atroche.org

My ugly Clojure solution

Sat, 25 May 2013 00:59:00 +0100

Along with reading the Joy of Clojure and building toy Pedestal apps, I’ve been working through the exercises on 4Clojure. Each problem builds on the skills gained in the previous one, and you get to see the best solutions after you’ve submitted yours.

Occasionally I get the most elegant solution right off the bat. Most of the time I’m humbled. Sometimes I’m downright embarrassed:

https://gist.github.com/5647270

This is one of the things I love about Clojure: if your code looks ugly, you’re probably looking at the problem wrong.

It’s problem 43, by the way. Get on it.

Practical Promises in JS

Tue, 02 Apr 2013 23:52:00 +0100

I enjoyed James Coglan’s article about promises in Javascript, even though his political statements re. the node platform’s design are silly.

Mikeal Rogers is definitely right on this one:

The platform must make decisions that encourage compatibility and discourage incompatibility in the ecosystem.

node wouldn’t be such an astounding success if they’d tried to force everyone to use promises.

But if you can ignore that part of it, the article makes a great read for anyone wondering how to fit promises into his or her workflow.

To learn Chinese faster I’ve been reading their books,...

Fri, 29 Mar 2013 04:25:00 +0000

To learn Chinese faster I’ve been reading their books, listening to their music and following Kevin Rudd on Weibo.

I’ve also been watching movies. And up until today, when I watched Blue Kite (the entirety of which is on YouTube, see above), I was disappointed. Crouching Tiger is fun but vapid. Farewell My Concubine (which won at Cannes in 1993) is beautiful, interestingly shot but ultimately a plodding, overwrought melodrama.

Blue Kite is subtle, touching and entirely human. It’s about a family living through the cultural revolution, and it hits hard. It gave me the same feeling as when I first read Animal Farm in primary school – you know the part where you realise just how bad things are about to get for everyone except the pigs and the dogs?

If you’re going to watch a Chinese movie, watch this.

One of the shots from our trial run of GIF Booth at the Toff in...

Thu, 28 Mar 2013 05:12:27 +0000

One of the shots from our trial run of GIF Booth at the Toff in Town. These guys kept coming back and doing static poses, even after we explained how GIFs work. The mind boggles.

It goes from the most boring GIF ever to a pretty decent one in about 50 milliseconds, thanks to Adam.

“Mobile” is already outdated

Wed, 27 Mar 2013 01:03:00 +0000

I just got back from a having a chat with some savvy advertising guys in South Melbourne. The creative director asked me about “the future of mobile”.

I responded enthusiastically but blandly: “It’s the future, that’s for sure. Just look at the numbers.” But there was a nagging thought in the back in my mind, one I couldn’t flesh out without disrupting the flow of our chat.

Luckily I had a half-hour bike ride north afterwards.

The phrase “personal computer” is amusing in 2013. Hipsters in Yarraville use it ironically when talking about their MacBook Airs. Why? Because every computer you see today is personal. But at one point, when our idea of a computer was twenty tonnes of steel in a vault, it was important to make the distinction.

We’re used to desktops, so we make a similar distinction with mobile. But soon everything is going to be mobile, in the same way that all computers now are personal. The term is destined for quaintness.

Sure, we might have some sort of desktop experience – screens, keyboards, and so on. But the little device in our pocket or on our face or embedded in our clothes will communicate with them wirelessly, and we’ll carry the important part – the computer — everywhere we go. Ubuntu’s new phone is a step in that direction.

And four-inch touchscreen displays are old news. Hobbyists and corporate research labs are already playing with cooler shit. One of the tamer examples, Google Glass, launches at the end of this year. You’re about to start seeing people wearing their computers on their face, at the supermarket. Maybe we’ll control it with this awesome “wearable gesture control” armband.

Meanwhile, neural implants are moving shockingly fast. Check out this clip of a paraplegic woman using her mind to operate a robotic arm. Or this MIT Tech Review article on implants for the blind. And if we can let blind people see, we can give normal people a neurally-superimposed Twitter feed.

William Gibson said “the future is already here — it’s just not very evenly distributed.” Guess what? We’re getting better at distributing. Whether via Kickstarter or Google, these toys are going to reach market way sooner than you think.

So everything’s going to be mobile. But what’s wrong with using the word itself? It’s handy. And it’s just a word, right? Except, words profoundly shape our discourse and our thoughts. We’ve got to choose them carefully.

Look at Dell. They were huge. They owned personal computing. But they had a fixed concept of what a Personal Computer is. You can just imagine a board meeting at Dell in the early 2000s, with a gray-haired executive sternly telling his underlings “We’re Dell. We make Personal Computers”, while pointing to this:

The reason Apple got it right both times is not because they “get” personal computers or mobile, but because they understand something more fundamental about technology.

Jobs used to say that the computer is a “bicycle for the mind”. With the Apple II and the iPhone, they weren’t trying to build a PC or a smartphone. All they did was build the best bike they could with the technology at the time.

The computers of the coming decade will surely be personal and mobile. But they won’t come close to resembling an iPad. Anyone who invests too much in the current conception of “mobile” is going to be quickly left behind.

Learnings from an art installation

Mon, 28 Jan 2013 00:34:00 +0000

A couple of days ago my friends and I installed a shed at Rainbow Serpent, a festival of ten thousand ravers and hippies in the middle of the Australian bush. Inside was an old TV that spoke and formed words out of flying stars. We watched wirelessly via webcam and had conversations with trippers.

It was fucking awesome. Here’s what I took away.

You should only rely on yourselves

The organisers of the festival were lovely, but had bigger things to worry about than us. With bushfires, overdoses, and a draconian shire council hanging over their heads they simply could not run around looking after us.

It was our responsibility to barter for hay bales, beg favours from tradies and perform all the other little indignities that don’t occur to you when you’re sitting on a couch in Melbourne.

The initial concept is only a small part of the experience

I can take credit for the basic idea and assembling a crew. But after that it was definitely a shared creation. We were faced with hundreds of questions like: should it be housed in a shed or a tent, or left in the open? And every time we answered one there were more hiding underneath. How big should the shed be? Do we paint it? All of it? Which colours? Creating something is a quintessentially fractal problem.

The same installation could’ve been implemented in an infinite variety of ways. It’s not the “what” that’s interesting or creative, but the “how” — the execution. Sas painting the shed like a galaxy, Owen finding and preparing the perfect old Sony CRT, Pat wrapping the structure in blinking fairy lights to draw the crowds — the whole thing is uniquely ours because of how we answered those never-ending questions.

Anything you can do at home, do at home

Trying to work, think and code on a laptop rapidly losing power in the blistering Australian sun in the middle of a dusty, windy field while throbbing bass pounds from gigantic subwoofers is … suboptimal.

So many problems only occured to us when we executed in real world conditions, like how the laptop would behave when the monitor cable got pulled out (thanks, wind) or how being encased in a TV inside a metal shed would mess with the router’s performance (a lot). In hindsight, we could’ve predicted some of those conditions and tried to simulate them at home, where we had supplies and quiet and cups of tea.

Murphy’s Law is only half right

Sure, lots of things went wrong. Our video calling software bugged out, roadies kept shutting down our bio-diesel generator and ravers used our shed as a place to snort 2-CB out of the wind. But for every bad coincidence there was a lucky break:

— The wind was strong enough to carry our shed off, but we made friends with a trucker who had spare rope for lashing it down

— We forgot some of the milk crates for raising the TV up, but the bales of hay we sourced from a nearby market stall were the perfect dimensions

— The TV that enclosed all our sensitive bits (laptop, screen, router, speakers) was the exact width of our shed, and looked like it was made to fit our LCD screen. And Owen found it driving past a random street corner in Richmond.

I’m not mystical in any sense of the word, but things have a tendency to work out when you’re doing something worthwhile.

Everyone loves interactive

From the festival organisers who took a gamble on us to the curious girl who ran off to bring back her wheelchair-bound friend, everyone reacted incredibly well. Even while we were setting up, just telling people about the idea got them excited.

Most of the art I see is disappointingly static. The only way you interact with it is by looking at it. When an installation turns you into a participant, it’s a special feeling — especially because we’re only just starting to explore this space as a culture.

Rainbow Serpent actually had a bunch of interactive stuff, from a insta-polaroid photo booth to a giant snake that ate cans to themed parties inside geodesic domes. It’s only going to get easier as tech gets better and cheaper. I’m excited.

People are forgiving

It didn’t matter that our visulisation lagged, or that our speech synthesiser couldn’t pronounce adverbs, or that sometimes the installation had duct tape over the entrance and safety cones inside.

The fact that we were trying to make something cool, for fun, for free meant that people would give it a shot anyway, and give feedback and offer to help us out. Which made it a success, despite so many things failing.

Making stuff takes lots of time and money

Our biggest limitations besides inexperience were hours and cash. We had to buy the shed, paint, petrol, food and so on, as well as put in hours coding and designing and building.

Our electronic setup was totally barebones. Our laptop was slow and old and (most embarrassingly) running Windows, and the whole thing was held together with duct tape.

When we resurrect this project in another time and space, I hope it’ll be with decent funding so we can put together a solid, polished unit so we don’t have to worry about setting things up on-site and can instead focus on the part that makes it interesting: the interaction between us and the participants.

You get better at it

The four of us had never made anything as a team, so it took us a while to learn who was good at what, and how we should settle disputes (like whether it was worth Owen risking broken bones to hang fairy lights from the top of the pole), and what we ultimately wanted.

We even got better at simple things that’ll carry over to next time, such as constructing, painting, ripping duct tape, convincing strangers, dealing with bureaucracy, porting heavy things and packing cars.

As Sas’s sister Melissa (an installation artist of renown) told us, art is just problem solving.

Next time we’re going to be even better.

I can’t wait for next time.

Thanks to Owen, Pat and Sas for reading over this post and for being incredible in general

A prototype speech synthesiser from Bell Labs in 1939. Found the...

Sat, 29 Dec 2012 07:45:08 +0000

A prototype speech synthesiser from Bell Labs in 1939. Found the video while researching my upcoming installation at Rainbow Serpent.

JavaScript Indentation

Thu, 06 Sep 2012 14:57:00 +0100

JavaScript doesn’t have a PEP8 equivalent. There’s no official rule about how code should be indented. Community styleguides (of which there are many) have no clear consensus. What’s a coder to do?

I wrote a script to clone the 100 most watched JS projects on GitHub¹ and determine the type of whitespace they use for indenting their .js files. Here’s a summary:

Compare this to the universal dominance of particular styles in other languages like Ruby and Python, and it seems bizarrely egalitarian. Is the JS community particularly fractious and disorganised?

The top ten projects tell a different story. Eight of them use two spaces (jQuery and three.js both use tabs.)

How do the results compare with advice from the most prominent styleguides?

Douglas Crockford: four spaces
Google: two spaces
jQuery: tabs
Idiomatic JavaScript: two spaces
GitHub: two spaces
NPM: two spaces
CoffeeScript output: two spaces

jQuery is used by more than half of all websites, and Douglas Crockford wrote Stack Overflow’s favourite JavaScript book, but it’s clear that they’re both out of touch with the JavaScript community.

The prevalence of the two spaces style is probably a result of the influence of the Ruby and Rails on web development, and of the callback-driven nature of browser / Node.js development. But I’d love to hear other suggestions as to why.²

For my next trick I’ll adapt my code (written in CoffeeScript, by the way) to investigate line length, trailing whitespace and other questions of JavaScript formatting. If you want to see where it goes, follow me on Twitter here.

Indentation is a matter of personal preference (it’s why they call it style, after all), but if you’re starting an open source project and want it to be consistent with community standards, indent with two spaces.

Join the discussion on Hacker News

Projects on GitHub might not reflect how most programmers (who work on commerical software) are writing their code. But seeing as most production JS is minified beyond recognition, it’s the best dataset we’ve got to work with. ↩
My friend and mentor Tim Diggins suggested investigating the influence that the server-side language has on JavaScript style. I’d guess that, e.g., projects using Python on the backend tend to use four spaces. ↩

Crowdfunding Political Donations

Sat, 17 Dec 2011 14:31:00 +0000

TL;DR version: Bottom-up Kickstarter for political issues. Let citizens pledge to donate money to political parties if they make certain promises.

Steve lives in London. The only political issue that he gets riled up about is gay rights, and more specifically same-sex marriage. This is understandable, because Steve can’t legally marry Nigel, his partner of ten years.

He wants the law to change. What are his options? He could:

Write a letter to his representative
Wait four years to vote
Sit at home and wait for market researchers to call him and ask for his opinions

But Steve is clued-in and cynical enough to know that when it comes to politics, money talks. Even if (as argued in Freakonomics) it doesn’t decide election outcomes, it certainly influences policy. Why else would corporations and individuals continue to donate so many millions?

The current government in the UK doesn’t have same-sex marriage on the agenda. The opposition party, Labour, have only made wishy-washy statements like this:

“It is right that we look at extending marriage equality for those people who want it. I welcome the announcement [of civil partnerships] and hope that progress can be made on this important issue as soon as possible.”

Steve could donate £50 to Labour, but for all they know he could be donating because he believes in higher tax rates for the wealthy. They don’t even ask why he’s donating. Ditto for the Liberal Democrats and the Conservatives.

So Steve goes to CrowdPolitics (a temporary working title, of course) and starts a pledge. If the Labour party promises to introduce legislation to legalise same-sex marriage, they’ll get his £50.

His friends are passionate about the issue too, and they tweet and blog and so on until there’s a significant pile of cash sitting there for the taking. Individual donations are huge. Of the $86 million Obama has raised for the 2012 election, 48% has been from “small individual” contributions.

If Labour won’t promise to introduce same-sex marriage legislation in their first term, Steve and co. will never get charged. Just like Kickstarter, it’s all-or-nothing. Labour isn’t against same-sex marriage. But the incentives aren’t there for them to start pushing the issue. Crowdfunding political pledges would provide clear, public, transparent and explicit incentives for them to do so.

Importantly, they never have to acknowledge that money is the reason why they made the promise. They won’t have to admit openly that their party policies are just another factor in a financial equation. They can make the promise and say it’s because they care about human rights and equality. It’s a win-win.

If Labour break their promise a few years later, they know that they’ll never receive another penny or vote from Steve or his friends. They’ll look like liars and cowards and frauds. Without an explicit statement of intent, it’s easy for them to backpeddle and make excuses and vague statements about their future policy.

The major problem I can see is this: who decides whether a promise fulfils a pledge, and how? The site moderators? The people who pledge? Will they each get one vote, or is it proportonal to how much money they donated?

I’d love some suggestions on how to solve that. Because I don’t see anything else getting in the way of making this work. Or am I missing something?

Join the discussion on Hacker News

Rapid vetting of startup ideas: IsThisCrazy.com

Tue, 20 Sep 2011 23:13:00 +0100

I have lots of dumb ideas for tech ventures.¹ When one hits me, I often think that it’s brilliant and world-changing and guaranteed to let me buy the solid gold house and rocket car I’ve always wanted. But after the psilocybin wears off and my friends calm me down, I usually see cracks in my plan², and put it on the shelf in my brain marked ‘not quite’.

This shelf is already pretty full. And the misshapen, broken toys lying there constantly beckon for me to pick them up and play with them, to turn them over in my mind and reconsider them from new angles. Because what if, right? What if there is a way to make them work, and I’m just not seeing it?

There’s part of me that doesn’t want anyone else to hear them. If they are as incredible as I hope they are, I want the accompanying fame and riches. I’d rather be Edison than Tesla. I’m a dragon who’s sitting on a pile of shiny things and jealously, suspiciously won’t let any jewelers inspect them for flaws. I’m also a guy who uses terrible metaphors.

Whatever I am, my curiosity is getting the better of me. I want to see if other people can innovate around the problems, or at least give me a quick, firm “What the fuck were you thinking?” There’s wisdom in them there crowds. It’s like Eric S. Raymond said (and possibly still says every so often): “given enough eyeballs, all bugs are shallow”.

Enough abstract talk. Let’s walk through how IsThisCrazy.com would work, with Uncle Quentin from the Famous Five³:

Quentin is struck by an idea for an online pet supplies store (thanks to Timmy, his niece Georgina’s loveable mongrel)
He condenses the basic idea down into a handful of sentences, and posts it from the main page of the app
After a little time, every famous one of the Famous Five gives feedback, in two ways: with TED-style ratings (‘fascinating’, ‘unconvincing’, etc.), and with freeform comments.
Seeing that the basic concept is not entirely without merit, Quentin provides more specific details (with images, videos, documents, etc.) on a separate page reachable from the summary view.
Julian and Dick are particularly keen on the idea, and, after seeing that Quentin is looking for collaborators with design and business skills, get in touch.

So Quentin ends up more confident in the potential of his scheme, more aware of possible problems, and with a core team of collaborators. This is the ideal scenario, but even if it got rejected unanimously it’d still be much better for him to see the abject folly in his idea instead of stewing over it for years, starting at Timmy and wondering “What If?”

I wish the site existed now so I could post this idea to it and see if it’s crazy or not.

The one I had tonight was for a kind of online poker where all of your winnings go to a charity of your choice (there’s a chance you’ll end up donating way more than you put in, and the only risk is that the money will go to some other entirely deserving cause) ↩
Maybe it’s already been done before, or maybe it depends on technology that doesn’t exist yet, or maybe no one in the world except me would ever use it. ↩
Because a) he’s a scientist and b) why the fuck not? ↩

Better Volunteering for Web Geeks

Sat, 17 Sep 2011 12:26:00 +0100

Summary for the impatient: let’s create a non-profit which feels like a startup, makes awesome stuff on the web for charities, and is exactly the kind of place where we’d volunteer.

Charities are great at building wells in African villages, but terrible at doing anything online. Fair enough: it’s not their speciality. But it does mean that they’re missing out on brilliant opportunities to get people engaged (and reaching for their wallets).

The biggest ones pay through the nose to corporate IT consultancies, but the vast majority just give their system administrator’s fourteen-year-old son free rein. Neither scenario is optimal.

Any of you could undoubtedly do better, even as a volunteer. But unless you’re a polymath with plenty of free time, you’re going to need a team to work with. And the staff at charities are generally overworked and undervalued, working with legacy codebases and bosses who don’t know the first thing about computers. Do you really want to spend your free time in that environment?

There needs to be a lean non-profit web agency that exclusively makes awesome stuff for other non-profits. You’d never have to deal with the bureaucracy or incompetence inherent in large organisations, and you’d work with smart people on the cutting edge of design, programming, marketing, and all the other disciplines needed to make a great online experience. It sounds just like a startup, doesn’t it?

If you’re going to save the world, there’s no reason why you can’t have fun and learn from your peers in the process. It’d be a great way to pad your portfolio or github profile (obviously we’d open source our work as much as possible) as well as meet potential employees, employers, co-founders, mentors and friends.

There are problems. First: the inevitably high volunteer turnover rate. When push comes to shove, your charity work is never going to take priority over your family life and or your job. Although high turnover means a steady stream of new faces and ideas, it also creates significant overhead as new volunteers are brought up to speed.

There are ways to mitigate this, aside from building the kind of place you’d prioritise over your job. Keeping a minimal paid staff would ensure that at least someone on the team has been there from the beginning of the project, and knows what needs to be done and who can do it.

But who pays their salaries? Soliciting donations from the public would be tough. Oxfam struggles to get people to reach for their wallets, and they use photos of starving children. Ours would only feature a fridge running low on diet coke. Corporate donations from tech companies are more likely. This is exactly the kind of thing Google would get behind.

We could also charge our clients fees. We’d make sure it’s cheaper than outsourcing it to an IT firm or hiring in-house staff. We could undercut the going rate hugely, because most of the work would be done by volunteers, and because we’re not looking to turn a profit. And who knows? There are probably government grants for this kind of thing.

The first few projects might need to be done for free, until we’ve built up a reputation and a portfolio to convince organisations that we’re worth it. We’re not going to be working with Oxfam right off the bat. We’ll start small, explore the space and find a viable ‘business model’ before scaling.

After a few projects, we’d have considerable expertise at tugging heartstrings, recruiting volunteers, increasing donations, and all other facets of the unique (but not totally unfamiliar) problem set faced by charities. Why would they want to hire anyone else?

Part of why I’m so excited about this is because I know that London is ready for this. It’s teeming with enthusiastic, talented designers and developers at startup hubs, hackspaces, dojos, meetups, user groups, and, most importantly, pubs. It also has more NGOs, non-profits and charities than any other city on the planet.

So let’s do it before those kids in Silicon Valley beat us to the punch.

(Discuss on Hacker News)

jam.ly

Fri, 15 Jul 2011 23:35:52 +0100

Not all bands meet in high school or university, or through friends. There’s a reason NME and Rolling Stone devote space to classifieds written by groups desperate for bass players, and why there are thousands of similar (but more wackily typeset) flyers flapping on noticeboards in Shoreditch and Williamsburg, the hipster capitals of the world.

The Internet is obviously a better means to this end. When was the last time you took a flyer or looked at the classifieds? The idea of a site to hook musicians up isn’t original, but it is so far poorly implemented. The ones I could find suffer from ugly design, an extremely narrow target demographic or having even fewer features than Craigslist. The market is clearly humungous. Let’s see if we can do better.

Call the site jam.ly for now. Its users are either existing bands looking for new members, or lonely musicians looking to join or form bands. Each user keeps a profile, and browses and searches through the others to find someone to jam with.

jam.ly automatically recommends the best matches to the user, to eliminate time spent browsing through profiles of people they’re clearly never going to want to meet. OKCupid (the only dating site run by Harvard maths nerds) is a good model here. It asks pointed questions (e.g. “How frequently do you bathe or shower?”), and matches users on how similarly they answer, along with some other crucial factors like gender and age.

jam.ly can run along the same lines, except with music-related questions and special attention paid to things like:

Skill: there are a lot of awful “musicians” out there, and time shouldn’t be wasted meeting them. Embedded videos in profiles should help separate the wheat from the chaff, especially if you allow users to rate each other.¹
Taste: some people are looking to start the next great Weezer cover band². Others want a platonic soul mate to get high with them and write a concept album about space. The site should hook into their Last.fm account³ to give a compatibility rating, but they’ll also list the kind of stuff they want to play.
Location: this is huge. How far away from do they live? Are they a five minute tram ride away or a 50km drive from one side of the city to the other? Obviously beyond a certain distance there’s no point in ever seeing their profile.
Commitment: how many hours a week are they will to put into this? How willing and able are they to leave their jobs and travel to other cities to tour, if it comes to that?
Equipment: do they already have a gig-worthy rig, or do they practise on a micro amp? Do they have a space to jam in, or any recording equipment?
Experience: past performance predicts future performance. If they’ve been in a successful band before, they should be able to provide references from that band with links to myspace or bandcamp or whatever it is the kids use these days.

That’s all I can think of for now. Tell me if I’ve missed anything important.

I know people who spend hours watching videos of other guitarists on forums and YouTube, giving feedback and criticism. This is just like that, except they’ll feel like they’re making even more of a difference. ↩
The current holders of that title are called Meezer, and they are my heroes. ↩
Here’s mine. By the way, Last.fm is based in the aforementioned hipster-mecca Shoreditch. They’re nice guys, but are still using PHP. No one told them that 1995 called and it wants its programming language back. ↩