3 ms·
I used to love the Web and APIs. Still think it’s great. But I think it’s far superior to download software (Merkle trees) onto your own computer automatically
by EGreg 8y ago
I used to love the Web and APIs. Still think it’s great.
But I think it’s far superior to download software (Merkle trees) onto your own computer automatically and run it in a sandbox with your own data, while doing end to end encryption. That way you don’t worry if the API goes down or starts charging more or the database is hacked and you don’t have to jump through holps to get 1000 keys for each new project. Also you can resell your project to regular people who can host it themselves without them jumping through all those hoops. And finally, hosting should consist of provisioning space in a giant abstract cloud that holds encrypted data, no need to set up Linux instances or whatever.
Which platforms are doing that?
- theamk 8y agoIf you get rid of "merkle tree" requirement, then you have just described database files hosted on github. It does plenty of advantages -- it scales good and if done right, it is immune to hacking or site going down. But it has some disadvantages as well -- the biggest is that client needs to do non-trivial, complex work, so you cannot just insert AJAX call onto your webpage. Also, you lose all control of data, so a competitor can trivially clone your service.
- EGreg 8y agoWell yeah, I am not talking about Capitalism in the traditional sense anymore (business builds a moat, provides a service to you and charges fees). I am talking about a collaboration economy (open source, science style). You may still be able to sell stuff, but not the stuff whose marginal cost to copy is zero. Information would be free. Asked who would generate it, the answer is increasingly “everyone”. News and Maps can be crowdsourced (see openstreetmap for example). The hard part is in the consensus model, ie data sets will be honeypots for people who want to alter the “truth as we know it” (like Stephen Colbert and Elephants on Wikipedia). I am not convinced that a for-profit top-down corporation is necessarily better for that than a flat, democratic crowd, provided we have interesting game mechanics built in (like high stakes challenges where you can lose a significant portion of your accumulated reputation if you don’t win by a large margin).
- ocdtrekkie 8y agoI definitely strongly prefer locally-stored data for a bunch of reasons, like the fact that nobody can analyze what queries you made on that data. But there is a lot of data that is regularly new and obviously has to be pulled from the internet regularly, such as news or weather data.
- EGreg 8y agoIsn’t sync/mirroring the proper tool for that? The data itself may of course be monetized by the various sources but if one goes down there can be others, as long as the protocol is resilient to that.