[Bobboau]

It's time for the Permanent Web

OMGWTFDRIVEBYPOSTOMGWTF!!!!!

[headdie]

So it's a variation on the idea of torrenting the web?

[Bobboau]

sorta, except it includes versioning and is focused more on persistence than simply distribution and is intended to be used for distributing websites and website content, rather than just simple file distribution.

[jr2]

I like this. 

[karajorma]

In general this seems like a good idea, but a couple of things worry me about it.

One of the big ones though is updating things. If the original link never goes dead you're missing a big ****ing clue that you're looking at something outdated. If I go to www.somewebsite.com/SomeModName/HereIsTheRelease.zip and I get a 404 the first thing I do is try www.somewebsite.com/SomeModName or www.somewebsite.com and see if the mod is still available elsewhere on the same website. Now unless someone actually goes to the trouble to update links, I'll still find the old version of the mod without any clue how old it may be.

Secondly, I do worry about how lazy it's going to make people about backing up stuff. After all, there's no need to back up your website when it's all distributed forever. Until of course a hacker gets hold of your website and makes all the copies expire.

[headdie]

So
a) how do you keep the files updates without incessantly polling the source
b) how to you or the system tell the source data from the duplicated data?

I have another point, how does this system relate to copyright and intellectual property law

[Luis Dias]

In general this seems like a good idea, but a couple of things worry me about it.

One of the big ones though is updating things. If the original link never goes dead you're missing a big ****ing clue that you're looking at something outdated. If I go to www.somewebsite.com/SomeModName/HereIsTheRelease.zip and I get a 404 the first thing I do is try www.somewebsite.com/SomeModName or www.somewebsite.com and see if the mod is still available elsewhere on the same website. Now unless someone actually goes to the trouble to update links, I'll still find the old version of the mod without any clue how old it may be.

Secondly, I do worry about how lazy it's going to make people about backing up stuff. After all, there's no need to back up your website when it's all distributed forever. Until of course a hacker gets hold of your website and makes all the copies expire.

These seem non problems to me. "Oh the problem of this stuff is that it's too robust AND too convenient", yeah? well sign me up

[AdmiralRalwood]

In general this seems like a good idea, but a couple of things worry me about it.

One of the big ones though is updating things. If the original link never goes dead you're missing a big ****ing clue that you're looking at something outdated. If I go to www.somewebsite.com/SomeModName/HereIsTheRelease.zip and I get a 404 the first thing I do is try www.somewebsite.com/SomeModName or www.somewebsite.com and see if the mod is still available elsewhere on the same website. Now unless someone actually goes to the trouble to update links, I'll still find the old version of the mod without any clue how old it may be.

This would seem to be the point of the (incomplete) IPNS system:

IPNS allows you to use a private key to sign a reference to the IPFS hash representing the latest version of your site using a public key hash (pubkeyhash for short).
[...]
IPNS isn't done yet, so if that link doesn't work, don't fret. Just know that I will be able to change what that pubkeyhash points to, but the pubkeyhash will always remain the same. When it's done, it will solve the site updating problem.

Secondly, I do worry about how lazy it's going to make people about backing up stuff. After all, there's no need to back up your website when it's all distributed forever. Until of course a hacker gets hold of your website and makes all the copies expire.

I'm not sure what this is supposed to mean, though; as far as I can tell, copies never expire. A hacker would have to gain access to every single node that holds a copy of the website and delete all such copies.

The real problem in my mind is what, exactly, incentivizes people to make those copies in the first place? One can imagine people wanting backups of, say, websites dedicated to their favourite space combat sim, but suppose...

Suppose someone makes a personal site for, say, their fanfiction collection that nobody reads. Nobody makes copies of this website, their originals eventually go down, and you're left with exactly the scenario that this scheme is supposed to avoid: broken links and error pages because the only copy of the website no longer exists. Now, one might reasonably ask, "if nobody read the stupid fanfiction, who cares that it's gone?" Well, maybe nobody read it because nobody hard heard of the person but then they later publish a novel, only they don't have backups of all their older work, so even though people might be interested in it now, no one can access it. Even supposing nobody ever had any reason to care about this person's terrible fanfiction, according to the original author, even terrible websites are a part of our history:

Whether eroding content is questionable crap or timelessly useful, it's still our history, and we're losing it fast.

So, are "little guy" independent websites like this hypothetical example purely dependent on the generosity of others? If so, well, all I can say is... let's see how long that lasts in the real world. Then again, perhaps this "worst case scenario" is no worse off than the status quo, so... I guess we'll see what happens.

[karajorma]

This would seem to be the point of the (incomplete) IPNS system:

IPNS allows you to use a private key to sign a reference to the IPFS hash representing the latest version of your site using a public key hash (pubkeyhash for short).
[...]
IPNS isn't done yet, so if that link doesn't work, don't fret. Just know that I will be able to change what that pubkeyhash points to, but the pubkeyhash will always remain the same. When it's done, it will solve the site updating problem.

I was on about the IPNS system in both of these actually. Although the IPNS system will allow you to change the public hash, that only works if you get people who bother to do so, I suspect a great many people won't bother. What the system needs is a way of updating how old the hash is. Then you would just make your html software update the ages and you would know if you're downloading from a link and it says it's 6 years old that perhaps there is a newer version.

Secondly, I do worry about how lazy it's going to make people about backing up stuff. After all, there's no need to back up your website when it's all distributed forever. Until of course a hacker gets hold of your website and makes all the copies expire.

I'm not sure what this is supposed to mean, though; as far as I can tell, copies never expire. A hacker would have to gain access to every single node that holds a copy of the website and delete all such copies.[/quote]

My point was that using IPNS a hacker might be able to set the file to a different one for your entire website. In theory that should mean that all current copies of the file would then be flagged as old and if not backed up by their owner, would vanish from the internet. If they don't vanish, we're back to my first issue.

[Bobboau]

they wouldn't vanish, they would just loose their human readable name. think of it like putting the entire internet into one huge version control system.