@kek ๐Ÿธ in #patchwork-next
Re: %a/l1he3uL

And once that is implemented (we call this out-of-order messages), you can do %I5E3g8a...

@kek ๐Ÿธ dug Not sure if someone has posted this here (I think I saw it linked recently)
@kek ๐Ÿธ dug Sometimes it's just a hidden hierarchy. Poorly defined rules is a _smell_ y
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ in #unix
Re: %utbx9WnA9

I'm using mksh for a while now. I didn't know about the implicit cd, thanks! What I like about it is that when using Alt+f/b and Ctrl+W, it respects - and / as word seperators. There's probably a way to do that in bash, but I hate configuring shells (that's why zsh is out for me)

@kek ๐Ÿธ in #decentralize
Re: %3WQ69u2Rd

but yeah, pubkeys as names makes so much more sense!

@kek ๐Ÿธ in #decentralize
Re: %3WQ69u2Rd

I think once they realize that it's just like email they'll shut up and realize that it's actually not so weird.

@kek ๐Ÿธ
Re: %fyAIBxWCQ

What would be super rad is to put the cryptpad stuff on top of IPFS' orbit

@kek ๐Ÿธ dug https://cryptpad.fr/ Consider it an alternative of google docs maybe is to
@kek ๐Ÿธ dug ## Anyone know how much each ethereum transaction costs the planet? I have in #ethereum
@kek ๐Ÿธ dug I've been using my trusty french press for the last couple of years. My oth in #coffee
@kek ๐Ÿธ dug hey, so tonight is the night we'll (maybe) be on TV NZ for "What Next?" wit in #root-systems
@kek ๐Ÿธ in #scuttlebutt
Re: %cNKDX8mOO

for the record: what @ezdiy describes is often called "context collapse"

@kek ๐Ÿธ pushed to %HGAxKZtUd...
@kek ๐Ÿธ in #scuttlebutt
Re: %ewQsrk9Bw

Yeah I'd also love to see use of multiformats. There are libraries for many languages and ssb really solves a lot of the problems they try to solve in an ad-hoc, somewhat hacky way. Invite codes are another example.

Also: IPFS uses CBOR but (of course it's IPFS) is AFAIK encoding agnostic.

@Aljoscha

content-adressable blobs, or simply no blobs and delegate to ipfs

What do you mean? That is already the case.

feeds should have a "type", just like messages (again, preferably hashes)

What kind of type? "RSS bot", "real person", "twitter bot"? Why not put that into an about message?

@kek ๐Ÿธ changed something in about
@kek ๐Ÿธ in #crypto
Re: %5ZIsdDwy1

There also is a Volume 2 which the author seems to find more valuable.

@kek ๐Ÿธ dug I liked the part with the dogs more than the part with the police. in #traveling
@kek ๐Ÿธ in #crypto
Re: %5ZIsdDwy1

My professor if a big fan of Oded Goldreich, we also use this in the lecture.

http://www.wisdom.weizmann.ac.il/~oded/foc-vol1.html

@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ

Examiner failing the turing test

turingtest.png

@kek ๐Ÿธ in #solar

Open MPPT by some freifunk (a mesh network) people

https://www.youtube.com/watch?v=NLP4MxQp8kk&feature=youtu.be

https://github.com/elektra42/freifunk-open-mppt

@kek ๐Ÿธ in #emojis
Re: %j+GvbKbPB

Where is no cars? Oh right, the consortium didn't think of that...

@kek ๐Ÿธ dug One of the main things that bothers me with Unicode emoji is the fact that, in #ssbc
@kek ๐Ÿธ dug ### view and publish [custom emoji](%RfQ2qb/wqQ/CfiwogMLFfNYJDwOFBe0+pZIP7y in #patchfoo
@kek ๐Ÿธ
Re: %lbfqJbj5+

I think BFT algorithms such as PBFT or HoneybadgerBFT should work, too. But they are also quite heavy

@kek ๐Ÿธ dug **Mujadara** ![mujadara.jpg](&Xyx38OXsKF1XNMRSjDnFpzctxtHTU214CuU5ltOBJYs in #cooking
@kek ๐Ÿธ in #solar
Re: %c2og3HFS9

paging @dominic @Piet @nanomonkey

@kek ๐Ÿธ in #solar
Re: %c2og3HFS9

Also: you don't ever need to care about cell balancing.

@kek ๐Ÿธ in #solar
Re: %Ulf7UYUjo

Considering that LiFePO4 is significantly (around factor 4.5) more expensive than a regular car battery of comparable capacity - what is the big advantage of LeFePO4? I mean yeah, it will not break as fast, I think it's a bit lighter and the energy density is higher, but it seems those don't really justify the price. Paying a thousand bucks for a 12V 200Ah battery seems kind of much when you could also pay two hundret.

@kek ๐Ÿธ dug I just want to feel like something I am putting out there matters. I don't in #shakes
@kek ๐Ÿธ
Re: %4CXLVEWQx

I have Android, but I'm not logged into a Google account. Will they still record my stuff? Can I see what they store? Because for the My Activity thing I need to be logged in.

@kek ๐Ÿธ in #scuttlebot
Re: %UE8H+Q6jK

Also on using blobs for Dropbox-like stuff: I'm not sure whether it's clever to direclty use a file as a blob. Maybe we could reuse some IPFS code (I guess that would be their unixfs) to implement sharding large files and directories into several blobs, though. Also, if people start using this stuff and moving large amount of data into the networks we will need garbage collection for blobs, and I'm not even sure that will be enough. Maybe the best way to go forward is to create blob servers that store blobs only from people on a whitelist. That would introduce a new single point of failure though, and I'm not sure if that is a good idea.
What would work is to tag blobs to only ask a certain set of peers for the blob, so you can only sync in a direct connection. Also maybe pubs could start facilitating those, either STUN style or by relaying the TCP streams.

Of course a password db will not be the problem. But I'm sure that if we have something for sharing files and directories, people will soon start putting music and movies in there, and those are big.

@kek ๐Ÿธ in #scuttlebot
Re: %UE8H+Q6jK

&[blobid].[hashAlg].[key].[encAlg]

I meant &[hash].[hashAlg].[key].[encAlg]

Though I think &[key].[encAlg].[hash].[hashAlg] would be better - you read right-to left, always know what to do and terminate when reading a &.

@kek ๐Ÿธ in #scuttlebot
Re: %UE8H+Q6jK

@cel that is right. They'd need to at least run some javascript to do the decryption. It's not a big deal if a local server sees the keys, but I don't want a pub seeing that.

Yeah but why would you send that request to a pub? The request goes to the UI web server.

@vendan one day I'd like to understand what javascript did to hurt you so much

1) slow
2) needs lots of ram
3) no proper types
3a) by extension, parameter checking in functions is terrible
4) a billon ways to do one thing, so everybody does shit different and in the you end up writing a bazillion polyfills

Also when you use JS for core functionality the site will be unusable with e.g. NoScript, which sucks.

not really about js but about getting ssb running:

the node vs electron version thingy is really annoying, because you need to jump through a thousand hoop to get a usable version of sodium. I was using ssb in a vm and I just couldn't use it when using the js fallback. Too friggin slow.


To me, JS looks like a nice language to prototype stuff. You can just cram a bunch of stuff together and see if it works. But to build something fast and robust it just isn't the right tool.


Oh yeah and on the id: I'd propose
&[blobid].[hashAlg].[key].[encAlg]

@kek ๐Ÿธ dug Welcome, Leo! I was curious on the aspects you found confusing. Was it th in #faq
@kek ๐Ÿธ in #new-people
Re: %QKtgTPdQN

The problem with web-based is that a website resides on a domain, giving you inherent centralization. Also the server is able to pass you any JavaScript, including "please give me your private key" and you wouldn't even notice! I guess ServiceWorkers could provide some remedy here.

I agree this is a lot better than the old web - one of my favourite projects that does crypto in the browser without leaking the key to the server is https://cryptpad.fr. This is great as a tool for transition, but IMHO the goal should be an actual program that you run locally.

@kek ๐Ÿธ dug # encrypted blobs forking from %GBLCZLIgQ8Tx2OdHZQw/+MZc6dgfCwPZ0DIJMiHAIm in #scuttlebot
@kek ๐Ÿธ in #offgrid
Re: %KpunZ6o4e

you're talking about 10mV (which is very little) and it's the least significant digit. It could be your multimeter, or just plain rounding.

Since balacing issues usually occur due to manufacturing differences wrt output impedance, I don't find it suprising at all that it's the same cell that has a different voltage.

@kek ๐Ÿธ dug Good news! It *"worked"*! I got a message from the computer to replicate to in #ssbc
@kek ๐Ÿธ followed @xj9
@kek ๐Ÿธ in #new-people
Re: %6sMcLKZLU

Also, in the future I can envision pubs being useful for regaining access to your log if you lose access to your local copy, so "archives" could apply in this case as well.

note: this has happened in the past.

@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ

I just stumbled over gowebview. It's supposed to make an android app from a web server written in go. I have no clue whether it actually works, but it might be possible to make an ssb android app from it, using go-ssb.

I hope I'll find some time to try this soonish, but I can't promise anything...

@kek ๐Ÿธ dug [@keks](@YXkE3TikkY4GFMX3lzXUllRkNTbj5E+604AkaO1xbz8=.ed25519) I think it w in #counterantidisintermediationism
@kek ๐Ÿธ dug great article. text: [meet-the-counterantidisintermediationists.txt](&YLEN in #counterantidisintermediationism
@kek ๐Ÿธ in #offgrid
Re: %UlrYlvkFx

Doesn't rewiring for balancing suck anyway? I would assume they get unbalanced within one charge, and rewiring on every charge is just not workable - at least I think so. But then, it seems a lot of people do it that way...

Here is a paper giving a nice overview of several techniques for balancing: A Review of Passive and Active Battery Balancing_ based on MATLAB-Simulink.pdf

The simplest design is to connect a large resistor in parallell to each cell. I guess it looks like it's not very efficient though. I tried calculating the losses but the last time I did EE is several years back and I'm not 100% sure I can treat a battery as a simple voltage source.

Any electrowizards around?

@kek ๐Ÿธ in #counterantidisintermediationism
Re: %N1/K6vyxj

@Soggypretzels

I was thinking about that as well as a service that would mirror an RSS feed onto the Scuttleverse for offline viewing.

That existed and was super spammy. Not sure maybe nowadays it wouldn't be so bad because the entire network has gotten very spammy (back in the day it was quiet enough to read every single post)

@kek ๐Ÿธ in #new-people
Re: %VCNmbENU6

@lojak well, if all pubs disappear you can only get updates by making direct connections - either through wifi or e.g. cjdns. Basically a pub is just a regular node that has it's ip address and pubkey public.

@kek ๐Ÿธ in #webgl
Re: %magXwZikf

It's just a jump to the left...

@kek ๐Ÿธ in #flume
Re: %xuVitZB9P

a certain action like "forget this after x days" is based on humans honouring promises rather than DRM style technological enforcement.

This!

@kek ๐Ÿธ in #flume
Re: %xuVitZB9P

The gossip protocol by itself currently is trustless because its neigh impossible to verify trust in reverse direction

Well, let's say subscribing is trustless, but publishing is not. You can't publish stuff to people you don't have a transitive "trust" (follow) relation with.

@kek ๐Ÿธ dug just for clarifiction, block was never about hiding your feed in the blocke in #flume
@kek ๐Ÿธ in #ethereum
Re: %bNc4mcv0i

I read REX and my head said

rex.jpg

@kek ๐Ÿธ dug Secure Scuttlebutt crosses The Atlantic ... in a published article! https in #counterantidisintermediationism
@kek ๐Ÿธ in #flume
Re: %hKY5uHiyP

also instead of doing

{
  "flag1": true,
  "flag2": false,
  "flag3": true
}

we could just put all the flags in an array: ["flag1", "flag3"]

@kek ๐Ÿธ in #flume
Re: %hKY5uHiyP

@ev

interested in a way to remove specific feeds from my database

that sounds like a local "kickban": remove it and don't accept it when someone offers it.

Maybe we shouldn't conflate too many different things in one message field. Let's be verbose here, we can still set several fields when clicking a button - in the end this is going to be client-specific behaviour.
I'd rather have a more or less clear definition of what database fields actually mean than let everyone interpret the db completely different.

@kek ๐Ÿธ in #flume
Re: %hKY5uHiyP

@ezdiy

To determine follow status now, you need to know only the most recent 'contact' message. For what describe, you'd need to scan the whole history of em.

Yes, but is that a problem? I thought one reason for switching to flume is much better reduce performance?

But yeah, let's agree on the semantics of follow/unfollow and block/unblock. Maybe "mute" is a better term for hiding from the UI and block should actually be about replication?

@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ in #flume
Re: %hKY5uHiyP

Reading the thread, I think I find what @ezdiy described as

weird tri-state thing (which is even more confusing)

would be the most intuitive for me:

  1. number of times I follow someone == number of times I unfollowed them => neutral
  2. number of times I follow someone > number of times I unfollowed them => follow
  3. number of times I follow someone < number of times I unfollowed them => block
@kek ๐Ÿธ in #flume
Re: %hKY5uHiyP

Will it completely stop replication? Maybe I'm just not interested in their stuff anymore, no need to completely stop replication.
Sometimes I follow people b/c I see replies to their posts but they are out of foaf range. When I received the post I sometimes unfollow them.

I thought there was a special "blocking" type? I would expect if I first follow and then unfollow, that everything is just like before I followed them - but I replicate stuff from people I never followed, right? So IMHO it should be just like that, and unfollowing more often than following should block.

I hope that makes sense.

@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ dug source: https://english.emmaclit.com/2017/05/20/you-shouldve-asked/ ![you- in #feminism
@kek ๐Ÿธ dug Basically, if I'm building some messaging system (including simple busses), in #elixir
@kek ๐Ÿธ in #elixir
Re: %r6yY2WK65

I've been meaning to take a close look at Erlang/Elixir and used Erlang for a little toy project I never finished. As a Go dev I found non-blocking IPC (Go channel communication usually blocks) a bit weird. Which one do you prefer for what cases? Can you recommend some resources?

@kek ๐Ÿธ in #scuttlebutt
Re: %UuDYPy7ae

I believe the undo thing could be used as a tie breaker when forks happen! If in doubt, use the undo message.

@kek ๐Ÿธ in #scuttlebutt
Re: %UuDYPy7ae

more options. not sure how efficiently they can be implemented using level's/bolt's iterators though.

truncate the past

post a message "i'm going to remove all my posts before %abc and I'd like all of you to do the same. Also it would be rad if you post a message that you also think that message is on my log". Not a prose message, but this could be the rough semantics.

undo/replace

post a message "undo my last message and replace it with this one: {..some nested mesage...}". The inner message has the same previous field as the one before.

@kek ๐Ÿธ

Go developer survey results

What is the biggest challenge you personally face using Go today?

67 (1.9%) java

@kek ๐Ÿธ dug I interviewed a great philosopher of technology a while back. I finally got in #technosophy
@kek ๐Ÿธ voted Maybe I will try this. But I think, it makes no real difference: Very few i
@kek ๐Ÿธ dug Maybe I will try this. But I think, it makes no real difference: Very few i
@kek ๐Ÿธ in #patchwork
Re: %hBben4J4Y

still 2GB is way too much

@kek ๐Ÿธ dug Hey hey hey hey now. Don't throw stuff on GCed languages just cause one is in #patchwork
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ dug this looks promising!!! https://ponzico.win/ in #hypercapitalism
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ changed something in about
@kek ๐Ÿธ followed @keks/vm
@kek ๐Ÿธ changed something in about
@kek ๐Ÿธ changed something in about
@kek ๐Ÿธ
Re: %N6lWEemCK

@Vendan

is wrong IMO, as that nick may never have been true for other clients

But you, @Dominic and me meet in person and I call him e.g. @knowing-nose, you would also hear that name.
The names I use for others are part of my expression. I think a tooltip on hove that tells you a bit more about the user without opening the profile would be nice though.

@kek ๐Ÿธ in #scuttlebot
Re: %shv9Si/sg

This is my primary use case for cjdns.

@kek ๐Ÿธ changed something in about
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ pushed to %gcqnwOUOX...
@kek ๐Ÿธ
{
  "type": "git-repo",
  "name": "tlog"
}
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ sent a private message
@kek ๐Ÿธ in #34c3
Re: %B8yYztqAO

โ€žAnd of couse, weโ€™ll make every effort to tame the convention hallโ€™s atmosphere and make it as cosy as our participants have come to expect over 33 years of congress history.โ€œ

I sure hope they succeed

@kek ๐Ÿธ changed something in about
@kek ๐Ÿธ in #patchwork
Re: %WU3gGJ0sU

Another use-case: I sometimes use them as bookmarks for myself. There is no UI for that yet (except manually querying on the command line), but at least the data is out.