Editing 4chan/Successor

: ''whatever happens moot had one major point he wanted to make with his farewell: 
There needs to be more anonymous imageboards equal in stature and vibrance to 4chan.''

== Basic Components of the Anonymous Imageboard ==

Let us revisit the crucial components of a 4chan style anonymous imageboard:

# Openness - It must be easy to access and to post. As refugees from Something Awful's paid account systems, Shii defined the value of an anonymous imageboard as easy and inviting to post without having to create an account.
# Anonymity - Moot and Shii valued the ability for people to live and speak without reputations, something that is restored by the existence of the anonymous imageboard. This became popular with sites like 2ch since they work along the Japanese tradition of Honne and Tatemae, where society itself forces people to mask their true beliefs https://en.wikipedia.org/wiki/Honne_and_tatemae
and thus made 2ch valuable to the entire nation of Japan as a place where people could release their inner feelings
# Critical Mass and Community Vibrance - This component is harder for an imageboard clone to reproduce. The community must be vibrant and speedy, such that you can expect a reply within a few minutes or hours of posting. Without a strong community, the board is as good as dead, as we see with many textboards or imageboards.

This triangle of features that 4chan offers has not been replicated easily in the English-language imageboard community.

7chan, 8chan, and 420chan are the few exceptions, but they may not easily survive an onslaught of 4chan refugees.

== Types of Boards ==

* '''Image-dense Boards''' (pr0n boards: /e/, /d/, /b/) - The NSFW p0rn boards are difficult to host, but also present a tremendous funding opportunity when it comes to ads. 4chan is close to the strength of PornHub in the smut market. This would be a boon for Bloo, who, as a /b/ and pr0n board archiver, already has the experience and the will to manage such a system.
* '''Pure Discussion Boards''' (/a/, /v/, /g/) - The SFW blue boards have strong cohesive communities that focus strongly on text discussion rather than sharing images. Images are still critical to expression and data sharing, but in the same way as a modern group chat. They are also more conducive to donations to keep their home and community alive.
* For low budget archivers, It may be necessary to initially stick to textboards and leverage image embeds from Imgur or whereever to reduce the space constraints
** We might be able to leverage IPFS in particular to host content in a distributed, low cost manner. Research this possibility.
** On the other hand, archivers with better budgets (and which already have the storage capacity) could host images and still keep archiving them as if they had come from 4chan.

== Successors must be Innovative ==

moot said in his farewell speech
if he created 4chan today he would have it on mobile 

i'm running this organization from a smartphone right now
mobile group chat is the future
before I had irccloud i was not accessible on irc
since i didn't have push notifications
better mobile integration is not just for normies, it tremendously increases ease of use, accessibility, and flexibility

moot raised yik yak as one of his ideals of this
it's not an anonymous imageboard clone since it uses location heavily
however it is a completely new expereince that brings in a whole new type of community and type of conversation
innovations in communication like these are not just interesting ,they are vital to acceptance 

since for example nobody can really move off reddit because nothing really is better
so: what innovations in anonymous boards can you think of that can fit with the same three principles but bring a new and probably better experience
internationalization could be key, since 2ch users yearn for something that is not only like 2ch but better than it
they settled for twitter basically
and are also trying reddit

== 4chan Archivers as Strong Candidates for Succession ==

I can foresee that a possible contigency plan where the 4chan archivers, with no more 4chan to archive, become the best placed successors to 4chan

This can be done simply by shifting the ghostposting function to allow full threads, rather than just single posts. The archived threads from 4chan would keep continuity with the original thread on 4chan, bringing in a natural community in an instant.

this strongly hinges on the ability of FoolFuuka (https://github.com/FoolCode/FoolFuuka) or whatever board system they use to be able to serve an imageboard community as it's new primary purpose as well as our ability to migrate users over to the platforms (as I highly doubt 4channers would turn to the archives looking for a place to call home (unless we have archives make announcements about it)).

FoolFuuka might not be sufficient for this need as things grow, so either before the crash or after it, we need to have a migration plan where we turn Fuuka SQL format into Lynxchan or Asagi or whatever imageboard we so choose, keeping the original threads displayed.

We also have to think about how these original threads are marked in comparison to new threads. 

== New Technologies ==

As the costs of hosting an anonymous imageboard are so high, we need to leverage new technologies to make it work in the modern day.

If no one imageboard has enough bandwidth/space/power to host everybody, the communities will be split across different sites. Sites can use federation to share the posts across different servers then (3) is improved. (rather than having lots of boards with a very small userbase you can have lots that all work together to produce one large community).

=== Mobile Interface ===

Moot stated that if he were to recreate 4chan today, a mobile focus would be crucial for such a discussion system. 

There have been many a 4chan app made in the market, and open source. We can modify them to provide a mobile view into our site, and perhaps enhance them to 

An alternative option is to implement support for the already (somewhat) documented API so that the only change app developers would have to make is to the endpoints they connect to get data from.

=== Retooling FoolFuuka ===

Immediately retooling FoolFuuka to transform archived boards into postable boards is actually very possible and rather simple. Some mods are needed to make it happen though.

<antonizoon> How would you conduct this shift? In modding foolfuuka
<antonizoon> Should threads be made in the same place as the 4chan original boards to enforce continuity
<antonizoon> And should there be a way to mark which threads originally came from 4chan
<VoH> I'd suggest adding a flag to posts that came from 4chan, there's already a post user field
<VoH> like a collumn that normally handles things like noting an admin/mod post, but at least temporarily add one for posts that were originally from 4chan
<antonizoon> Great. How about retooling the archive to make ghostposting into normal thread posting
<VoH> Even simpler, in the board controls you have a setting that sets the board as an archive or a board
<VoH> the slightly harder part is going through the DB and un-ghosting the threads
<antonizoon> Unghosting?
<VoH> uhh, due to how foolfuuka handles a thread past its age cutoff
<VoH> if its ghosted an additional collumn is populated with a timestamp of the last ghost post i think
<VoH> lemme pull it ip
<VoH> yeah there's a "time_ghost" and "time_ghost_bump" in the _threads table
<antonizoon> Ah yes because ghosted posts don't bump?
<VoH> yup
<VoH> unless you're on the ghost board
<VoH> I've got some duplicate boards on testarchive I can fiddle with
<npleb> moving from asagi archive to ffuuka board is as simple as removing triggers/procedures and setting it as board in ffuuka

== IPFS ==

https://ipfs.io/ipfs/QmNhFJjGcMPqpuYfxL62VVB9528NXqDNMFXiqN5bgFYiZ1/its-time-for-the-permanent-web.html

We may want to try putting together a test IPFS imageboard, just to get a feel for how it would work. It is possible that IPFS can be used as the file backend (not for storing the thread content: but maybe we can serve backups from there)

This way, the costs of bandwidth or even DDOS are mitigated for the images, which have broken the backs of many an archiver.

note that this does not reduce the cost of hosting and storage space: rather, IPFS is more like a cloudflare replacement, dramatically reducing bandwidth costs. The more gateways you add, and the more fellow archivers that join the network, the lower the bandwidth gets.

One way to mitigate this is to provide load balancing IPFS nodes that people can volunteer to provide. These nodes serve a similar role as a standard load balancing server seen in large websites, but instead of contacting our own datacenter, they search for data through the IPFS network. It also provides a unique opportunity for people to donate servers rather than money.

=== Examples ===

Neocities is a website host that puts text and data into IPFS. It has a rather high amount of traffic on IPFS, and is the proof of concept model for a large imageboard based on IPFS.

glop.me is a pomf clone file host thatkeeps all files in IPFS. We should interview the admin about how he set it up and how it reduces his costs. it's actually quite heavily used.

https://github.com/2yrs/Pomf

There is even an IPFS pastebin: https://github.com/victorbjelkholm/ipfsbin
-should point out glop also has a paste funciton http://glop.me/paste.html

Most of all, to solve the client side problem of having to install IPFS, it could be entirely implemented client side in Javascript for people to access the IPFS network directly.

https://github.com/ipfs/js-ipfs

That is the ideological basis behind IPFSChan: which stores the text on IPFS too. Not sure if this is sustainable: but what is sustainable is putting JSON threads archived into IPFS.

https://github.com/disgustingwall/IPFSchan

=== How IPFS works ===

When a file is put onto IPFS, it is given a cryptographic hash (computed based off the contents in the file, sort of like a checksum).

This means that attaching a file onto IPFS that has the same contents of another file on the network results in the same hash. 

To search for a hash, it quickly looks for nodes that have the data that matches the hash using a Distributed Hash Table. 

When the data is received, it's contents are verified with the same hash used to search for it.

=== Inherent Dedup: Image Hash ===

IPFS has the ability to provide "inherent dedup" since each file is accessed by hash.

However, this dedup feature is lost if files are put into containers, since the container also stores the filename + extension. (comment: dedup is not lost, the container object will still point to the same object within)

Thus, what we will do is just serve the bare hash, and let the browser do the content detection.

This makes serving arbitrary content harder such as zips, but what we can do is work together with reit's glop.me for normal IPFS file hosting.

An example of a bare hash (right click and save it): https://gw.x1n.pl/ipfs/QmbMF1pFwvDWeskNmrBiuSzCDwJwciDtm57soJ16Qoyfwv

You can see that this file displays anyway in a browser, since the browser can detect image file formats regardless of extension and use the correct one when saving, as you see in imgur. The ipfs gateway is also detecting and sending the content type.

An example of the same file but with an arbitrary filename, which requires a different hash: https://gw.x1n.pl/ipfs/QmW5qhzdkKv2HpCBjCeJ8j96jSTzG9USni528YVoxXiVGm/ff500a72c9b419730344dfbb5a86682c.png

=== Handling "Deletion" in IPFS ===

Out of legal and ethical constraints, some sort of "deletion" must be presented, even though "deletion" from the network is technically not possible in IPFS. e.g. illegal hashes and such.

To implement this, banned hashes could simply not be served to users at all. With Glop.me, perhaps with nginx rules that don't allow certain urls to be accessed.

# /etc/nginx/conf.d/glop.me.conf
location ^~ /ipfs/QmAAAAAAAAABBBBBBBBBBCCCCCCCCCC { return 404; } # block the hash QmAAAAAAAAABBBBBBBBBBCCCCCCCCCC from being served to the user

This works purely because of the conjecture that we mentioned above (that most people won't have IPFS installed). If they do have it installed and grab data/media through the network directly, there is nothing we can do. But on the other hand, the liability is not on us because it's not coming out of our nodes.

The last part of handling deletion is the bitswap blocklists. In order to avoid serving or hosting banned hashes at all, we can subscribe to a central authority that can post lists of dmca's or blocked hashes. Unfortunately, given that this blacklist is literally a form of censorship, this central authority achieves a tremendous amount of leeway in enforcing the will of whoever's right to be forgotten. It remains to be seen how this situation can be solved.

The central authority: https://github.com/ipfs/refs

https://news.ycombinator.com/item?id=12787728
https://www.reddit.com/r/ipfs/comments/3m351b/discussion_permanent_content_dmca_and_illegal/

research into addressing this issue: https://github.com/ipfs/notes/issues/21

=== Upload Process ===

For the most part, on most services incoming bandwidth is free (like on AWS). This is because the vast majority of volume goes out one way: outbound.

Thus, it is no problem to accept large amounts of incoming files, and upload it to IPFS from there.

=== IPFS Nodes ===

Since IPFS requires a whole new approach to the Internet, it takes a gateway node for normal internet users to access IPFS files. This was the achilles heel to widespread adoption of onion networks: Tor and I2P simply required effort far beyond what normal users want to put in, and will forever remain niche networks.

One of the dilemmas with IPFS is that you still need to serve the images to the users. If they have IPFS installed they can pull them from the IPFS network. 

If they don't (and 99% of users do not) then we still have to serve the images to them via your own IPFS gateway. 

Which is easy, you just have an Nginx proxy pass to the IPFS daemon anywhere in the world. This server should have some cache space.

On the flip side: we can set up gateway nodes across the world that basically acts as a CDN, and this costs much less than building a giant NAS where everyone connects direct to it.

Handling the raw storage itself is still a concern, but here's one way we could approach it: https://trello.com/c/o1eBKKP3/146-ipfs-home-nas-long-term-storage

=== Thread Archival ===

While live threads might not be easily be hosted on IPFS since they require many concurrent changes and instant availability (we could keep it hosted on a database), Archived threads are another story. 

Archived threads not only do not change once the thread is closed: they should never change (since that would be tantamount to rewriting history). Also, IPFS reduces the strain on databases and SSDs since popular content is obtained from nearby peers in a distributed manner. Finally, the images linked to using IPFS are still on IPFS. 

Thus, using IPFS as a flat file archival system for archived JSON threads is a perfect use case. It's just a question of how to do it.

1. The thread is opened for discussion, and a JSONB entry is made in the database. Images are previewed on the server as thumbs, which are served conventionally.
2. Discussion happens in the thread until the thread closes. A lot of access and writes occur on the JSONB entry in the database.
3. The thread is closed as it has fallen under the post or bump limit. The JSON thread is then archived to IPFS and linked to by hash.
4. A user can view the thread with a client side renderer by providing an IPFS hash, which grabs the JSON thread and full size images (no thumbs needed in this system)

Notice that thumbs are not archived, and are used only on the server to reduce bandwidth costs when serving live threads. This is because once the thread JSON and full images are on IPFS, the bandwidth problems that require sharing thumbnails are mitigated, and there isn't as big of a need to host redundant image data.

=== TXT Records with IPFS: IPNS (IP Nameservice) ===

https://ipfs.io/ipns/gw.x1n.pl/

6:40 PM <•Sunako> Huh, neat
6:40 PM <•Sunako> You can put a hash into a TXT record
6:40 PM <•Sunako> And IPNS can resolve it like that
6:40 PM <•Sunako> With legacy DNS
6:41 PM <•Sunako> $ dig +short TXT gw.x1n.pl
6:41 PM <•Sunako> "QmbMF1pFwvDWeskNmrBiuSzCDwJwciDtm57soJ16Qoyfwv" ( check out https://github.com/ipfs/go-dnslink )

== NNTPChan: Modern Anonymous USENET ==

https://github.com/majestrate/nntpchan/

https://2hu-ch.org/thread-6a5153ae138db794d7a0ed33224812ab803480ea.html

A chan which is designed to be fully decentralized by using the NNTP service, akin to USENET but leveraging cryptographic signing. Could be a great way to have a freeform anonymous board.

The linked thread in question also has very good discussion on the use of IPFS, any downsides.

https://2hu-ch.org/img/FB2UGBGOTHQLWC74BC2I3V4E7DQPQMP2LSAVYNZ6GAVJENC6J4QHZEGJM4FCKRX4YD5SMTFQ2EIJIIGSPF5VMLHAYJKEG76J2WI74YA=.png

but what uguu (the author) said is that IPFS is too slow  for propagation
so if we used IPFS instead the problem would be that image posted on one board would not show up on another until a while later
you'd be seeing the post/mesasge but a blank image until propagation happens

However, when he says that I think he means if you used your own low bandwidth home server to display IPFS images. 

IPFS does not remove the need for good servers. The seed node needs to have good upload speeds

As you saw with sunako's pictures (linked above), images we know were never uploaded before, they showed up quite quickly right?

That's because his server acts as the main node and it is pretty fast already without ipfs so you get the speeds too.

== Frameworks ==

=== FoolFuuka ===

FoolFuuka is the engine commonly used by 4chan archivers. It has a ghostposting feature that makes it an able imageboard on it's own: can this feature be extended to instantly shift gears and ressurrect those communities?

* Pros - it's already developed, which means we can quickly make instances of it.
* Cons - No longer maintained, is a mess, could become untenable with huge amounts of real world users.

== 4chan Archiver Transition Scenario ==

=== Type 1: Continuity ith ===

== New Features ==

=== Faster, Compact, Instant update. Like Group chat ===

</pre>
8:27 PM <•antonizoon> how would a more instant chat affect how anonymous boards function
8:27 PM <•antonizoon> no more refreshing and thus less load
8:28 PM <•antonizoon> each thread is a self contained temporary room
8:28 PM <•antonizoon> but otherwise the same as 4chan
8:28 PM <dscb> <r3c0d3x> personally I'd find it more enjoyable so that I can see threads evolve in real time
8:28 PM <dabs> possibly more exuberant (or turbulent). could maybe tell sooner if its gonna be worthwhile for the user
8:28 PM <•antonizoon> how would that affect it
8:28 PM <•antonizoon> yeah
8:28 PM <dabs> atob.xyz is a great example actually
8:29 PM <•antonizoon> ah yes
8:29 PM <•antonizoon> that seems to be a good approach. it's something in between group chat and image board
8:29 PM <dabs> yeah
8:29 PM <dabs> https://en-m.tanobb.com/
8:29 PM <•antonizoon> with reduced lines for posts making it much cleaner on mobile too
8:29 PM <dabs> yup yup
8:30 PM <dabs> tanobb is an example where userbase is less integrated i think
</pre>

http://atob.xyz/p/93248

== Infrastructure Needed (Sunako's Analysis) ==

What making a replacement for 4chan?
a successor

yeah, if you have the money

I estimate 4chan's bill at probably 400-800USD/M For colo/power only

8000-12500~USD for bandwidth(0.008 to 0.0125/GB x 1PB per month) if they pay for actual bandwidth costs(which I doubt)
However they use quadranet who offer 1gbps unmetered connections for around 700USD/M
In theory you'd need four to reach 1PB monthly while maxing them out, lets say five though which ends up 3000-5000USD
There are cheaper options though, ask me(Sunako) for more info if you want

5000~USD for cloudflare enterprise

They own the servers or well, moot bought them a long time ago presumably sold to hiro

Most of their gear is at quadranet In LA

if you wanted to leverage the same facility even
I mean, I've said before about ipfs, it is a good idea
but it in actual practice is non-existent
For a large project I mean
I wouldn't depend on it
I'd just replicate 4chan's own efforts
And go with cloudflare

They aren't bad for the money really
moot spent a long time trying to find the right solutions
Up to his departure even
And in the end it just goes back to cold hard bandwidth

=== CloudFlare ===

Cloudflare is a requirement to survive DDoS attacks. Moot uses their enterprise plan, which can be expensive, but works.

https://news.ycombinator.com/item?id=6682625

=== Human Capital Costs ===

<pre>
8:53 PM <•Sunako> Honestly considering we don't know the inner workings of Hiro's 4chan empire, I can't estimate people costs
8:53 PM <•Sunako> But since he himself seems to be useless
8:53 PM <•Sunako> You can assume at the least one dev and one sysadmin
8:53 PM <•Sunako> They are probably only on a contractual as needed basis though
8:53 PM <•antonizoon> are moderators volunteers
8:53 PM <•Sunako> Yeah
8:53 PM <•antonizoon> janitors are
8:53 PM <•Sunako> Mods have always been unpaid
8:53 PM <•Sunako> As far as I'm aware
8:54 PM <•Sunako> I'm going to assume hiro maybe only has one full time-ish guy around
8:54 PM <•Sunako> who maybe does dev and sysadmin
8:55 PM <•Sunako> And hires freelancers as needed
8:55 PM <•Sunako> For other work
8:55 PM <•Sunako> Since that is the most efficient way to do this
8:55 PM <•antonizoon> isn't security a tremendous concern
8:55 PM <•Sunako> On his (supposedly) tight budget
8:55 PM <•antonizoon> i would add one of those as well
8:55 PM <•antonizoon> since it is a miracle they've survived this far
8:55 PM <•Sunako> What like exploits in the software?
8:55 PM <•Sunako> I wouldn't bother considering it
8:56 PM <•Sunako> Assuming the sysadmin does his job 4chan is significantly better off than sites in its league
8:56 PM <•Sunako> Since they only allow image uploads as far as I'm aware
8:56 PM <•Sunako> And I guess swf files
8:56 PM <•Sunako> They have a pretty small target area for exploitation in software
8:59 PM <•antonizoon> on that topic how large of a dev team would this need
9:00 PM <•antonizoon> i would say that something snapchat only hires 40 people
9:00 PM <•antonizoon> which is on a similar level
9:01 PM <•Sunako> 4chan's software has been in a mostly "done" state as far as I've heard from the moot days
9:01 PM <•Sunako> For quite a while
9:01 PM <•Sunako> As in only security fixes and very rarely new features are added
9:01 PM <•Sunako> So I'd say they only do freelancers as needed
9:01 PM <•antonizoon> given the amount of work needed to keep foolfuuka, kusaba, and the like relevant
9:01 PM <•Sunako> Maybe one resident dev?
9:01 PM <•Sunako> foolfuuka hasn't had any work in a while
9:02 PM <•antonizoon> maybe it might take a little more than that? it's not open source either unlike them too
9:02 PM <•Sunako> The last dozen or so commits are super minor things
9:02 PM <•Sunako> Mostly pulls too I think
9:02 PM <•Sunako> I mean you can look at things like vichan too
9:02 PM <•Sunako> or tinyboard for examples
9:02 PM <•antonizoon> how about
9:02 PM <•Sunako> They don't receive a lot of work
9:02 PM <•antonizoon> infinity next
9:03 PM <•Sunako> Isn't that dead?
9:03 PM <•antonizoon> they paid one guy or two to make the whole thing
9:03 PM <•antonizoon> is what i meant
9:03 PM <•Sunako> I haven't looked at it
9:03 PM <•Sunako> Oh yeah
9:03 PM <•antonizoon> and it was all a waste
9:03 PM <•Sunako> Totally doable
9:03 PM <•Sunako> hence why I said maybe 1 resident dev
9:03 PM <•Sunako> I would not be surprised at all if Hiro said they have no in-house devs though
9:03 PM <•Sunako> and all the work is freelanced out
9:04 PM <•antonizoon> ok, so 1-2 sysadmins, 1 developer (at least some kind of paid position  when needed if not full time as it is proprietary)
9:04 PM → dabs joined (~dabs@no.bully.pls)
9:04 PM <•Sunako> It isn't impossible one person occupies all the needed roles
9:04 PM <•Sunako> For a place like 4chan I can see it
9:05 PM <•antonizoon> so for example what's a full time wage you would accept monthky
9:05 PM <•Sunako> Me? Jesus you picked the wrong person to ask that
9:05 PM ⇐ dabs quit (~dabs@no.bully.pls)
9:05 PM <•Sunako> I have a CCAr from cisco
9:05 PM <•antonizoon> well you would be most similar to the sort of skillset needed 
9:06 PM <•Sunako> Instead of what I want
9:06 PM <•antonizoon> to keep something that gets attacked this much to stay alive
9:06 PM <•Sunako> I'll tell you what the lowest offer I've ever got
9:06 PM <•Sunako> It was around 300k~
9:06 PM <•antonizoon> yearly?
9:06 PM <•Sunako> Yeah
9:06 PM <•Sunako> monthly would be insane
9:06 PM <•Sunako> lmao
9:07 PM <•Sunako> Average dev/sysadmin/etc I think is in the 100k ballpark
9:07 PM <•Sunako> Depending on experience
9:07 PM <•Sunako> Someone I know who just started a while back at Microsoft
9:07 PM <•antonizoon> that's $25,000 a month for one senior sysadmin at minimum
9:07 PM <•Sunako> Fresh out of college got like 120k
9:07 PM <•Sunako> for java dev
9:07 PM <•Sunako> I think
9:08 PM <•Sunako> Actually, was it c++
9:08 PM <•Sunako> Maybe both
9:08 PM <•antonizoon> $10,000 a month for him
9:08 PM <•antonizoon> entry level full time dev
9:08 PM <•Sunako> 80-150k is the median I'd say
9:08 PM <•Sunako> For the average fresh out of college dev
9:08 PM <•antonizoon> if you have two sysadmins and three developers i would say that $80,000 sounds about right for capital costs
</pre>

=== Domains ===

I personally like gandi
If you aren't doing anything super illegal
They are France based and try really hard to not be assholes
I think that is even their company motto
"No bullshit"
And in practice I've seen it hold up
They are very neutral with domain issues
And generally only fulfill their legal obligation and nothing more
Which is usually just to facilitate communication between a complaint lodger and the domain holder

Now we just have to decide what to call this chan

=== Textboards ===

some discussion boards require only textboards. what do you foresee the bandwidth costs without the images
Basically nothing
With modern compression text costs next to nothing
You could run it on a 20 dollar vps
If it is just text

I'd go with a few hundred GB if you want to keep the text close to indefinitely
If you go with the standard 404 page 10 business almost nothing
Maybe a few dozen GB

As for how many could it serve
I mean as many as your server can realistically handle
More than I would think would even visit
On say a 20USD VPS maybe like 20k users depending on your config
Maybe 10k If it is crap php 
And you use hhvm for speed
write it in golang, it's better than PHP.

== Overchan ==

How this will work is that this organization will set up an Overchan like Amezou (but improved), which looks like a normal Fuuka homepage, but redirects users to chans that provide a 4chan board inspired community: as well as the archivers that provide ghostposting for continuity.

Some of these boards are Federated, which mean that they pool costs together. Others are external links to better fitting communities, such as the most vibrant communities of 8chan. Finally, other boards from good altchans can be recommended this way.

=== Continuing Federated Archivers ===

The 4chan Archivers are well placed to function as direct successors to 4chan boards, since they contain the archives of past threads and already have viewership. Certain archivers can be federated into a joint organization that can pool resources and costs to keep each other active. Not all archivers might want to be imageboards, but the ones that do should bring together their efforts.

    Desuarchive - All boards

      /a/, 

    Nyafuu - All Boards


=== 8chan Boards ===

8chan has many, but not all vibrant communities on it's site. Not everything can go there, because those communities haven't developed yet. But we can link to them if so. This can also offload much of our sites.

The continuity of 8chan communities can be shaky, because it relies on having good moderation. We have to choose the boards that are top in the ratings, and have had a long history of being there.

    8chan /b/

    8chan /v/

      8Chan's /v/ is chosen for it's close affinity and common heritage with 4chan's /v/. It's where most of those people would go anyway.

    8chan /pol/, /leftypol/

    8chan /tech/

    8chan /r9k/


=== NSFW Boards ===

Certain IRL NSFW boards are a special instance, since they would come under the control of Bloo for ad renevue. They require a lot of bandwidth anyway as they are image heavy, so offloading these would reduce the strain on the surviving archivers.

    /b/, /d/, /h/, /hc/, /hm/, /r/, /s/, /soc/


=== New Communities ===

The Overchan, as only a listing of boards, can also rank them and show new incoming places to go.

== Candidates ==

=== Likely Interested ===

* Bloo's 4chan archivers are a perfect place for all the NSFW boards.


=== Not Interested ===

These 4chan archivers will likely call it a day and simply provide the SQL backups.

* 4plebs - npleb just isn't interested in this one.