BTW, Horde is a distributed supervisor, i.e. analogous to what regular Elixir OTP supervisors do, except that it can restart processes on different cluster nodes. It also provides a distributed process registry.
Daniel Azuma of Google gave a great 39 minute talk at ElixirConf 2018, "Docker and OTP: Friends or Foes?" [3], where he shows Horde (and DeltaCrdt) in action to keep a multi-player tank video game going while the Docker container it's running in gets killed and and the Elixir processes get respawned on another container. The state is persisted across Elixir nodes in other containers in an instance of DeltaCrdt.
(Again, the talk is really good -- if the above sounds interesting, you'll want to check it out [3]).
[0] https://github.com/derekkraan/delta_crdt_ex
[1] https://moosecode.nl/blog/how_deltacrdt_can_help_write_distr...
https://doc.akka.io/docs/akka/current/typed/distributed-data...
It's a bit like Riak or etcd but can be used across organizational and trust boundaries similar to e.g. a cryptocurrency. Anyone can run LF as part of the same network and both data privacy and security are maintained.
However, as he told me a more sophisticated persistence layer is needed. Looking forward how it's emerging
Cool stuff for sure.
Here’s an insightful podcast about Actual (hosted by the author of Tailwind.css), http://www.fullstackradio.com/126
GUN is running in production with 8M monthly users.
On large sites like HackerNoon and with Internet Archive.
What makes you say it is unstable?
If there is anything, please report it and we'll get it fixed.
Edit: Re grandparent: Will try to add "dark mode" to docs, in meanwhile, check out the docs on plain-text GitHub wiki ( https://github.com/amark/gun/wiki/Graph-Guide ) which is where everything pull from anyways.
These people aren't me. (Thanks to them for kind comments, tho!)
I have a couple throw away HN accounts but I do not use them to talk about gun.
I use a distributed testing tool to simulate any setup or edge cases we find, so we can fix it.
So this would be useful to replicate.
Also, a big fix was released early October.
Last release was several months ago.
Previous HN discussion on the mini-retrospective might be interesting: https://news.ycombinator.com/item?id=19886883
A good example is inserting indentation in response to, say, the "Enter" key. In a synchronous model, this is pretty straightforward. You calculate the indentation as essentially a function of the contents of the buffer and the delta (the insert of the newline), then any other typing gets processed after that. In a CRDT model, the additional typing can be interleaved in any order, and the merge operation has to converge to the correct results. You can think about how bad that is: the user could type a hundred lines of code, complete with complex editing operations including cut and paste, deletions, undo, and then the auto-indent plugin wakes up, calculates the diff, and sends it to the core. Meanwhile, the user has typed another hundred lines, including more cut'n'pasting of the affected lines. This is perhaps unrealistic, but the CRDT model requires it. And I think it is possible to solve it (I wrote some "rope science" articles, then there was more discussion in several issues), but it's a hugely intricate puzzle compared with just applying it synchronously, and the benefit is not worth it. It's quite practical to compute indentation for any realistic source file in under a millisecond, if you've got a nice fast language.