5 ms·

Well sure, but who owns the database? Where is it hosted and who vouches for its accuracy? What happens if it goes down or its maintainer goes out of business?

Well sure, but who owns the database? Where is it hosted and who vouches for its accuracy? What happens if it goes down or its maintainer goes out of business? Of course, all these questions can be addressed in one way or another with conventional database tech. But blockchain has some undeniable appeal in an application like this.

dmitriid 9y ago
It doesn’t. Because you assume that blockchain will magically identify and attribute every single of 30 million songs in Spotify’s database.
- dmitriid 9y ago
  Oh. I just realized another thing. 30 million songs can easily fit into a database running on a Macbook Air. Stick it on a medium-sized server, and you can already serve thousands of users. Blockchain meanwhile... something something ... 4 txns/second ... 1 terrawatt per txn ... up to several days to clear a txn when network is congested
  - daics 9y ago
    UMG, Warner, Sony and Spotify cannot share bills for running a common database on AWS today without creating yet another corporation and relinquishing power over the data to it. Blockchains are not the next big thing in databases. They are here to revolutionize how existing organizations co-ordinate, enter into contracts with each other, and even allow individuals themselves co-ordinate directly with each other without intermediaries through novel org structures. You are trading computational scalability for social scalability. http://unenumerated.blogspot.com/2017/02/money-blockchains-and-social-scalability.html http://unenumerated.blogspot.com/2017/02/money-blockchains-a...
    - dmitriid 9y ago
      > Blockchains are not the next big thing in databases Blockchains are the next big thing in nothing. They are not here to revolutionise anything [1] > hey are here to revolutionize how existing organizations co-ordinate, enter into contracts with each other etc. Nope. There are about zero things in blockchain that help with that. Because at the end of the day someone has to do all the job of, you know, adding all the info about who owns what percentage of what song in which region of the world. Oh wait. Which song was that? A Japanese LP that's 2 seconds longer than the original single released in Europe? Or that remastered song on a "Best of album" that was published by a different publisher than the remastered version published on the "Remastered" album that is published by a different combination of publishers than the original 1960s album that is different from that singular French copy of a concert in 1995... All that for a single song that ends up being attributed to the same singers and songwriters, except that one cover by all the same people sans that one guy, and except that different cover that will be attributed to the same songwriters, but a different singer, and except... All that info has to be: standardized, assigned to every single one of those 30 million songs, and any new releases should contain the same standardized info. You know, something the publishers could agree on right now, and they don't. So how in the seven hells is blockchain going to help? [1] https://hackernoon.com/ten-years-in-nobody-has-come-up-with-a-use-case-for-blockchain-ee98c180100 https://hackernoon.com/ten-years-in-nobody-has-come-up-with-...
      - lt 9y ago
        > Blockchains are the next big thing in nothing. They are not here to revolutionise anything [1] I beg to differ [1] > So how in the seven hells is blockchain going to help? It's not a silver bullet, and most things are not ones that were not possible before. But it facilitates some combinations, makes some things a lot more practical, and opens space for innovation. In this case, in particular, what comes to mind is: - there's a replicated consensus of all this attributions. Every single company agrees and has a shared database of the rights and the licensing. - the disintermediation: there's not necessarily a need to have a middle man managing this informations and agreements. - auditable, notorized, unforgeable history: you have this immutable record of when songs were released, who held their records, who and when licensed them, etc. - open information: if you design this system as an open network, any interested party can join and get the information it wants without needing APIs, etc. All this without getting into tokens and handling the payments and distribution of royalties on the chain, and other innovations that the capabilities of the blockchain can bring. [1] https://blockchaintechguide.com/a-blockchain-based-future/ https://blockchaintechguide.com/a-blockchain-based-future/
  - lt 9y ago
    > Blockchain meanwhile... something something ... 4 txns/second ... 1 terrawatt per txn ... up to several days to clear a txn when network is congested Don't mistake Bitcoin's limitations for Blockchain limitations. These are characteristics of the consensus mechanism, of Proof of Work in particular.
zaphar 9y ago
If by blockchain you mean a merkle tree then sure I might agree with you. If by blockchain you mean one of the various crypto currencies with a consensus protocol then I don't really see the appeal. A merkle tree has some useful properties for verifying the integrity of your data cryptographically. As such it is a useful way to distribute a ledger. But in this particular domain there is no double spend problem to worry about so proof of work and most other consensus protocols are... wasteful at best.

dmitriid 9y ago

It doesn’t. Because you assume that blockchain will magically identify and attribute every single of 30 million songs in Spotify’s database.

dmitriid 9y ago
Oh. I just realized another thing. 30 million songs can easily fit into a database running on a Macbook Air. Stick it on a medium-sized server, and you can already serve thousands of users. Blockchain meanwhile... something something ... 4 txns/second ... 1 terrawatt per txn ... up to several days to clear a txn when network is congested
- daics 9y ago
  UMG, Warner, Sony and Spotify cannot share bills for running a common database on AWS today without creating yet another corporation and relinquishing power over the data to it. Blockchains are not the next big thing in databases. They are here to revolutionize how existing organizations co-ordinate, enter into contracts with each other, and even allow individuals themselves co-ordinate directly with each other without intermediaries through novel org structures. You are trading computational scalability for social scalability. http://unenumerated.blogspot.com/2017/02/money-blockchains-and-social-scalability.html http://unenumerated.blogspot.com/2017/02/money-blockchains-a...
  - dmitriid 9y ago
    > Blockchains are not the next big thing in databases Blockchains are the next big thing in nothing. They are not here to revolutionise anything [1] > hey are here to revolutionize how existing organizations co-ordinate, enter into contracts with each other etc. Nope. There are about zero things in blockchain that help with that. Because at the end of the day someone has to do all the job of, you know, adding all the info about who owns what percentage of what song in which region of the world. Oh wait. Which song was that? A Japanese LP that's 2 seconds longer than the original single released in Europe? Or that remastered song on a "Best of album" that was published by a different publisher than the remastered version published on the "Remastered" album that is published by a different combination of publishers than the original 1960s album that is different from that singular French copy of a concert in 1995... All that for a single song that ends up being attributed to the same singers and songwriters, except that one cover by all the same people sans that one guy, and except that different cover that will be attributed to the same songwriters, but a different singer, and except... All that info has to be: standardized, assigned to every single one of those 30 million songs, and any new releases should contain the same standardized info. You know, something the publishers could agree on right now, and they don't. So how in the seven hells is blockchain going to help? [1] https://hackernoon.com/ten-years-in-nobody-has-come-up-with-a-use-case-for-blockchain-ee98c180100 https://hackernoon.com/ten-years-in-nobody-has-come-up-with-...
    - lt 9y ago
      > Blockchains are the next big thing in nothing. They are not here to revolutionise anything [1] I beg to differ [1] > So how in the seven hells is blockchain going to help? It's not a silver bullet, and most things are not ones that were not possible before. But it facilitates some combinations, makes some things a lot more practical, and opens space for innovation. In this case, in particular, what comes to mind is: - there's a replicated consensus of all this attributions. Every single company agrees and has a shared database of the rights and the licensing. - the disintermediation: there's not necessarily a need to have a middle man managing this informations and agreements. - auditable, notorized, unforgeable history: you have this immutable record of when songs were released, who held their records, who and when licensed them, etc. - open information: if you design this system as an open network, any interested party can join and get the information it wants without needing APIs, etc. All this without getting into tokens and handling the payments and distribution of royalties on the chain, and other innovations that the capabilities of the blockchain can bring. [1] https://blockchaintechguide.com/a-blockchain-based-future/ https://blockchaintechguide.com/a-blockchain-based-future/
- lt 9y ago
  > Blockchain meanwhile... something something ... 4 txns/second ... 1 terrawatt per txn ... up to several days to clear a txn when network is congested Don't mistake Bitcoin's limitations for Blockchain limitations. These are characteristics of the consensus mechanism, of Proof of Work in particular.

zaphar 9y ago

If by blockchain you mean a merkle tree then sure I might agree with you. If by blockchain you mean one of the various crypto currencies with a consensus protocol then I don't really see the appeal. A merkle tree has some useful properties for verifying the integrity of your data cryptographically. As such it is a useful way to distribute a ledger. But in this particular domain there is no double spend problem to worry about so proof of work and most other consensus protocols are... wasteful at best.