Business News Digital Education & Events Industry People The Great Escape 2016

CMU@TGE: The data pioneers – Robert Kaye, MusicBrainz

By | Published on Tuesday 14 June 2016

Robert Kaye

Look out for insights, advice and viewpoints dished out at this year’s CMU Insights @ The Great Escape conference here in the CMU Daily throughout June. This week, some of the takeaways from the data focused strand.

MusicBrainz is an open source database of music metadata, launched in 2000. Since then, its users have helped to build a vast database filled with information on artists and their releases, including ISRC and ISWC codes and other musicians that the performers are related to. Organisations including the BBC and Google use the database as a provider of music data. Speaking at CMU@TGE, Founder Robert Kaye explained that the database is a repository for what he refers to as “public metadata”.

“Public metadata are the types of things you can find on the back of a CD”, he said. “So that’s public information that’s out there somewhere. ISRCs are sometimes on the CDs, and we can collect them through there. Lots of our contributors will paw through this or that website – people know where all of the music societies websites are – and just one after another enter all this data in”.

What the platform does not have, however, is the music rights data that is not made publicly available, such as ownership splits between collaborating songwriters. But, said Kaye, “it wouldn’t make sense for us to do this, because any time that you have an open system – and we’re aggressively open – well, an open system that dictates how money flows is a recipe for disaster, because then every day you’ll have ten Beyonces showing up saying, ‘Hey! I’m Beyonce’. That’s a real problem”.

“I would like to see a system that’s built on MusicBrainz to create that kind of rights database”, he added. “I’d really love to see that, but it would very clearly need to be a separate project that is not an open project”.

Now, many among you probably just thought, “Aha! The bloomin blockchain! That’ll do that” Well, not so fast, buzzword fiends.

“I really buy into the motivation of the blockchain”, said Kaye. “It’s a really good idea to make this information publicly available, publicly verifiable, publicly searchable. We desperately need such a system. By simply saying, ‘Hey, I’m the creator of this piece of music’, and stating that before you even release it, is just about the only way to avoid conflicts that you might have later, that, years on, are going to be solved by a judge, who doesn’t really care about these things”.

There’s a ‘but’ coming though, and that’s the problem of the complexity in any such system. As a “marginally competent computer scientist” Kaye noted that the basis of any project of this type is that “you have to prove the rigor of the algorithm you want to use”.

“When you’re looking at a distributed public ledger [like the blockchain], it’s comprised of three things because there’s three words in it: distributed, public, ledger. Okay, ledger is easy. Public? Yeah, we know how to do public. No problem. Distributed? Exceptionally hard”.

He continued: “Computer scientists have been tinkering on this really hard for the last ten years not making a lot of progress, because out of the P2P era, which is where MusicBrainz was born from, there was so much talk about distributing things, but can anybody name me one distributed system that you know of that’s not Bitcoin, nor is it BitTorrent? Anybody know of a distributed system? There are zero hands going up, because these things just simply don’t exist, because they’re effectively incredibly difficult to write”.

While the success of Bitcoin is often cited as proof that a music-related blockchain system could work, Kaye also questioned whether Bitcoin is actually as successful as many think it is. It isn’t fast enough to match the quantity of transactions that go through a mainstream financial system he said, and therefore “Bitcoin, as it stands in the blockchain, simply doesn’t scale. It doesn’t work. So it’s not a proven system”.

However, he conceded, scaling is less important for music’s potential blockchains than it is for Bitcoin. “Scaling’s obviously important, but it doesn’t have to be on the same scale as a financial system. It can be significantly less, because we’re only doing so many licensing deals, as opposed to paying for French fries a million times a day”.

But there’s still a but. “The problem with these types of systems is, if you’re building a cryptography algorithm, it’s not the sort of thing where you can say, ‘Oh, well if we find problems we’ll fix them later’. I mean, how many of you guys know that What’s App is now end-to-end encrypted? That’s fabulous, that’s really good, but if it’s based on faulty cryptography and that cryptography is broken, all of your messages that you thought were secure – all of your sexting messages and so forth – no, they’re all going to be uploaded to a server and you’re all busted. That’s really bad”.

Referencing PledgeMusic’s Benji Rogers, who has become one of the leading proponents of the blockchain for music, and who keynoted in the same strand at CMU@TGE, Kaye said: “I love Benji, he’s a great guy, but he’s not a cryptographer. He’s not even a computer scientist. So the things that he talks about, and the things he strings together into a coherent set of ideas, sound really nice, but are technically not as sound as they need to be. We have to do something. It’s a really good idea, and I really operate on that principal, but if you’re talking about cryptography you can’t actually move forward unless you have rigorous proof that your system is actually going to work. And that’s not here yet”.

So, a healthy dose of cynicism in the music industry’s big data and blockchain debate, which we’ll hear more of later this week.