[kickscondor]

So - there is no server-executed code - it all runs in the browser and the folder can only access itself anyway, which can't happen unless you have the private key.

Access control in Beaker is through that private key - you need it in order to edit the 'dat' (name for a synced folder). So, no, there aren't a lot of complex permissions available - but you can also separate an app into several dats and use a master one to manage the permissions of those. Not terribly complex, but it's actually surprising how much you can do. (It's tough to wrap your head around not having a server - but it's actually true.)

But help me out - I think alot of the Web falls into this category:

* User logs in to edit their data (has private key to their dat). * User shares their data (blog, photo feed, whatever) with others (who don't have the key). * Those others merge all incoming feeds into a single master feed.

You could replicate YouTube, Facebook, Twitter this way - usually there are not complex permissions in these apps, are there? (Not that you'd want to replicate them...)

[zmw]

Umm, having everyone’s data as a flat directory as opposed to an aggregated database sounds terribly inefficient... You need to somehow build a distributed, decentralized database on top of that flat structure, right? Otherwise your Twitter is just a microblog publishing tool plus a direct crawl RSS reader...

Maybe Twitter is too specialized an example. What about any kind of search? You do need an index, and someone still has to own that index, and “donate” computing power to update that index. You own your self-hosted data, like many of us already do, but there will still be gatekeepers, e.g. Google for our current web.

EDIT: I realized that with a clever enough architecture and probably much more computing power than necessary in a trusted environment, no one needs to own the index. But it seems way more advanced than this protocol. (I’m completely new to this stuff so please excuse my naive skepticism.)

[kickscondor]

No, no - I understand on several fronts: first, there is just so much technology these days, it's tough to find anything that isn't just a fleeting thing; also, you're absolutely right that you can't just solve everything with a distributed filesystem.

I also am not sure what yourself (or newnewpdro) are looking for in the web or what appeals to you - for me, Google simply doesn't work for me - sure for technical issues it does, but it is basically Stack Overflow search in that department. If I'm looking for personal blogs, I can't just type "personal blog" into Google and find anything worthwhile - it's all clickbait of a fashion. The best way I've found of finding blogs is either to look through Pinboard tags or to click around on other blogs until I eventually get somewhere. It's horribly inefficient - but it's rewarding when I get there. I'm making a personally-edited blog directory to try to aid discovery - and yeah I actually think there's a lot we can do if we all did more grassroots search and directories. Anyway, that's my perspective - wondering what you're looking for in this thread. Have enjoyed your other questions above (below?)

[newnewpdro]

> Access control in Beaker is through that private key - you need it in order to edit the 'dat' (name for a synced folder).

You're referring to write access, which is a small subset of access control.

How do you restrict read access to a group of specific people? Encrypt the data and distribute keys to the privileged parties? How does revocation work?