Unhosted: a protocol for decentralized web apps

fair enough, still, it never hurts to have free software alternatives to commercial software. in installed software, we are already used to having the option to use linux, OpenOffice, mysql, etcetera. in hosted software, currently, the only alternatives to commercial websites are other commercial websites. with unhosted, we're simply trying to copy free software's success from installed software to the web, and that way we hope to expand people's options to vote with their traffic.

You don't see anything wrong with that? Really? Then, either you're an ignorant fool, or you're okay with (among other things) Gmail analysing your whole correspondence, Hotmail looking at the cookies of its partners' websites, Facebook's ever-changing un-privacy policy, Google keeping web searches for months in a not-at-all-anonymous way, and users ignoring a large parts of all that. Or you didn't really try to see what's wrong. Or you're lying to yourself, or to us. Or you casually threw a cached thought to make your point. Or something.

Also, if people don't vote with their traffic, that's because they're hooked. Immediate costs (switching) are always overestimated, and long term benefits (privacy) are always underestimated. Only perceived benefits and costs can influence a decision. People's perceptions are off, therefore they make bad decisions, therefore big companies own them.

Finally, if we had symmetrical, unfiltered broadband from the start, along with easy to use mail and web servers, then self hosting would have been ubiquitous by now. Gmail, Blogger, and YouTube wouldn't even exist (search engines still resist to decentralization, though). That is the way the entire internet (not only the web) intended to remain (not just be: back when it wasn't widespread, it was fully decentralized).

You are tired of seeing this refrain? I, am tired of seeing comments that assume it's false, without having the guts to state it clearly. Really, do you actually think the web was intended to be roughly a limited number of high traffic web site? Side question: do you think that's better than a more evenly distributed model? Or even good? Personally, I think it's unintended, worse, and bad. You know some of my arguments. I have another one: raw efficiency. A centralized web creates choke points in the network, and hurts peerage agreements. That is costly to ISPs, which then are tempted to relinquish Internet Neutrality to recover the loss. Carry out that trend to its ultimate conclusion, and the Internet as we know it will be replaced by an AOL-like network. Such a thing will still be called "internet", and it will still be IP based. But it would have lost it's most interesting property: letting the people write freely.

> and it's natural for users to flock to something that is good

In fact, it's natural for users to flock around the dancing bunnies.

http://en.wikipedia.org/wiki/Dancing_pigs

You know, I'm not one to complain about duplication, but I think it's a bit excessive to have the same thing submitted three times in less than a month.

Sorry, didn't know it was a dupe. I figured hackernews would have told me when I tried to submit it.