Oblomovka

2007-10-05»

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?

|

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.)

2007-09-29»

I'm in the air, wheeling into San Francisco, having just finished William Gibson's Spook Country. I brought it with me on a trip to Canada, because you should read later Gibson on airplanes and in slightly-foreign Western hotels, just as you should read Ballard in airports and light industrial parks, and William Burroughs off your face on purest horse.

I like Gibson in exactly the way you thought I would, so I'm not sure I can say anything unpredictable here. The potted-review I came up with when I was about twenty pages will do: This feels like Eighties Gibson, writing about our recent past as his envisaged near future. Set in 2006, it has geohacking, retro-fame, rogue states, cold war warriors gone white hot with rage: they're all written about as though extrapolated from 1985, instead of marked back from 2007. I feels like a 20th century dystopia, which sadly doubles up as rather optimistic from our point of view.

There, gnomic enough for you? I'm trying to be awkward. I was mildly irritated all the way through by a spoiler I'd read in a review -- which turned out not to be a spoiler at all, but an inept phrasing by the critic. No-spoilers are even worse than spoilers, because you don't even have that grim sensation of following through. You just read to the end of the book, and then go "Hey, wait, wasn't it supposed to turn out they were all otters all along?"

I'm still having dreams -- which may, now I think about it, due to my steadily worsening stomach (it may be horrendously nasty gut-rot, but I doubt it: I think it's just IBS turned psychosomatically psychopathic. I'm seeing a gastro in a fortnight. I'm sure he'll just recommend a change of diet. Gastro! The menu!).

Last night I dreamt I was in a jeep in South America with Cory, planting explosives to covertly excavate out a new, spare, Panama canal for the US. The day before I was a sort of inept Professor Xavier, doing childcare for a bunch of superpowered preschoolers and having to defend them from some bigger supercriminal kids. Lots of soccer-coach encouragement of them to shoot percussive sonic blasts while I cowered behind them. This is a parental anxiety dream, but more exciting than most.

I was in Canada to meet with privacy activists. I can't give you their names because obviously we all met in darkened rooms wearing blindfolds. I did get to see Michael Chertoff give a keynote though. Boy did he misread the audience. Never ever tell an international conference of data protection and privacy commissioners that you can scan a fingerprint at the US border, and match it to a print on a document found in a safe house in Europe. Because while you're sitting there thinking "hooray for l33t national security tricks!", they're thinking: what the hell else are you doing with that tech?

I guess we're all in a fucking jeep driven by a science fiction author now.