Oblomovka

2007-10-30»

Back from a whistle-stop tour (in that it lasted about as long as I can whistle) of University of Michigan-Dearborn, and Reed College, Portland, doing my regular speaking job explaining EFF issues to college students.

I do these gigs about once a month, so if you know any students (or perhaps turn into one after 11AM), contact Samara and get me to come talk at your school. All the monies go to EFF. I like the format; I have a standard pitch, but I have room to throw in extra material or explore topics the audience are interested in. At Michigan, the talk was organised by WUMD, and we had DJs play a set beforehand, so we ended up talking a fair bit about artists’ rights and distribution. At Reed, I was invited by the college’s Free Culture chapter, so we spent some time dissecting public policy. Both times, the audience was fantastic, and we all had a great time. It’s nice to do some more public speaking, after quite a few years of not much performance.

I also liked the more random questions after the main event. We spent a long time at Dearborn talking about Valley culture. In one of the city’s hookah bars (Dearborn has a really distinctive and self-assured arab community, from what I could glean), we meandered over mesh networks and data havens, mainstream game companies and whether it sucked to work for them, why Ruby was better than C++. “Is it true that the guy who wrote BitTorrent wants to live on an oilrig?”, they asked. I explained that was two different people, with entirely different livejournal accounts. The organiser at WUMD runs the radio station, studies at the college, works in his father’s video shop, and then at the weekend runs a clothing company with his cousin. He wants to study law, and is an EFF member. I still love the future, and glad I get to travel to meet it so much.

Comments Off on never not blog if you have too much to say

2007-10-25»

We should have had a plan to do something with the NTK site after we trailed off updating it, but I don’t think we could have come up with something as deeply wonderful as ntk.org.uk, part of the current No To Knives campaign in the UK.

Things I like about this site:

• It’s all in Flash.
• It’s aiming to get 1,000,000 people to pledge not to carry knives on the hard streets of Britain
• So far, they’ve got seven.
• I can’t link to anything on it, because it’s all in Flash.
• But that’s okay, because the terms of conditions would ban me from doing so anyway:
  

  You may link to our home page, provided you do so in a way that is fair and legal and does not damage our reputation or take advantage of it, but you must not establish a link in such a way as to suggest any form of association, approval or endorsement on our part where none exists. You must not establish a link from any website that is not owned by you.

• Also, I had to quote that bit, because they also say:
  

  We may revise these terms of use at any time by amending this page. You should check this page from time to time to take notice of any changes we made, as they are binding on you. Some of the provisions contained in these terms of use may also be superseded by provisions or notices published elsewhere on our website.

• Amusingly, the terms and conditions are the only page that I can link to, because it’s the only page not in Flash.
• …unlike their accessibility page, which you can only get to in (inaccessible) Flash.

And I haven’t even started on my questions about why the Royal Armouries of all people want everyone to stop carrying knives. Is this like how only the Queen can own swans?

Comments Off on i’m in ur country, pollutin ur namespace

2007-10-18»

Sorry about the extended pause: I just spent a week in Beijing. About the only thing wrong with Beijing is that everyone abbreviates it to “BJ”. Oh, that and the repressive and autocratic regime: which, incidentally, I am more optimistic about. (Understand that I could scarcely have been more pessimistic about it, given my formative civil liberties event was waking up to the Tianenmen Square massacre.)

Bear in mind of course that I am randomly extrapolating from barely a week of interaction, but There was a lot less respect than I was expecting. Lots of teasing and undermining of authority, from giggling waitresses teasing their stern boss, to extended, loudly gesticulated debates between hotel managers and bellhops, between out-of-town cabbies and bemused police officers. It felt more like New York than a cowering, centrally-planned potemkin Olympic village it’s sometimes painted.

Even experiencing the Great Firewall filled me with my perverse sense of hope. (At work, my self-given title is “World’s Most Conflict-Averse Activist”, but I am a also a fair runner-up for “Digital Rights’ Biggest Gap-Toothed Optimist”. At a Canadian privacy conference last month, a TV crew filmed me specifically because I was the only pro-privacy speaker who didn’t seem actively suicidal).

The Great Firewall, in brief, is appalling. Appalling enough for it to be one day abandoned, I hope. It damages the entire experience of the Internet — not just for obvious political searches, but for everyday business, too. DNS queries fall over regularly, connections drop for random but broadly discoverable reasons, searches are arbitrarily banned. A resident and I spent a few minutes whittling down the use-case for a new and frustrating Google block. It turned out the PRC was blocking any Google search with the Chinese pinyin syllable “zhēng” in it. That’s like someone blocking any search term that contains the English word “trip”. My colleague had discovered it because they were searching for a Chinese company that included that syllable – Zhēngtu Networks, local creators of the second most popular MMORPG in China. Google has been down in the last few days; so has YouTube. This isn’t just breaking Google’s service; it’s breaking the whole functionality and usefulness of the Net.

The Firewall divides China into a local domain with okay connectivity and rampant self-censorship, and the “foreigner” domain with crummy degraded performance and arbitrary blocking. As a policy, I don’t think that can work. You can build a walled garden as big as half a continent, and you’re still going to painfully suffer competitive disadvantage to your trading partners. If the US had declared itself a Net isolate from the rest of the world, the Net would have died on the vine. As everyone who has every tried to pick out what they thought everyone would need from the Internet has learnt, it’s not about having “enough”, it’s about having all of it. You don’t know which part of the Net you need, because everyone else is finding different parts that they need — and you need them.

I think far more sinister is the developing technology that silently drops parts of the Internet with little observable effect on even quite close neighbours. For instance, British Telecom’s Cleanfeed, which can block individual URLs within a domain, and leave the rest untouched. A manifestly broken Internet will provoke all kinds of debate, offline and on. A silently censored Net may encroach until the lack of debate becomes perfectly natural.

No narrative dreams in China, just lots of mental processing of my attempts to learn and understand Chinese ideograms. Like semantic tetris, you close your eyes, and the radicals leap up out of the visual noise.

Here’s the Chinese for Internet, or “cyber”:网际 (wǎng jì). It’s the image for “net”, appended to the ideogram for “edge/boundary/between”.

I like its X X eyes, as though the sign for Internet has its own embedded emoticon.

Comments Off on peking ducked out for a bit

2007-10-09»

I’ve now had a few nightmares, I am sorry to report. Recurrent nightmares, or at least endless rehashing of the NIGHTMARE THAT IS MY BOURGEOUS LIFE was what led my subconscious to forgo the whole dreaming thing for the last few years. It’s not too bad, though – my last nightmare had dragons in it. We hid under the kitchen table; very effective. Then, Mike Myers turned up and filmed the new Austin Powers in my bedsit. He’s uch more humble in real life. And by “real life”, here I mean “in my dream”.

So, a lot of people smershed my mentioning of the old P2P revolution in the last entry with my general thinking about the future move to the edge. I hadn’t intended to make a direct comparison, but it’s worth noting, as many did, the changes in the network since the glory days of 2001, and how that would effect creating edge services now, as the P2Pers were trying to do then.

First, and most obviously, the massacre of dial-up users is proceding as planned. There are still plenty of them out there, but we no longer have to feel guilty if we do not care about them. More importantly, always-on Net connections are pretty much everywhere. Even dynamic IPs are generally fairly static in the medium term. Hooray!

Second, the realisation by most protocol-designers that it’s an HTTP world, and that we just try and communicate in it. HTTP was a fantastic fit for the early net, so good in fact the modern Net has now co-evolved to be a good fit with HTTP. Nobody cares a goat’s fig about NAT because it don’t mess with the HTTP — and that’s one of the main reasons the Internet is now so NATty. That makes everybody lean toward HTTP to work well in this new infrastructure. One of the reasons why REST stuff just works is that it lives in HTTP space, so there are no sudden moves. P2P protocols have always had HTTP elements, but I think it may be true that whatever develops next in the peer-to-peer realm will just look like local webservers talking in a RESTy way with other webservers – or to talk with humans or other userspace applications (I count humans as a “userspace application”).

A corollary of this is that I think you really have to just deal with everything else that comes with HTTP — including DNS. All the old P2P dances have sexy URLs of their own devising, and hashes, and DHTs, and all that jazz — but without popping up in DNS-land, these servers are just invisible to everyone. Even Microsoft’s own P2P DNS-a-like, PNRP, falls into this category. Sure, your machine may be announcing to other Microsoft peers that it’s My-Computer474342.pnrp.net, until Microsoft does the obviously sensible thing and starts resolving those addresses in standard DNS (please somebody write in and tell me they do) , those addresses are just burial plots in a walled garden.

(John Gilmore once proposed a decentralised solution to this problem, whereby the toppest level domain would actually determine which protocol the IP service should use to find the rest of the domain. The current TLDs would be grandfathered in as ‘.com.icann’, ‘.net.icann’, etc. You wouldn’t need to change any URLs, because search domain ‘.icann’ would be default. But ‘My-Computer743473.pnrp.’ would run using Microsoft’s P2P name-finding algorithm, ‘7a7898bef783ed731aaf.bittorrentilikehashes.bittorent.’ would find a bittorrent resource, and so on. Obviously this would reduce ICANN’s role to one Postel-looking geek adding a list of protocols onto a list and chatting to the BIND guys. Since that geek wouldn’t need a multi-million travel budget, it will never happen.)

But I digress. The key point here is that if you can control your own DNS, and your can control your own webserver, you’re pretty much ready to go as a generic everything server on the Net, whether you’re hanging off the edge, or partying in the affluent middle zones.

Oh, you say, but what about NAT! What about the unreliability of the edge! What about if the kitchen table was made of wood, and the dragon just went RAAAARGGGGH and breathed fire over the top of it?

These, and other nightmares, I will discuss after this word from my sponsors.

The Democrats are currently considering caving to the White House on granting retroactive immunity to the telecommunication companies for breaking the law and spying on your phone calls. It’s all going to happen this week. If you haven’t already, please call the Congressional leadership and tell them not to give an amnesty for lawbreakers.

Comments Off on h-t-t-p, you know me

2007-10-04»

The two background themes of this blog conspire: my digestive problem is keeping me awake, and stopping my dreams. Well at least I’m not fitfully asleep, dreaming that there’s a small weasel biting the left side of my trunk or something.

Lots of great conversations with people about my ongoing flailing ideas here. I am awful at replying to email, because by the time I’ve found the reply button, there’s another email to read and oh, bright shiny blog thing, but I did read them all. Even the guy who said that I’d just rediscovered Ray Ozzie’s Groove (sorry if I was a bit rude in my reply, Andre).

What made me rub my hands with glee was that all of the replies were by people who I know are much smarter than me, which means I’d managed to fulfil my primary aim of expressing an idea so irritatingly vaguely that better heads will fill it in for me.

A telltale of my favourite smart people is that they don’t prematurely pessimize, which is to blindly announce “Well that would never work because X, Y, and Z”. Buzzkill. No, my kind of smart people go “Well, you’ll have to fix X first, which I think you could do by doing A, B, and — oooh, I bet we could solve ‘Z’ with some string and that doorknob over there! Let’s go!”

However, to speed things along, I’m now explaining to such people there’s a class of problems that I don’t even want to fix in this thought experiment (which, to remind everyone, is — what happens if we push to the edge everything that we’re currently throwing onto Google Documents and other Web-based services). Examples of this class of problem in my gedankenexperiment are:

• There’s not enough bandwidth for a home or mobile servers.
• Customer-level connections are too unreliable for services.
• Hard-drives are too fragile to lug around with you.

These are examples of problems that I hand-wavily announce will bore themselves to death. That is to say, I don’t want to talk about them, because I believe they are very dull, and I am confident there are clever people who don’t find them quite as boring as me will solve them for me.

There is risk here. You do have to be careful of what problems you assume will die of boredom, because sometimes they turn on you and bore your entire future vision to death instead.

NAT traversal is a good example of that. NAT traversal is a tremendously dull topic that was far too boring for most of the people excited about P2P technologies in 2001 to think about for very long (although the ones that did find it fascinating kept the rest of us up until 4AM drawing funny diagrams). They had a revolution to lead! Endless opportunity lay just beyond the horizon!

P2P was what Web 2.0 was supposed to be, incidentally, five years earlier, almost literally (the Web 2.0 conference came from Emerging Technology which came from the ashes of P2Pcon). Sadly, P2P never developed escape velocity, and the entire fledgling industry collapsed more-or-less into BitTorrent and Groove, and that was that. NAT traversal was one of the problems that still hinders it, as is the fact that client PCs generally don’t act like servers, but vanished off and on the networks in irritating ways. By the time you’d coped with constantly self-dismantling networks and impossible to reach edge nodes, I understand most P2P developers wanted to gnaw their own legs off in tedium. The endless opportunity had to be endlessly postponed while everyone fixed this one last problem with getting the Network to work over firewalls, and with constantly changing dynamic IPs, and a whole rats-nest of other dull issues.

If you want a more modern way of thinking of the risks of a boring problem, think of the utterly dull issue of cross-platform JavaScript compatibility. An entire generation of AJAXian prototypes died on intranets because they weren’t cross-platform, and it took decent JS frameworks and know-how built by Stakhanovite miners in the dark pits of tedium.

But we prevailed! The problem, pinned down by the corpses of endless headslapping programming hours, finally died of its own boredom, and JavaScript ultimately came into its own. About seven years later than anyone imagined.

Boring problems can heavily delay the arrival of the future, but they don’t really change the game.

So because we are all Buckminster Fullerish futurists here, let’s airily discount them. Our problems with bandwidth, at least in the United States, are down to awful, creaking monopolies, that will slowly die choking on their own gorged subsidies and foul bellhead toxins (and if not, there’s always China). The fragility of harddrives isn’t going to last another generation.

The unreliability of consumer connections, though. Um. I don’t know whether this is a problem that will die or be fatal. One could argue that it was what actually *did* kill the P2P unboom. Certainly, unreliability is something that the Internet is supposed to deal well with, and when it doesn’t, we could certainly do with some deliciously generalisable solutions. It’s not like it’s not a problem if you keep servers where they’re supposed to be, in yonder cloud. When your main server goes down, what do you do? And can you do that when your edge server drops off the Net a couple of minutes every day, or a bunch of seconds every hour?

Oh, all right. Have your damn comments. You’re just going to pile on and say you don’t have the slightest idea what I’m talking about, and have I tried peppermint tea, aren’t you?

|

Comments Off on death by boredom

2007-10-02»

I have had more vivid dreams, but the last one was a long debate between me and my friends because I was using mutt 1.0.2.1 and it was calling “mailbox deadlocks” on their servers. Nothing more draining than waking up after dreaming an imaginary one hour conference call, especially when you realise you have a real one hour conference call in a few minutes. So I shan’t talk about that.

I’ve been spending some time explaining in a hand-waving fashion my instincts about moving to the very edge. Usually I keep this stuff close to my chest until I’ve thought it all out, for fear of looking like an over-obvious idiot. But over time I’ve seen a lot of obvious idiots become fantastically smart just by letting it all hang out online, so I will bore you with my half-baked, poorly styled, not-very-viral ideas as they occur.

Brief summary of the thesis: I’m bored of this current revolution, so I’m doing the cheap trick to help plot out a possible next one, which is to reverse and take to its extreme one of the obvious contemporary trends. My question right now: given that we’re entrusting so much data and control now to the cloud and the server-farm, what happens if we pull the other way, and swing more power out to the edge, and the end-user? How far can we go with that?

Intermediaries have been what I’ve been considering today. Browsing EFFish issues, I see a lot of problems which are caused by the distance between an intermediaries’ goals, and that of its customers. When your hosting provider, includes as part of their terms and conditions that they reserve the right to take you down if you cause problems with them (or even criticise them)

Intermediaries don’t have to be corporate though, nor middle-men. Pooling resources in a communal way can have problems, too (witness my dream, where my mutt process brings down everybody else’s accounts on a communal co-loc). And even having a home server doesn’t seem to fit how I imagine protecting data and providing user power. There are interactions and privacy that exist within a home, and between friends.

I guess what I’m imagining is the single-person server: holding and electively sharing your data with other single-person servers. I don’t see this as substantially different from people having their own phones. Indeed, phones are already powerful enough to support that.

(The 21st century question about this is — what are the energy costs? I’m not going to have that argument for a while, because I want to find out more about the nature of decentralised energy systems.)

Comments Off on intermediaries