54G series have enough space to store a Facebook-type site, even if it only stored information for the one user. I suppose it could possibly store the barebones social network information, but you couldn't store any photos or videos or even a significant activity history.
posted by jedicus at 5:32 PM on March 19, 2010

jedicus: "b) social networking site users are 'sticky' and aren't actually interested in moving to another social networking system."

Friendster and MySpace will be delighted to hear this news.
posted by mullingitover at 5:36 PM on March 19, 2010 [7 favorites]

I think we'll have to wait for social network interactions to become a little more meaningful than Farmville and glorified e-mail.

I've found and reunited with several dear, dear friends that I lost contact with because I was a complete burnout in my early 20s via Facebook. One of those friends has lived with me for the last few months, and if it wasn't for us finding each other on Facebook she probably wouldn't have had a place to go. Every time someone says how useless and goofy social networks are, I can only say "maybe for you".
posted by DecemberBoy at 5:38 PM on March 19, 2010 [13 favorites]

Friendster and MySpace will be delighted to hear this news.

Friendster has 115 million registered users (i.e., its peak theoretical user base). MySpace has something like 125 million registered users. Facebook has over 400 million active users. Even if every single Friendster or MySpace user ever has moved to (or at least joined) Facebook, that would only account for a little over half of Facebook's active userbase.

Apparently "Over 90% of Friendster's traffic comes from Asia. In Asia, Friendster has more monthly unique visitors than any other social network," which further suggests that different sites cater to different markets, within which the users are indeed sticky.
posted by jedicus at 5:47 PM on March 19, 2010

Looking at their site traffic trends in Alexa, if both sites don't have a drastic turnaround they won't have any visitors at all by the end of 2010.

It seems a little fishy to pick out Asia as an indicator of social network penetration since Facebook is blocked in Vietnam and China.
posted by mullingitover at 5:59 PM on March 19, 2010

It seems a little fishy to pick out Asia as an indicator of social network penetration since Facebook is blocked in Vietnam and China.

Eh, I just went with the stats I could find easily. Given the nature of network effects and language barriers it does seem reasonable that social networks would sort themselves along geographic and linguistic lines, though. As another data point, consider Orkut, whose membership is apparently 51% Brazilian, 20% Indian, and 17% US.
posted by jedicus at 6:03 PM on March 19, 2010

Something something Orkut?
posted by Rock Steady at 6:07 PM on March 19, 2010

Crap. Should have refreshed before making a lazy joke.
posted by Rock Steady at 6:07 PM on March 19, 2010

You know someone did a study, and among those people who plugged in their Wall warts, there was a marked increase in mortality rate during the subsequent 6 month period, when compared to those who did not plug theirs in. Although some have suggested a common correlative factor, particle physicists working at CERN have theorized that the alternate life which some people have chosen to build on Facebook's cloud servers shares a quantum duality with the corresponding user's real life. According to the theory, when the virtual existence is terminated unexpectedly, there is a slight but significant probability that due to superdimensional coupling, an analogous effect will be produced in what we refer to as real life.

Critics said this theory was flawed because no such pattern was detected in the data belonging to other social websites, but the authors contend that nobody took any of the other ones seriously. "Well, nobody who mattered," one of them blogged.
posted by nervousfritz at 6:11 PM on March 19, 2010 [2 favorites]

This is awesome, although I don't really feel like it alleviates the problem at its source - the problem is that, just like piracy is a side-effect of copyright protection being a distinctly artificial condition of data, privacy also is. If a user can display information, they can save it. I'm just waiting for someone to write a Facebook API clone which saves (and versions, so you can see every change and blip) all the data it encounters - it breaks the TOS, but there are ways around that kind of thing.
posted by tmcw at 6:12 PM on March 19, 2010

Finally we are free of corporate control! Just plug this device into your router, which sends your private data to your friends through... um... a series of tubes, I guess? Which are owned by locally-owned mom-and-pop outfits... no that's not it... they're owned by a handful of multinational telecommunications corpora... oh, god damn it! So close...
posted by AlsoMike at 6:12 PM on March 19, 2010

I suppose it could possibly store the barebones social network information, but you couldn't store any photos or videos or even a significant activity history.

I thought that's what flickr and youtube and cloud storage is all about. I need bupkiss in the way of on-board storage.

As another data point, consider Bebo: dead SMN walking.
posted by five fresh fish at 6:27 PM on March 19, 2010

Hey, I solved the problem of global conflict. All everyone has to do is relax and be excellent to each other. Done!
posted by The Lurkers Support Me in Email at 6:35 PM on March 19, 2010 [3 favorites]

So instead of Facebook being able to analyze our habits from their logs...our ISP will be able to analyze our habits from their logs. How is this an improvement?

If you intend the data to be public, there's no problem with stashing it on someone else's server.
posted by miyabo at 6:43 PM on March 19, 2010 [1 favorite]

I thought that's what flickr and youtube and cloud storage is all about. I need bupkiss in the way of on-board storage.

How is keeping your photos on Flickr and your videos on YouTube better from a privacy point of view than keeping your photos and videos on Facebook?
posted by jedicus at 6:45 PM on March 19, 2010

jedicus: "Ooh, you mean like my Linksys WRT54G? Awesome.

No, I don't. None of the models in the WRT54G series have enough space to store a Facebook-type site, even if it only stored information for the one user. I suppose it could possibly store the barebones social network information, but you couldn't store any photos or videos or even a significant activity history."

So more like my DNS 323. Well, you'd still have to deal with a firewall, and an internet connection to handle it. And then you'd need new friends, that understand VPNs and PGP.
posted by pwnguin at 6:51 PM on March 19, 2010 [1 favorite]

Surely I can't be the only mefite who already has one of these hardware devices. I have a Pogoplug, which is very similar to the wall-wart described in the article. It sports the same processor, is relatively cheap ($99), and uses USB attached storage. In looking at the website, I see they have updated the latest model to feature a rather garish pink, but the technical specs are close the those of the unit I bought a couple of years ago.

And it's hackable.....
posted by grimjeer at 6:52 PM on March 19, 2010 [2 favorites]

No, I don't. None of the models in the WRT54G series have enough space to store a Facebook-type site, even if it only stored information for the one user.

Uh, yeah. But the storage could easily be added. An 8 GB micro-SDHC card is $13. And a 3.5" 500gigabyte drive is like $50. And $99 for a 1.5TB drive.

Storage costs are low. However, I think it's more realistic for everyone to buy raw computer time on something like EC2 then use it in their homes, which means your site would be unavailable if you lose power, your internet connection, whatever. Virtual servers are cheap, and you can encrypt the system contents if you want. Although I suppose being a virtual machine, it would be possible to extract the encryption keys at runtime. But that's far more difficult then bulk snorkeling up unencrypted data right out of a database. If the cops really want your data, they can always get a warrant for your house.

But either way, people should have control of their profiles, and be able to run them on whatever platform they wish: Home PC, cloud server, embedded server (like this thing) or have it hosted on an ad supported system like this.

Friendster has 115 million registered users (i.e., its peak theoretical user base). MySpace has something like 125 million registered users.

I’m a "registered" user on both of those sites. I haven't been on either one of them this year. I can't even remember the specific last time I even logged on to myspace. The number of "Registered" users is a completely useless metric.

The fact that facebook has 400 active users doesn't mean anything if something much better comes along. But people left friendster because the site took forever to load (almost a minute at the peak). People left myspace because it was fugly as hell. Facebook doesn't have those serious problems, but if there's a compelling reason for people to leave, they will.

But that reason clearly isn't "privacy concerns". Most people don't care, or are not and never will be aware of it. So, the only way to get people move off FB is to out program, outdesign, and out market them. That's something I think people should be working on, but I don't see anyone trying to do that.

So instead of Facebook being able to analyze our habits from their logs...our ISP will be able to analyze our habits from their logs. How is this an improvement?

Why do you think your ISP can't do that now? Facebook doesn't use HTTPS for anything but the logon. Your ISP can see everything you do on there if they wanted too. And in fact, ISPs DO analyze their customers traffic and sell aggregate data about what sites are most visited. How do you think sites like compete.com and quantcast get their data? 'Ole Alexa still uses it's toolbar, though.

So in fact its' not a question of "Facebook or your ISP" It's "Facebook and your ISP or no one, because the traffic will all be encrypted"

How is keeping your photos on Flickr and your videos on YouTube better from a privacy point of view than keeping your photos and videos on Facebook?

Uh, it's not. But something like Amazon S3 would be much more private, since Amazon is in the business of selling space, not selling advertisement, and S3 is used by corporate clients for "serious business". They're not going to sell you out. They'll respond to supenas, I'm sure, but that's get back to crypto, which you can use with S3 if you want.
posted by delmoi at 7:05 PM on March 19, 2010 [3 favorites]

If you intend the data to be public, there's no problem with stashing it on someone else's server.

Most people don't put their data on facebook to share it with "the public" but rather with their friends
posted by delmoi at 7:06 PM on March 19, 2010

And then you'd need new friends, that understand VPNs and PGP.

And $99-$135 for a SheevaPlug, Pogoplug, DNS 323 or the like, plus the cost of running it and time to maintain it.

I mean, seriously, Let's say you wanted replace Facebook with this thing. Let's charitably assume that the average Facebook user lives with 3 others in the same residence, so we'll need 100 million of these things. Let's assume Moore's Law and economies of scale bring the unit cost down to $50. That's $5 billion in hardware and $1 billion per year in electricity. Even if you assume that half of all residences could run the server on an existing machine, it's still an astronomical cost.
posted by jedicus at 7:07 PM on March 19, 2010 [1 favorite]

Uh, funny, because if you look at what I was asking in 2008, I was trying to think about shit like this:

http://ask.metafilter.com/86157/Coop-based-social-network-startup-advice

(yeah, people got mad at that post for seeming axe-grindy, but I think my point was pertinent, and with the rise of FB as the dominant social network and their powers, well I think it's kind of prescient)

The idea I had was basically p2p social networking. When I saw Opera's UNITE I figured this was the sort of tech we need. Honestly there's no reason for a separate device. But we do need some way to integrate server and browser. Networks should really be bottom up. Allow for discovery of friends through mutual friends.

It would take a lot of work to pull this off. But if the net becomes more totalitarian in the future/more commercialized/more of both of those... Well, I think ultimately, this may become a route that some people take.

It would never be a facebook killer. And yeah, it might be just like linux -- underground in popularity. But it may be strong enough to have resiliency against outside forces AND enough of a following to have sustained development.

We can get all snarky about his proposed solution (which I, too, think is a bit flawed) but I think his idea and desire really holds some good points that we would do well to heed.
posted by symbioid at 7:25 PM on March 19, 2010 [1 favorite]

Let's assume Moore's Law and economies of scale bring the unit cost down to $50. That's $5 billion in hardware and $1 billion per year in electricity.

What a bizarre argument. Yes, it costs money to run servers. It also costs money to run servers hosted by facebook, and not only that, it costs money to run the internet routers in your home, and it costs money to buy and run your PC. These things cost money. So what?

Why does it matter that it would cost "$5 billion" a year, distributed among hundreds of millions of people? That's economic activity that creates jobs, the money isn't flushed down a toilet or disappear. We, as a society, need people to spend money on something, why not this, instead of buying virtual crap for farm-ville that goes back into Facebook to pay for itself.

But anyway, Like I said, you could run the software on a virtual server just as easily as you could on this plug thing. That would reduce energy consumption quite a bit. Only people who really wanted it in their home would need to use a plug, or run the server as a process on their desktop PCs. Or hell, you could even run the server on a cellphone, which would also save a lot of power.

The plug thing is just a convenience to make it easy for people to setup. In reality, it's just amorphous software that can run anywhere. Whereas facebook's software only runs on facebooks servers.

-- So to summarize there are several ways to do this that don't use a lot of energy and don't cost much

1) Run it as a process on your Cellphone (which is low power)
2) Run on a virtual server in the cloud

And there are ways that you can do this that require less setup and are convenient.

3) Run it on a plug
4) Keep your portable profile on a central service, like facebook (but with the ability to move your data)

4 would also energy efficient.
posted by delmoi at 7:52 PM on March 19, 2010

How is keeping your photos on Flickr and your videos on YouTube better from a privacy point of view than keeping your photos and videos on Facebook?

Opps. I didn't realize that those were part of the privacy issue. In which case I have to ask how keeping your photos and videos on any server is supposed to be private? If it's on a computer attached to the internet, it's available to the internet. It's just a matter of how much effort someone wants to put into getting it.
posted by five fresh fish at 8:08 PM on March 19, 2010

What a bizarre argument. Yes, it costs money to run servers. It also costs money to run servers hosted by facebook, and not only that, it costs money to run the internet routers in your home, and it costs money to buy and run your PC. These things cost money. So what?

Do you think it costs on the order of $1 billion per year to run Facebook, not counting software development or system administration? Or that Facebook runs on $5 billion worth of hardware? The argument isn't bizarre: it's a straightforward statement that this program, even under extremely generous assumptions, would cost vastly more than a centralized service to very little advantage. In fact, given the security issues a distributed setup run by amateurs would create, I would argue that it would be a substantial step backwards.

But anyway, Like I said, you could run the software on a virtual server just as easily as you could on this plug thing.

And I'll grant that that's a decent idea, at least compared to Moglen's.

1) Run it as a process on your Cellphone (which is low power)

Given the frequency with which your data would be accessed (i.e. many, many times per day), that would seem to be a good way to run your cellphone's battery into the ground. Also, it would be pretty inconvenient for your friends not to be able to see your information whenever your phone was off or in a dead zone. Maybe some day phones will have the power storage and communications capabilities needed for being an always on portable miniserver, but I don't think we're there yet. Anyway, you'd still have the security issues inherent in trusting millions of end users to run their own server competently.

Also, to his credit, Moglen addressed a privacy issue with storing the data on a cellphone: in many countries it's trivial for law enforcement to arrest you on some slight excuse and clone your phone.
posted by jedicus at 8:17 PM on March 19, 2010

I have to ask how keeping your photos and videos on any server is supposed to be private? If it's on a computer attached to the internet, it's available to the internet.

There's a difference between theoretical privacy (i.e. anything connected to the Internet is probably vulnerable) and practical privacy (e.g. a locked-down Facebook profile is more private than a public Flickr stream). I would argue that a personal p2p social networking server (virtual or otherwise) run by the average Facebook user is actually less private than Facebook because the personal server is more likely to be vulnerable to security holes than Facebook (assuming that Facebook's admins generally do a better job of keeping the system secure than the typical Facebook user would, which I think is a fair assumption).

So Moglen's plan doesn't actually necessarily help users keep information more private. Yes, it keeps information from the prying eyes of Facebook, Inc but probably less private from malicious attackers. It perhaps keeps information slightly more private from the government, but sadly (and probably wrongly) the government's going to get what it wants no matter where it's kept, so I'm not sure that's really a selling point at this juncture.
posted by jedicus at 8:27 PM on March 19, 2010

Do you think it costs on the order of $1 billion per year to run Facebook, not counting software development or system administration? Or that Facebook runs on $5 billion worth of hardware?

Look that's just an insane question. You were talking about 300 million users. that works out to an additional $3.33 a year in electrical costs per user. I would guarantee that many of those people leave their lights on, and do other things that waste far more energy then that. Buying this, plus replacing one lightbult with a CFL bulb would reduce energy consumption overall, so if it makes you feel better we could ship it with a CFL or LED lightbulb.

Furthermore, even if people turn off their computers at night, hardly anyone turns off their routers, which could easily be modified to host these applications, just by adding a few dollars of hardware for flash memory and a USB port. Adding the software to a router wouldn't result in a net increase in electricity use at all.

would argue that a personal p2p social networking server (virtual or otherwise) run by the average Facebook user is actually less private than Facebook because the personal server is more likely to be vulnerable to security holes than Facebook

That's specious reasoning. It would be easy to build an dongle that was well secured, locked down, and automatically updated with patches to keep it secure. No, the average person isn't going to know how to configure a secure version of Apache, but the idea is that these things will be configured out of the box.

Like I said: millions of people, pretty much everyone on the internet already runs a server connected to the internet: Their router. And while security holes have been a problem from time to time they are not very vulnerable at all. and they require almost no actual effort on the part of the user to "maintain".

Given the frequency with which your data would be accessed (i.e. many, many times per day), that would seem to be a good way to run your cellphone's battery into the ground. Also, it would be pretty inconvenient for your friends not to be able to see your information whenever your phone was off or in a dead zone.

It was just an example, but the point is a cellphone is a good example of a device that's very energy efficient. I wouldn't want to run this on a cellphone, since obviously you could lose your phone.

But the 19 watt power draw is a lot, and you could probably design these with more energy efficient chips. And I agree the $99 cost is really high for what we're talking about. But I don't think it would add much cost to add this as a feature on existing routers.
posted by delmoi at 8:47 PM on March 19, 2010

Friendster has 115 million registered users (i.e., its peak theoretical user base). MySpace has something like 125 million registered users. Facebook has over 400 million active users. Even if every single Friendster or MySpace user ever has moved to (or at least joined) Facebook, that would only account for a little over half of Facebook's active userbase.

Friendster and MySpace were mentioned as examples of social networking sites that have lost users, possibly (even presumably!) to Facebook. Pointing out that Facebook has users that also came from elsewhere (where "elsewhere" includes "no previous social networking site") is completely irrelevant.
posted by kenko at 8:59 PM on March 19, 2010

Let's imagine I want to send a message to my friends. Under Moglen's plan I'd encrypt the message under a scheme that only my friends could decrypt. How is this better than posting a similarly encrypt message to Facebook (maybe a super-future-Facebook that made this easy to do)?

Sure, Facebook would know who looked at the message in the latter case. But my ISP would know who looked at the message in the former case.

Unless we were all using an anonymity network...which prevents anyone from knowing who looked at the message in either case.

So what's the point of deploying a fleet of personal servers?
posted by miyabo at 9:04 PM on March 19, 2010

And while security holes have been a problem from time to time [routers] are not very vulnerable at all.

Routers don't normally run front-facing webservers. In fact, consumer routers usually don't expose any services to the outside world at all. Furthermore, routers don't normally contain much in the way of valuable personal information in and of themselves. I think adding such a service would increase both the risk of vulnerabilities and attract more attackers.

Pointing out that Facebook has users that also came from elsewhere (where "elsewhere" includes "no previous social networking site") is completely irrelevant.

That's true, but the point remains that Friendster and MySpace (and Orkut for that matter) still maintain large userbases within particular markets that Facebook has not penetrated as successfully. For Friendster that appears to be Asia. For Orkut it's Brazil and India. Furthermore, Facebook is now so large and its network effects so dominant that it's highly unlikely that a grassroots effort that requires upfront investment of time and money on the part of each individual user could challenge it.
posted by jedicus at 9:12 PM on March 19, 2010

His statement of the problem is really quite lucid, and while his proposed solution demonstrates a certain disconnect from Joe Citizen, at least he's trying.
posted by phrontist at 9:17 PM on March 19, 2010 [2 favorites]

What he's saying is a little bit similar to what I've been saying for a long time, that it makes a lot of sense to just run your own servers, if you know how they work. Unfortunately, at the present time, it takes a very high degree of technical proficiency to do it well.

I don't think the hardware's really an issue; as delmoi is saying, 19 watts per household is going to be noise in the overall electrical load in this country. But the software to make that work seamlessly would be a bitch to write. It has to be super-reliable and super-secure, because the intent is for everyone to have one. That would make it Priority Target Number Uno for the bad guys. They'd loooove them some tiny, silent Linux servers with direct Net exposure. Without moving parts, the owner might not know their server was hacked for months.

Another issue is scalability; if you get really popular, the typical anemic upload speeds in this country probably wouldn't cope well. Most providers have large caches in their modems, and this means that heavy uploading makes downloads glacier-slow, and vice-versa. So if your PlugBook page gets popular, your Internet is going to get very slow, and past a certain point, you just won't be able to have any more friends at all.

This is a truly difficult problem, and worse, it's a truly difficult problem aimed purely at end-users. That's about the worst possible use case for Free software; you can count the number of truly successful end-user Free software projects on one hand. (I can only think of two offhand: Firefox and Ubuntu.) This takes the kind of laser focus and engineering that you typically have to pay for; people doing this as a hobby, in their spare time, probably won't have the focus to spare to do genius-level work. And that's what this would take... ordinary brilliance is unlikely to suffice.

I'd love to see this happen; I think it would be fantastic. But if, as Moglen is saying, the new generation just doesn't care about privacy, then it's going to be a lot harder to get the manpower to get the job done. The grizzled, seasoned coders might be interested, but most of them have jobs and families these days, and expecting something this transformative out of that community may just be wishful thinking.

Apple could do something like this, but they're so anti-freedom that I can't imagine them ever putting the resources into it. They just, flat, do not do things that don't benefit Apple anymore. Google might be able to, they have the freedom focus, and they certainly have the raw coding capability, but I don't think they have the UI expertise. And I also have trouble imagining Google putting millions of dollars into something to actively turn people away from their services.

tl;dr version: Great idea, but the people with motivation probably don't have the focus, and the people with focus probably don't have the motivation.
posted by Malor at 9:25 PM on March 19, 2010 [2 favorites]

His statement of the problem is really quite lucid, and while his proposed solution demonstrates a certain disconnect from Joe Citizen, at least he's trying.

I agree with that completely.

What's interesting about his solution is that it's basically a telephone approach to the internet. Consider your home telephone as an interface to a router that also stores your friends' contact information that you can use to contact them directly, or exchange messages. In the early days of Telephony, I believe you bought your telephone from your phone company, just as now you may buy your router from your ISP. The difference is that back then, AT&T or Ma Bell didn't own your contact list or serve you ads based on the data they keep on you.
posted by cell divide at 9:38 PM on March 19, 2010

Let's imagine I want to send a message to my friends. Under Moglen's plan I'd encrypt the message under a scheme that only my friends could decrypt. How is this better than posting a similarly encrypt message to Facebook (maybe a super-future-Facebook that made this easy to do)?

There's not, except the face book that exists today does not implement crypto, so your friends would all need to download some crypto tool.
posted by delmoi at 10:57 PM on March 19, 2010

(there should not be a space in 'Facebook' above :P)
posted by delmoi at 11:00 PM on March 19, 2010

We're having a party, right now on my farm on Farmville. You're all invited.
posted by doctorschlock at 11:12 PM on March 19, 2010

This is a truly difficult problem, and worse, it's a truly difficult problem aimed purely at end-users. That's about the worst possible use case for Free software; you can count the number of truly successful end-user Free software projects on one hand. (I can only think of two offhand: Firefox and Ubuntu.)

Open Office. Android. Audacity. mplayer. VLC. gimp. MythTV (helped by Mythbuntu).

This takes the kind of laser focus and engineering that you typically have to pay for; people doing this as a hobby, in their spare time, probably won't have the focus to spare to do genius-level work. And that's what this would take... ordinary brilliance is unlikely to suffice.

Eh. I don't think the problem's as hard as you do. I can envision several solutions that would seriously mitigate the problem.

But you're making the mistake that many people do: that open source means hobbiest or non-profit. There are plenty of corporate-sponsored open source projects out there. They all make money for their respective companies in various ways.

For instance, if you built PlugBook, you might also offer your services (at a fee) for datamining it or writing applications on it, for instance.
posted by Netzapper at 1:05 AM on March 20, 2010

There's not, except the face book that exists today does not implement crypto, so your friends would all need to download some crypto tool.

Agreed! What I'm saying (admittedly not very succinctly) is that privacy is a software problem, not a hardware problem.
posted by miyabo at 8:40 AM on March 20, 2010

assuming that Facebook's admins generally do a better job of keeping the system secure than the typical Facebook user would, which I think is a fair assumption

A-ha-ha-ha! Gigglesnort! Suffice to say, I disagree with your assumption.

Another issue is scalability; if you get really popular, the typical anemic upload speeds in this country probably wouldn't cope well.

If you get really popular, then you aren't concerned about the privacy of your photos and shit, and you move onto a commercial system, be it Facebook or a private hosting service.

19 watts per household is going to be noise… But the software to make that work seamlessly would be a bitch to write. It has to be super-reliable and super-secure, because the intent is for everyone to have one.

You mean like enabling the Apache web server — a couple clicks operation — on a Mac Mini? Reliable, secure, and draws only 13W when no one's hitting up the server. Which, for the type of private situation we're describing (serving photos to friends and family), is going to be the situation almost all the time.

Or one can chuck a hundred bucks at Webfaction and get plenty of server space and a ton of WordPress and other instant-play software.
posted by five fresh fish at 9:26 AM on March 20, 2010