Stumbled upon this (disclaimer, I'm not affiliated and don't hold any special interest):
Aether is an app you install to your computer to connect to Aether network. This network is made of different boards (forums) where people post and discuss things. On the surface, it's fairly similar to Slashdot, Metafilter, Reddit, or any other community site on the Internet.
The different thing about Aether is that it doesn't have a server somewhere. The only thing the app does is that it finds and connects to other people using Aether. In other words, it's a distributed, peer-to-peer network.
This makes it impossible to censor, and renders its users anonymous. It's useful for people concerned about privacy, or pretty much anyone who doesn't want to be watched and catalogued for every word they write on the Internet (so, pretty much everybody).
It's also temporary. Whatever you post disappears after six months. It's designed to be an ephemeral space, and it's focused on now, rather than the past. Other people can still keep copies of what you wrote, but it won't last forever in the network itself. They also won't know who you are.
Community moderated, distributed and anonymous. Almost to good to be true, but... how do you know it is actually _gewg that's posting?
Original Submission
(Score: 0, Funny) by Anonymous Coward on Sunday July 05 2015, @08:33PM
i am _gewg
_gewg
(Score: 2) by KilroySmith on Sunday July 05 2015, @08:40PM
It'll be interesting to see how they solve the storage problem.
Let's see - if a billion subscribers all send a 1KB message each day (1 TB), six months worth of messages is about 182 TB. Based on my 30 second read of the website, every message is passed to every user on the Aether network, so each user needs to store 200 TB of data. Hmm, that won't work today, especially if those 1KB messages become 1GB movie rips.
Perhaps each user sets a data limit - "I'm willing to to store 100 GB of data for Aether". Messages would spread through the network and get destroyed on individual machines based on the owner's data limit. That could work - but message delivery reliability would drop drastically as time passed.
(Score: 4, Insightful) by M. Baranczak on Sunday July 05 2015, @09:47PM
Let's see - if a billion subscribers all send a 1KB message each day (1 TB), six months worth of messages is about 182 TB.
A fella wiser than myself said: "Scalability is easy. The hard part is convincing a million people to use your shitty site." Twitter has about 250 million users, you think this thing will get 4 times more users than Twitter?
(Score: 0) by Anonymous Coward on Monday July 06 2015, @02:32AM
Clearly he picked a billion because of the easy math matching it to a TB. Of course if you want to screw up everyone else you can start posting large files (or a bunch of smaller files) and let their hard drives fill up with crap. It will be easy to poison this well.
The reality is this type of system will turn into another way of sharing files with plausible deniability: I didn't download that. I don't have any control over what anyone else posts. I can only delete it from my computer once it arrives and I see what it is. That's why it's a deleted file on my hard drive.
(Score: 3, Interesting) by joshuac on Sunday July 05 2015, @09:52PM
Your assumptions have Aether getting 27+ times the number of installs as the number of people who have registered on Reddit over a decade and then daily commentary from every one of those users.
A massive success would be more like 1/27th the Reddit userbase and more reasonable posting rate may be a little closer to only the 90th percentile posting 7 times per week, with a rapidly dropping off interaction rate for the remainder.
Put another way, if Reddit (very roughly) gets about 550 thousand comments per day and Aether is a humongous success at 1/20th the size of Reddit, and also only needs to store about 180 days of comments, that's about 5 million comments, or a whopping 5MB assuming 1KB per plain-text posted comment.
Disk storage isn't going to be a problem.
(Score: 2, Disagree) by dusty monkey on Sunday July 05 2015, @10:04PM
Put another way, if Reddit (very roughly) gets about 550 thousand comments per day and Aether is a humongous success at 1/20th the size of Reddit, and also only needs to store about 180 days of comments, that's about 5 million comments, or a whopping 5MB assuming 1KB per plain-text posted comment.
With math like that, there is no problem that can't be solved. Problems seem to get exponentially easier the more you figure on them.
- when you vote for the lesser of two evils, you are still voting for evil - stop supporting evil -
(Score: 1) by joshuac on Monday July 06 2015, @05:10AM
How did you you already know about my mathematical solution for world hunger already? ;)
Sorry about that math, obviously I hadn't done even a quick proofreading before posting and going out the door.
I do still standby the obvious; Aether won't have a billion users nor every user commenting every day.
(Score: 2, Insightful) by Anonymous Coward on Sunday July 05 2015, @10:43PM
Don't forget how readily text is compressed. 75% compression is common with generic methods. Tweak it to be text-centric and 90% is quite possible.
(Score: 2) by kaszz on Monday July 06 2015, @01:18AM
7zip seems to be the compression to use for text (as wikipedia uses it).
(Score: 3, Informative) by jcross on Monday July 06 2015, @12:06AM
Additionally, you can easily shard the site into independent "sub-aethers", and users would only be downloading/storing/uploading content from the ones they're actually interested in. Systems that can be subdivided like that tend to scale very easily.
(Score: 2) by frojack on Monday July 06 2015, @03:23AM
Three statements cause me problems:
1) The different thing about Aether is that it doesn't have a server somewhere.
2) It's also temporary. Whatever you post disappears after six months.
3) Disk storage isn't going to be a problem.
I can't get to #2, or #3.
because I can't get past #1.
No, you are mistaken. I've always had this sig.
(Score: 0) by Anonymous Coward on Tuesday July 07 2015, @06:19PM
Why can't you get past #1? Has the concept of peer-to-peer networks passed you by?
#2 will obviously be controlled by the client, people can make copies of the content and keep it longer if they want to, but you won't be able to sync anything older than 6 months form other clients.
#3 Where did you get this from?
(Score: 0) by Anonymous Coward on Tuesday July 07 2015, @06:23PM
Excuse my comment for #3, I read back again and see where you got it from.
(Score: 3, Insightful) by stormwyrm on Sunday July 05 2015, @11:33PM
A billion users posting a message a day? I don't believe that there has never existed any online forum that is so large and so active. And even if you had a billion users, chances are that less than 1% of those will be so active as to post even one message a day. While I do read SN every day, I probably post closer to once every three or four days.
So just now a thought just occurred to me: in what way is Aether better than and different from Usenet? From Wikipedia [wikipedia.org]: "One notable difference between a BBS or web forum and Usenet is the absence of a central server and dedicated administrator. Usenet is distributed among a large, constantly changing conglomeration of servers that store and forward messages to one another in so-called news feeds. Individual users may read messages from and post messages to a local server operated by a commercial usenet provider, their Internet service provider, university, employer, or their own server."
Numquam ponenda est pluralitas sine necessitate.
(Score: 2) by VLM on Monday July 06 2015, @12:19PM
in what way is Aether better than and different from
Or fidonet or frost on freenet or zillions of similar ideas over the decades...
This helps with estimates. My guess is it'll be less popular that fidonet was at its highest peak, yet lower barrier to entry than frost on freenet, so hundreds of thousands of active users is realistic. Then the math is less than a million posts a day, less than a thousand days a year, and a K per post would seem to imply less than a terabyte per year.
Of course looking at or reading a post from a year ago seems unlikely-ish unless a really unusual user culture develops, so you really only need a couple gigs to hold a couple days. In a very usenet-ish manner you can set your expire to match your disk space, more or less.
Also if its anything like usenet, people only literate in Chinese will have very little conversation with people only literate in English, so some sharding strategy based on language would help traffic and storage quite a bit.
I haven't bothered digging into the dead github to see how it works, but its possible the network design will melt down before storage issues crop up, anyway. Hope they aren't doing something dumb like a mesh, or something that somehow scales O(x**4) or O(something ** x)
(Score: 2) by NCommander on Monday July 06 2015, @01:40PM
USENET isn't completely decentralized. The technical term for it is store-and-forward. News has to be stored on a central server, and then fed to other servers, it's not like P2P where shit goes from user to user directly. If you want a local USENET feed, you need to either suck it down from an upstream server, or get a peering agreement from another. The best way to think of it is as a massive mirroring system.
Still always moving
(Score: 1, Insightful) by Anonymous Coward on Monday July 06 2015, @03:01PM
What is the difference between a locally installed server that can peer with other servers (including other locally installed servers), and a locally installed client that does the same? Apart from the word used to describe it, I'd say nothing.
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @08:49PM
I don't download and keep a lot of stuff locally.
I am, however, a real packrat when it comes to bookmarks.
Stuff that disappears and makes my bookmarks useless does bug me.
When Google changed (and changed and changed and changed) its USENET archive, that was irritating.
After their most recent(?) "improvement", I just gave up on them.
it's focused on now
Hipsters.
-- gewg_
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @09:04PM
Bad taste in pedandery, "whatever" would have fitted the Subject.
(Score: 5, Touché) by wonkey_monkey on Sunday July 05 2015, @09:15PM
Bad taste in pedandery
I believe you meant "pedantry."
You're welcome.
systemd is Roko's Basilisk
(Score: 1, Interesting) by Anonymous Coward on Sunday July 05 2015, @09:12PM
There are bookmarklets and services to automatically backup pages of interest to one of the many online services that do it and that way you can bookmark those and always have access.
(Score: 2) by c0lo on Sunday July 05 2015, @09:21PM
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @10:04PM
Yeah, that will work for some purposes.
...though the cloud thing isn't necessarily perfect either.
You may have noticed that I am fond of the Google Cache of pages.
Those allow me to highlight portions of the page which I think are especially cogent.
Posting a link to that marked-up Cache (along with a link to the original) is technique I like to use.
Paradoxically, over time, that gives another point of failure if Google loses track of the page or changes their 12-character identifier for that page.
(Sites switching over to https-only is a recent wrinkle in this paradigm.)
OTOH, having a record of significant words in the text has allowed me to find the page again or another copy of the text when the original can't be found.
Pages that e.g. are behind a database search page[1] or are forbidden by a robots.txt won't get cached, so those may be other gotchas.
[1] Often, copies of pages are still at Archive.org, but a Google Cache of those doesn't exist.
-- gewg_
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @10:22PM
People archive 4chan; they can archive this newfangled thing.
(Score: 3, Insightful) by kaszz on Monday July 06 2015, @01:28AM
USENET posts disappeared after 2 weeks or whatever the local administrator thought was enough of retention time.
Anyway, there's a lot of hipster designed and administered things out there both in hardware and software. Any sane person does best to avoid them at all times if possible.
Seems we are now very near the point of re-inventing USENET in some form. This time eyes behind your shoulder will be surveying every post.
(Score: 2) by NCommander on Monday July 06 2015, @08:43AM
This was something I was seriously investigating for SN. I ran an experiment if it would be possible to spool the database into INN (netnews server), and proved that the basic idea was sound, but there were enough edge cases to prove that using INN wasn't practical for two way communication. What I want to see for SN is it to be able to continue if we ever go down, and my intent now is to implement NNTP in perl, and allow downstream servers to pull from use via NNTP XFER. As such, it could even be spooled into USENET itself, with "from USENET" posts coming in as AC.
Still always moving
(Score: 2) by TheRaven on Monday July 06 2015, @09:02AM
sudo mod me up
(Score: 2) by NCommander on Monday July 06 2015, @09:12AM
I got somewhat farther than you I guess; my plan was to set the soylentnews.* hiearchery to have the Moderated flag on it, and use a mail-gateway to manage posts. THe problem is there was no way to hook into INN's authetication code to make it validate against the SN database, and then automatically generate a mail header that validated a user to a post. The two systems were so far disconnected to each other that I couldn't figure out a decent way of doing it without a massive hack. Plus INN sucks to run.
I've looked at implementing NNTP, the problem are the article control numbers; NNTP *really* wants to its own system of managing it, and that means mapping tables to keep everything sane. Maybe I'll hack up NNTP code for the 14.07/08 rehash release though. The pain part will be handling STARTTLS in perl. I'm already cringing thinking about it.
Still always moving
(Score: 2) by NCommander on Monday July 06 2015, @11:30AM
Actually, after thinking about this again, its going to have to use INN if we want to spool onto USENET and get posts, or else I need to re-implement a mail-to-post gateway from scratch. I'm reading some of my old code from when I last experimented with this to see what I can do ...
Still always moving
(Score: 0) by Anonymous Coward on Monday July 06 2015, @03:21PM
Maybe a better solution than an integrated NNTP gateway would be an HTTP API that allows access to all the features also available with the user interface (but nothing else). This would then allow to write interfaces to other systems like NNTP without messing with Rehash internals, possibly even by third parties who wouldn't want to dive into Rehash itself. This would also have the advantage that once that API is available, any such development and deployment can go on independent of Rehash development/deployment, so e.g. an update of the NNTP gateway would not affect the main site.
Anyway, if you post to Usenet, make sure that the posts follow Usenet conventions (especially: no HTML; HTML is not appreciated on Usenet).
(Score: 2) by mr_mischief on Monday July 06 2015, @04:49PM
USENET and FIDO were the first two things I thought of when I saw the summary.
(Score: 2) by GoonDu on Monday July 06 2015, @03:24AM
4chan is similar to that concept in terms of ephemerality, it took the userbase roughly 7 years (when storage becomes way much cheaper) to create an archive site that automatically archive all the posts from respective boards. I can see the same thing happening to this as well, a server simply needs to connect to the Aether network and download scrape the posts.
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @10:04PM
Has Soylent news ever received an order under Section 215 of the USA Patriot Act?
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @11:09PM
I hear nothing but Crickets.
(Score: 2) by takyon on Sunday July 05 2015, @11:50PM
I'll get a response for you. It might run in a meta story.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 4, Informative) by NCommander on Monday July 06 2015, @12:06AM
No, we have not. We need to set an actual canary up, just been busy.
Still always moving
(Score: 0) by Anonymous Coward on Monday July 06 2015, @04:09AM
Ha. I guess I was trolling a little bit, but thank you for the answer. Cheers ! :)
(Score: 3, Funny) by Appalbarry on Sunday July 05 2015, @10:31PM
This makes it impossible to censor, and renders its users anonymous.
Yeah, Right. Haven't we heard that one before?
Someone go ask Dread Pirate Roberts....
(Score: 2) by coolgopher on Monday July 06 2015, @02:55AM
Inconceivable!
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @10:47PM
It's also temporary. Whatever you post disappears after six months. It's designed to be an ephemeral space, and it's focused on now, rather than the past.
A lot of newspaper-sponsored forums already work like this, although they weren't intended to. Old stories and old posts mysteriously vanish, out of reach of a Google search.
Same with YouTube. Try posting comments on someone's uploaded video of copyrighted content (regardless of whether it appeared a deal was cut) and then check back in three years. Chances are, that upload will be gone, along with all the comments, and will be replaced by someone else's upload of the same tune or video clip.
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @10:56PM
IMDB comments disappear too, but it depends on the specific movie - posts to low volume titles last a long time.
(Score: 0) by Anonymous Coward on Sunday July 05 2015, @11:30PM
out of reach of a Google search
On those sites, are new comments added to the -top- of the stack such that old comments are pushed to another page?
If so, when you make a comment (or find a useful comment), go to another forum where posts are permitted (but which does NOT use the nofollow tag) and post to a low-traffic thread.
In your post, add a link to Page 2 (and Page 3...) of the comment thread you would like to be indexed by search engines.
Give it a while to be discovered and indexed (which may take a month or more) then try your search again (using the site: parameter).
-- gewg_
(Score: 0) by Anonymous Coward on Monday July 06 2015, @03:06AM
You know the ones, who think they're opinions are so lofty because they post under a named anonymous account instead of posting from a random anonymous account. They'd never go for it because if we're all anonymous, they'd have to actually make their points based upon their comments and arguments! Can't have that!
(Score: 0) by Anonymous Coward on Monday July 06 2015, @03:45AM
Never synced. Must be SN'ed.
So: sudo apt-get remove aether
(Score: 2) by q.kontinuum on Monday July 06 2015, @04:41AM
Sorry, needed to paraphrase due to subject length limitation.
How about users generate a key-pair for each of their assumed identities? If you want something attributed to a specific identity, sign it with its key. If you don't, don't.
Registered IRC nick on chat.soylentnews.org: qkontinuum
(Score: 4, Interesting) by NCommander on Monday July 06 2015, @08:44AM
On my very long TODO list, I was debating implementing a 4chan like tripcode system for ACs. Avoid the hassles of an account, but still be verifiable.
Still always moving
(Score: 2) by q.kontinuum on Monday July 06 2015, @12:11PM
Sounds interesting :-) I would be curious how many people would actually use that, and especially how long they would stick to the same tripcode. I have a feeling that, while there are legitimate reasons to stay anonymous (or pseudonymous), a higher percentage of ACs just doesn't want to be judged by previous contributions, sometimes for obvious reasons.
Registered IRC nick on chat.soylentnews.org: qkontinuum
(Score: 5, Interesting) by VortexCortex on Monday July 06 2015, @05:48AM
how do you know it is actually _gewg that's posting?
The problem with anonymous forums is that they're easily infested with shills who leverage underhanded tactics to control the conversation, like consensus cracking, false flagging, demotivation, distraction, opinion "nudging" such as arguing two extremes while a 3rd shill presents a "moderate" yet propagandized response, etc. [digitalnewsasia.com]
Having seen the underbelly of the beast where a trolling industry has sprang up that allows everyone from governments to political parties, from to corporations to "non profit" organizations, to pay for shills to shape online discourse I recognize that the value of a purely anonymous system is very low, especially when it doesn't actually provide anonymity -- network analysis will tell you who posted what. Since Tor still has a hard time providing anonymity [thestack.com] then I don't hold out much hope that TFA's offering isn't worse. It's useful to be able to pick a pseudonym, perhaps just temporarily. Fortunately, so long as you're not ridiculously limited in message length then a client-side solution can be provided by simply integrating a PGP signature with each post. Unfortunately, there's a middle ground where one would like a verifiable pseudonym (see: tripcodes [name / fingerprint] on message boards like 2chan, et. al.), but without trusted signatories (or a centralized site) it's hard to provide such features (trust graph doesn't cut it for per-conversation verifiable poster ID -- centralized sites can just provide a topicID HMAC'd with poster IP address).
Due to a lack of moderation a purely anonymous posting forum does not scale well. As soon as it gets popular it's turned to shit. This is because the guarantee of anonymity creates a headache for self-moderation, i.e., spammers, shills and griefers can easily flood a conversation without adequate filtering being possible. On most centralized online forums a moderator team essentially provides a default filter you can not opt out of. With a decentralized approach the end users must be given adequate tools to apply their own filter -- Sharing (and merging) filter rules thus makes moderation opt-in: Users can crowdsource and select a "view" that is tailored to fit their desired discretion; Don't like the "moderation" then select a different moderator filter, or create your own filter (or drop them all to see completely uncensored mayhem with a side of spam). Being able to even temporarily filter all but signed posts is thus very useful -- so useful that any distributed anonymous message platform really should have a filter such as this. Come on now, let's learn from IRC's +v (voice) +m (mute) and etc. options rather than rediscover this particular wheel. Without moderators each filter is simply applied client side. Instead of moderators (or their bots) giving people they recognize or like a "voice", the client can maintain such a list and apply it to the unfiltered hive of messages on their own machine -- but that only works if pseudonyms or message signatures are also allowed instead of pure anonymity.
Nothing I'm saying here is new or unknown. Any modern anonymous messaging system that doesn't have such facilities is a decade sized step in the wrong direction. See also: TOX. [tox.im] Which is a great decentralized messaging platform that is also unfortunately missing many features. You'd think in today's day and age we could have at least just update usenet + PGP with some ephemeral IRC features and a better decentralization strategy... but we apparently can't since everyone wants to make us re-roll the whole party's character-sheets. Fortunately, some of us realize that it's the damn centralized Web's fault we don't have privacy online and are doing something about it by baking in decentralized data distribution (NDN) [named-data.net] With NDN baked in, a "reddit" implementation (along with damn near everything else) would have TFA's "anonymity" built in, but also support many more features like optional persistence. All the past decentralized / distributed data systems are just hacky kludges; Named Data Networking is the solution we need, i.e., the Internet itself needs to help out with decentralization rather than inefficiently putting all of the burden on the endpoints.
(Score: 3, Insightful) by darkfeline on Monday July 06 2015, @08:56AM
Isn't this basically Usenet? Some of the technical details might differ, but fundamentally it sounds like the same thing.
Join the SDF Public Access UNIX System today!
(Score: 3, Interesting) by FatPhil on Monday July 06 2015, @12:03PM
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 0) by Anonymous Coward on Monday July 06 2015, @02:49PM
Yes, that's exactly what I thought, too, when reading the summary.
Well, actually some people archived Usenet messages; one public archive actually became prominent and was later bought up by Google. But then, I wouldn't bet on nobody archiving messages on Aether either. Even if the developers don't intend it, the very nature of a distributed system makes it dead easy to do.
(Score: 2) by mr_mischief on Monday July 06 2015, @04:51PM
It's similar to Usenet or FIDO. It's not really a new idea. It's just a new implementation.