The tech industry is one of the most valuable sectors on the planet, but it ultimately rests on the unpaid labour of an alarmingly small amount of hobbyists.
That reliance is the blessing, and the curse, of open source software – coding projects put up on the internet for anyone to use, freely, in their own work. Some open source software solves simple problems elegantly enough that no one wants to redo the work unnecessarily; others tackle complex tasks that few have ever attempted.
This is no secret. In August 2020, you may have seen webcomic xkcd comment on the problem, with a cartoon depicting “all modern digital infrastructure” as a wobbly tower of bricks relying on “a project some random person in Nebraska has been thanklessly maintaining since 2003”.
Or, a year later, a viral tweet by Druthers Haver:
the most consequential figures in the tech world are half guys like steve jobs and bill gates and half some guy named ronald who maintains a unix tool called ‘runk’ which stands for Ronald’s Universal Number Kounter and handles all math for every machine on earth
The jokes are based on reality. In 2016, a developer was told to change the name for a piece of software of his called kik, after the creators of an unrelated messaging app with the same name asked for the name to be assigned to them. In response, he unpublished kik, as well as everything else he’d made – which included a small piece of code called left-pad, which was used by thousands of other programs. Owing to the way left-pad was distributed, through a service called a “package manager”, those programs began failing in minutes, as they tried to update themselves to a new version of left-pad that didn’t exist. And then the programs that incorporated those programs began to fail as well. It took two and a half hours to catch and fully fix the problem.
Before that, there was heartbleed. In 2014, researchers found a crippling bug in a piece of software called OpenSSL, which websites used to enable the encryption standard that keeps credit cards and passwords safe online. Maintained, again, by a few volunteer developers, OpenSSL solved a common problem simply and easily, meaning that it was the go-to way of adding good-enough security to your website. Except it wasn’t. The developers hadn’t noticed a bug in the code which could be used to access data that was supposed to be encrypted – and which had been sitting there for more than two years. Seven years later, the same basic thing happened again, in a different open source utility called Log4j.
Giving software away for free is great for a whole host of reasons – but quite bad at funding continued development of that software. There have been loads of attempts to fix that, from models of development where the software is free but the support is paid, to big companies directly hiring maintainers of important open-source projects. Lots of projects have ended up in a tip or supporter-focused model (remind you of anyone?), which can work for big complex tasks but falls down for some of the simplest – yet most widely used – pieces of work.
And then, there’s this:
A project called tea.xyz promised people they could “get rewards for [their] open source contributions”, complete with a flashy website describing how it would “enhance the sustainability of open-source software”.
So far, it’s achieved the exact opposite. Promising to reward open source contributors with crypto tokens, the project asked users to verify their access to open source projects by merging in a YAML file containing their crypto wallet address.
What was supposed to happen: a crypto-based system would give you micropayments for working on open source software, and redistribute tokens from users to contributors.
What actually happened: crypto fans flooded open source projects with low-effort alterations in an effort to become the “owner” of various highly used projects, resulting in already overwhelmed volunteer coders having to sift through spam and worse rather than work, or raise money.
It’s good that there’s growing awareness of the problem, but the solution clearly needs some work.
The curious case of the hundred bucks
Apple has been hit with its first-ever fine from the EU. At “over €1.8bn”, it’s straight in at number three in the bloc’s highest penalties ever. From our story:
The fine is nearly four times higher than expected in a move by the European Commission to show it will act decisively on tech companies who abuse their dominant position in the market for phones and online services.
The European competition commissioner, Margrethe Vestager, said a smaller fine would have been nothing more than the equivalent of “a parking fine” and was designed to act as “a deterrent” to such practices for Apple and others.
As a result of the anti-competitive practices, the public ended up paying more than they should have for music streaming, she said.
“Apple’s rules ended up harming consumers. Critical information was withheld so that consumers could not effectively use or make informed choices. Some consumers may have paid more because they weren’t aware that they can pay less if they subscribed outside of the app.”
The EU’s rules set the level of the fine at “15-25%” of the affected revenues, which would have been peanuts given the scale of Apple Music in Europe, so the regulator added a lump sum of €1.8bn to top it up to an amount that may actually provide some deterrence.
after newsletter promotion
We’ve covered Apple a lot here at TechScape recently (earlier on this very fine, and on the company’s compliance with the digital markets act, coming into force on Thursday), so I won’t dote on it too much. Except there is something weird here.
In its response, Apple vowed to appeal the ruling, and was insistent that its App Store model provides good value for developers – allowing the likes of Spotify to grow to the scale it is without ever paying a commission on its revenue from iPhone users.
And there’s something bugging me. Here are a few quotes from Apple’s lengthy public reply to the EU:
If a developer sells physical goods, serves ads in their app, or just shares an app for free, they don’t pay Apple anything.
When it comes to doing business, not everyone’s going to agree on the best deal. But it sure is hard to beat free. But free isn’t enough for Spotify.
Today, Spotify has a 56 percent share of Europe’s music streaming market – more than double their closest competitor’s – and pays Apple nothing for the services that have helped make them one of the most recognizable brands in the world.
Despite that success, and the App Store’s role in making it possible, Spotify pays Apple nothing.
And again – Spotify pays Apple nothing.
Except … Spotify, like every business on the App Store, pays Apple a $99 annual fee to be part of the Apple Developer Program.
In the context of a clash between two multibillion-dollar companies, $99 a year rounds to zero, of course. But, with 34 million registered developers, this is a business worth at least $100m a year to Apple (since some developers, including nonprofits and government agencies, do get a waiver).
I get it – saying “Spotify pays Apple nothing” is a much stronger lobbying position than “Spotify pays Apple just $99 a year, the same as every other developer”. But only if it is, you know, true. And Apple makes hundreds of millions of dollars a year from charging that fee to developers as standard, which complicates the narrative that the App Store is only funded by commission on sales.
And, of course, Spotify pays Apple millions of dollars a year for the devices it uses to create software for the company’s App Store. You cannot develop iOS apps without a Mac. Is it possible that Spotify acquires every Mac and iOS device it uses in the course of its operations second-hand, deliberately avoiding handing any money to Apple? Sure, I guess. Is it likely? Come on now.
I tried asking Apple how they squared this, and a spokesperson repeated the claim that Spotify paid $0 to Apple. When I asked if I could explicitly write that “Apple claimed that Spotify is not charged the developer fee”, though, the company stopped replying to my emails. Spotify had no such bashfulness, and confirmed that they pay the fee like all major developers.
There’s a chance this is simply me getting unhealthily obsessed with a point of principle. As a journalist, I value truth, and bristle at spin that, in my opinion, crosses the line into falsehoods. Maybe none of this matters. But… it doesn’t look good, does it?