Monthly Archives: May 2008

Five nines? No nines at all, more like

Twitter error message

A very long time ago I ran a web-based email service (allow me to tell you about it one day). It was moderately successful and, before the latter unpleasantness it had well over a million users and substantial traffic and brand awareness. I learnt one really big lesson from that particular experience, though: never run an email service. It’s a mug’s game. The problem is that email is essential infrastructure.

For email users it’s like dial-tone. Pick up the phone: if you don’t hear dial-tone what do you do? Do you say: “Hey, no problem, I’ll try later”? No. You say “what the fuck’s wrong with the phone?” “Hey everybody! the phone’s out!” “Shit. Did society collapse? Was there a nuclear bomb?” and so on. No one is sanguine or relaxed about a phone outage. Likewise with email. If clicking ‘send’ doesn’t work first time or if you get no email at all for fifteen minutes you’re pretty soon popping veins in your neck.

Phone networks and email systems have to be reliable. In the telecoms world they call it ‘five nines’. They mean that a phone network has to up 99.999% of the time and they engineer their systems to deliver this. Email systems are now engineered to the same standards. And it’s not cheap because the 80/20 rule applies.

Keeping your network up for 80% of the time costs about 20% of your systems budget—piece of cake. The difficult final 20% costs 80% of your budget (and it’s actually probably more like 95/5). And that’s before you’ve spent a quid dealing with the legion of bottom feeders firehosing your servers with spam. Like I said, it’s a mug’s game. And this explains why email provision is consolidating fast and why even big in-house systems are being outsourced to specialists.

Which brings me to Twitter. As you know, Twitter’s clever for all sorts of reasons. I’ve gone on about Twitter here before: I think it’s the most important application to appear on the Internet for years—possibly since the web itself. Seriously, I do. But it’s especially clever because the Twitter experience has been engineered so that users aren’t really bothered if it’s not working. Even big users (I would count myself as a big user) can live without it for a few hours, even for a day or two.

Nobody uses Twitter for anything important and, although it supports direct messages between users, it’s principally about the buzz of daily life—so even a longish period of extreme flakiness like the one we’ve just seen barely spoils the experience. Twitter’s remarkable achievement is to be important enough to produce addiction but nowhere near important enough to produce a phone call to the complaints department. Twitter’s a long way from five nines and it doesn’t matter at all.

What is Speechification?

Here are some words I wrote to prompt me in a meeting that Russell and Roo and myself went to at the BBC yesterday. It was like the nicest pitch meeting you’ve ever been to. Lots of important and interesting people from the Internet side of the BBC asked many questions and made many suggestions: a really open and positive reaction to what we’re doing with BBC stuff at Speechification and Watchification. Thanks especially to Jem Stone and to Sophie Walpole for putting it all together.

What is Speechification?

It’s curation. For years now visionary-types have been saying that pretty soon people will be curating media. Well, we’re actually doing it. The world’s just been waiting for a large enough and accessible enough bank of content to play with. So we roam the corridors of the BBC’s archive selecting the stuff we think is really excellent, unusual or important and putting it on display at Speechification.com.

It’s cheeky. We definitely push at the edges of what it’s OK to do with BBC content and we do this not because we’re pirates or vandals but because we want to exert some gentle pressure on the corporation’s leaders to do the right thing about rights, access, archiving etc.

It’s shareholder activism. Greens who want to influence the behaviour of smokestack corporations buy a handful of shares and show up at the annual general meeting to heckle the board. We don’t own shares in the BBC but we’re licence fee-payers so we’re at least stakeholders: opinionated, supportive stakeholders.

It’s a celebration. The BBC’s speech output is one of the glories of British culture. I don’t want to sound smarmy but listening to an evening of great programmes on Radio 4 is a privilege: the kind of condensed emotional and intellectual experience that leaves you smiling without knowing exactly why. We want the world to know about this stuff.

It’s unofficial PR for the neglected stuff. Mark Damazer is on record as saying that Radio 4 doesn’t do a good enough job of marketing its own output. He’s dead right. More than once we’ve featured programmes at Speechification that go out at ungodly hours, don’t have any useful information at bbc.co.uk and weren’t important enough to warrant a press release. We’re often the only people to write about a show anywhere. For these shows we’re an unpaid marketing department. We should bill them.

Five reasons it might not be so bad to have Boris as Mayor after all

1. The Mayor doesn’t have much to do anyway. He may have an £11B budget but it’s really only half a job. New Labour deliberately hobbled the Mayor’s office from the beginning by retaining control of everything important at the centre and providing no direct tax raising powers. The remarkable thing about Ken’s tenure has been how much impact he’s been able to have with control of public transport and bugger-all else.

2. He’s funny. He is funny isn’t he?

3. He’ll be a one-term Mayor. Nothing he’s proposed is achievable within budget. He’s backing away from the Routemaster idea (which has been authoritatively rubbished). Central Office cagily supported Boris but not his policies. With all the big reforms already firmly entrenched Boris will struggle to make an impact. As budgets rocket and policies evaporate disillusionment will set in.

4. Cameron will cut him loose. He may be better behaved now but you can’t innoculate a boob like Boris against gaffes. Can it be long before he alienates bus drivers or pensioners or people who live in Penge?

5. He’ll get bored. Boris evidently has the attention span of a nine year-old boy. As his concentration lapses he’ll drift off. Pretty soon he’ll forget where he works and after a year or two Londoners will be able to pick someone else.