| Time |
S |
Nick |
Message |
| 01:42 |
|
pdurbin |
prologic: are you a bigger fan of IRC or gopher? |
| 01:42 |
|
prologic |
Haha |
| 01:42 |
|
pdurbin |
What? It's a serious question. :) |
| 01:44 |
|
prologic |
Lol |
| 01:44 |
|
prologic |
Hard to say. |
| 01:44 |
|
prologic |
I do quite like the simplicity of Gopher though |
| 01:45 |
|
prologic |
Hard to pollute content with lots of UI noise |
| 01:45 |
|
pdurbin |
gotcha |
| 01:45 |
|
prologic |
Better for me and information discovery and accessibility |
| 01:46 |
|
pdurbin |
I feel like my original goal was to learn just enough IRC to participate in freenode, which is where all the open source people hung out at the time. This is maybe 2006 or so. |
| 01:46 |
|
prologic |
and now? |
| 01:50 |
|
pdurbin |
IRC has grown on me. I don't know if you caught some of the discussion in #opensourcedesign yesterday, but I like how IRC rewards patience. |
| 01:53 |
|
prologic |
I can’t quite say IRC has grown on me since I’ve been using it since 1997 |
| 01:53 |
|
prologic |
If anything I’ve grown with it |
| 01:53 |
|
pdurbin |
What do you like better than IRC for what it does? |
| 01:54 |
|
prologic |
Nothing |
| 01:54 |
|
pdurbin |
heh |
| 01:54 |
|
* pdurbin |
wonders if bear likes XMPP better than IRC |
| 01:55 |
|
prologic |
It’s actually funny modern chat systems are just IRC with the fancy UI |
| 01:57 |
|
pdurbin |
sure |
| 01:57 |
|
prologic |
In recent years the only unique thing I’ve seen is Google Wave |
| 01:57 |
|
pdurbin |
plus some features like store-and-forward/scrollback |
| 01:58 |
|
prologic |
But that didn’t take on |
| 01:58 |
|
pdurbin |
I was sort of into Wave at the time but if no one cares about it, neither do I. I wasn't exactly inspired by it. |
| 01:59 |
|
prologic |
Maybe but storing things for later retrieval or push isn’t really new |
| 02:00 |
|
pdurbin |
Sure, there are IRC bots for all this stuff. |
| 02:00 |
|
pdurbin |
workarounds |
| 02:00 |
|
pdurbin |
kludges |
| 02:01 |
|
pdurbin |
For me, IRC plus public logs is pretty much all I need. |
| 02:01 |
|
prologic |
Bbs |
| 02:25 |
|
bear |
I like XMPP more for it's non-chat like features, if the main requirement was chat, i would stick with IRC |
| 02:26 |
|
bear |
my fav protocal is actually NNTP |
| 02:26 |
|
bear |
it's all the things that these fancy block chain folks are working towards so hard |
| 02:27 |
|
pdurbin |
heh |
| 02:28 |
|
pdurbin |
bear: you wouldn't pick one of those newer fancier open source chat solutions? |
| 02:29 |
|
bear |
I haven't looked under the covers at most of them, I may - but it would have to have some strong new features to tip me over |
| 02:30 |
|
pdurbin |
Ok. Would you ever run prologic's IRCd server? :) |
| 02:30 |
|
bear |
I was considering do just that |
| 02:31 |
|
bear |
having a modern take on an ircd using gopher - that has my interest |
| 02:31 |
|
bear |
something that focuses on the core parts of what chat is |
| 02:32 |
|
pdurbin |
wait, is his ircd built on gopher? wtf |
| 02:32 |
|
bear |
isn't it? |
| 02:32 |
|
bear |
oh poo |
| 02:32 |
|
bear |
you talked about gopher above and the Go icon is a gopher... |
| 02:32 |
|
pdurbin |
wouldn't surprise me ;) |
| 02:33 |
|
bear |
but yes, having a golang ircd interests me - especially if netsplits can be handled with maybe some of the dup/merge capabilities that NNTP had ;) |
| 02:35 |
|
pdurbin |
Yeah, I don't run any services based on Go but I'm willing to give one a whirl. philbot is written in Perl (mostly not by me) and it's rock solid. |
| 02:36 |
|
pdurbin |
I bet my kids would love plush gophers I see people bring back from Go conferences. |
| 02:37 |
|
bear |
a lot of our tools and some services are written in Go - we are finding it to be as stable as anything else you would trust production with |
| 02:38 |
|
pdurbin |
Cool. What languages did you favor before Go came out? |
| 02:38 |
|
bear |
Python |
| 02:39 |
|
prologic |
back |
| 02:40 |
|
prologic |
sorry I couldn't bear tapping out responses on my shitty iPhone/iOS |
| 02:40 |
|
prologic |
tiny ass keybaord |
| 02:40 |
|
prologic |
what I was going to say that things like playback aren't really a hack/work-around as such |
| 02:40 |
|
prologic |
ZNC for example is one such solution that works quite well |
| 02:41 |
|
prologic |
Some of the newer IRCv3 specs which a lot of IRC servers and Client software support are also making some of these "things" less clunky |
| 02:41 |
|
bear |
yea, IRCv3 is supposed to have cleaned up a lot of things - unicode support, line lengths and the like |
| 02:41 |
|
bear |
better service handling |
| 02:42 |
|
prologic |
[18:30:40] <pdurbin>Ok. Would you ever run prologic's IRCd server? :) |
| 02:42 |
|
prologic |
[18:30:55] <bear>I was considering do just that |
| 02:42 |
|
prologic |
^^^ haha awesome :) |
| 02:42 |
|
prologic |
lemme know if you do! |
| 02:42 |
|
prologic |
[18:32:32] <pdurbin>wait, is his ircd built on gopher? wtf |
| 02:42 |
|
prologic |
[18:32:37] <bear>isn't it? |
| 02:42 |
|
prologic |
[18:32:44] <bear>oh poo |
| 02:43 |
|
bear |
it's on my list of things to explore this holiday break |
| 02:43 |
|
prologic |
^^^ Well Go :) Whose mascot is a Gopher :D |
| 02:43 |
|
bear |
yea, that and the conversation earlier made my mix the two |
| 02:44 |
|
prologic |
[18:33:48] <bear>but yes, having a golang ircd interests me - especially if netsplits can be handled with maybe some of the dup/merge capabilities that NNTP had ;) <-- I haven't filed an issue about this yet but I'm actually going to try to use concensus/raft for server<->server linking at some point |
| 02:44 |
|
prologic |
with the hopes that resolving netsplits will be easier/ebtter |
| 02:44 |
|
bear |
using raft makes a lot of sense - netsplits would be self-healing |
| 02:44 |
|
bear |
s/would/could/ |
| 02:45 |
|
prologic |
And yeah Go is the better Python |
| 02:45 |
|
prologic |
quite frankly |
| 02:45 |
|
prologic |
not my words; the words of some of my colleages at FB |
| 02:46 |
|
prologic |
but I actually affirm the same belief personally since hackong on some non-trivially sized proejcts and written non-trivial amounts of Go since ~Nov/Dec 2015 to now |
| 02:46 |
|
bear |
I keep hearing that, I haven't yet personally done enough Go coding to make the call - but a lot of people whom I respect are saying good things about it |
| 02:46 |
|
prologic |
*nods* |
| 02:46 |
|
pdurbin |
I thought Go was a better Erlang. :) |
| 02:46 |
|
prologic |
as someone that has done over 15 years of Python myself |
| 02:46 |
|
prologic |
~2012 or so; so whatever that is |
| 02:46 |
|
prologic |
yeah its a pretty sure thing for me at least |
| 02:47 |
|
prologic |
pdurbin well Go is a better <X> :P |
| 02:47 |
|
prologic |
they just did a bunch of things right IHMO |
| 02:47 |
|
bear |
i've seen some rough edges for Go - but they also exist for Python, so it's a wash |
| 02:48 |
|
prologic |
yeap |
| 02:48 |
|
prologic |
circling back on irc servers and raft |
| 02:48 |
|
prologic |
I haven't given it too much though besides thinking about basically replicating state |
| 02:49 |
|
prologic |
and discovering other nodes (effectively servers) |
| 02:49 |
|
prologic |
I might even dispart from tranditional approaches of the servers having to explicitly connect to each other via any PASS/CONNECT server<->server protocol |
| 02:50 |
|
bear |
a lot of that is/was required because of the contentious nature of irc networks |
| 02:51 |
|
prologic |
I might actually spend this thanksgiving weekend (when not conversing with family/friends eating/drinking/etc) and code up prologic/cadmus (a port of circuits/irclogger) |
| 02:51 |
|
bear |
if you are working from a trusted client connection, you could loosen those restrictions |
| 02:51 |
|
prologic |
And... make it distirbuted/clustered with raft |
| 02:51 |
|
prologic |
so that there are multiple instances on a network on different servers |
| 02:51 |
|
prologic |
and the bots discover each other via a predefined channel and join a cluster |
| 02:52 |
|
* bear |
nods |
| 02:52 |
|
prologic |
the idea being that the logs are replicated and agreed upon from different parts of the network |
| 02:52 |
|
bear |
services just become another flow of client-to-client traffic |
| 02:52 |
|
prologic |
well |
| 02:52 |
|
bear |
sorry - I went down a small branch from what your talking about :) |
| 02:52 |
|
prologic |
on prologic/eris which runs on irc.mills.io I'm actually taking that approach to "services" |
| 02:53 |
|
prologic |
not actually writing services in a tranditional server<->server appracoh with "fake nicknames" |
| 02:53 |
|
prologic |
but just simple bots that do simple things |
| 02:53 |
|
prologic |
so far just one; prologic/soter that just does two things persists channel topics and channel modes |
| 02:53 |
|
bear |
a micro services approach |
| 02:53 |
|
prologic |
the caveat is that the bots actually require IRC Operator privileges |
| 02:54 |
|
prologic |
yeah pretty much |
| 02:54 |
|
bear |
then logging becomes a request to the services for a range of log entries and each server replies with what it can provide |
| 02:54 |
|
prologic |
haha |
| 02:54 |
|
prologic |
no all good :) |
| 02:54 |
|
prologic |
I like discussing these things |
| 02:54 |
|
prologic |
and I really want to hack on this some more |
| 02:55 |
|
prologic |
prologic/eris's main focus btw is some different/modern design decisions of course |
| 02:55 |
|
prologic |
but also a focus on privacy/security |
| 02:56 |
|
prologic |
in practical terms (as of v1.5+) this means if you connect via TLS you get +zZ which means you cannot talk to anyone (private message) connected on a non-TLS port to the server and vice versa |
| 02:56 |
|
prologic |
I will extend this to channels as well soon so channels can be set with a +Z as well requiring all participants to be connected via TLS (+z) |
| 02:58 |
|
bear |
will be interesting if you expand that later to include auth state |
| 02:59 |
|
prologic |
what do you mean? |
| 02:59 |
|
prologic |
Also do let me know if you do plan on and actually get an instance of eris running for your purposes |
| 02:59 |
|
prologic |
I'd love the development to be driven by "not just me" |
| 03:00 |
|
bear |
so I would connect with TLS (gain my +z) and then auth (gain +r) and so then I would only be able to talk to others |
| 03:00 |
|
bear |
or say in #general which could be open to anyone |
| 03:01 |
|
prologic |
Ahh yes |
| 03:01 |
|
prologic |
you mean some kind of +R mode; where only authed users can talk to each other |
| 03:01 |
|
prologic |
and +R for channels where only authed users can join/talk on channels |
| 03:01 |
|
bear |
right |
| 03:01 |
|
prologic |
+Z should probably extend to also block joins of channels too if you're not securely connected |
| 03:02 |
|
prologic |
got it |
| 03:02 |
|
prologic |
I actually like all this and will finish adding +Z for channels and +R for users/channels once I'm done with SASL (being worked on now) |
| 03:02 |
|
bear |
that mirrors what XMPP can do with MUCs and may also resembles voice permissions |
| 03:02 |
|
prologic |
I'd also like to have some notion of 2FA/U2F for SASL too |
| 03:02 |
|
bear |
yes please |
| 03:03 |
|
prologic |
absolutely |
| 03:03 |
|
bear |
this also makes end-to-end encryption much saner |
| 03:03 |
|
prologic |
let me file issues for all three |
| 03:03 |
|
prologic |
if at least to document how I plan to design them and to track them |
| 03:04 |
|
bear |
:nods: |
| 03:04 |
|
bear |
ugh - slack has invaded my habits |
| 03:05 |
|
prologic |
lol |
| 03:06 |
|
prologic |
I actually really despise both Hipchat and Slack |
| 03:06 |
|
prologic |
if not for the fact they are both bastardized crappier IRCs |
| 03:06 |
|
prologic |
At least Slack actually provided a half-assed IRC gateway into Slack uggh |
| 03:06 |
|
prologic |
I prefer the kind of approach(s) irccloud.com took |
| 03:06 |
|
bear |
agree completely |
| 03:07 |
|
prologic |
build a goddamn better client and improve the protocols (they are heavily involved with IRCv3 specs) |
| 03:14 |
|
pdurbin |
Gitter has a nice IRC bridge. I use it all the time. |
| 03:15 |
|
pdurbin |
I've never really used HipChat much. I tried Slack's IRC bridge a while ago. I can't imagine using it now because now my team is so into Slack that I feel like I should just embrace the full experience. I even installed the Slack app on my phone. |
| 03:15 |
|
pdurbin |
But I think of Slack as a work tool. IRC is for fun. :) |
| 03:20 |
|
* pdurbin |
looks at https://en.wikipedia.org/wiki/Cadmus |
| 03:20 |
|
pdurbin |
prologic: cadmus is your god of writing? |
| 03:21 |
|
prologic |
exactly right :D |
| 03:22 |
|
pdurbin |
"the carrier of the letter" |
| 03:22 |
|
prologic |
precisely |
| 03:22 |
|
prologic |
I'm (if anythinb) naming supporting irc bots/services that go along with eris well :) |
| 03:22 |
|
prologic |
eris soter and now casmus |
| 03:23 |
|
pdurbin |
Is https://en.wikipedia.org/wiki/Soter_(daimon) the page to look at? |
| 03:26 |
|
pdurbin |
"the personification or daimon of safety, preservation and deliverance from harm" |
| 03:27 |
|
prologic |
bear https://github.com/prologic/eris/issues |
| 03:27 |
|
prologic |
documented 4 new isseus |
| 03:28 |
|
prologic |
pdurbin that's the one! |
| 03:28 |
|
bear |
cool - i'll give it a look at end of week when I start holiday |
| 03:28 |
|
prologic |
the connotation here is Soter is the protector of IRC Channels :) |
| 03:28 |
|
prologic |
bear 👍 |
| 03:29 |
|
pdurbin |
https://en.wikipedia.org/wiki/Eris_(mythology) says "goddess of strife and discord". What could that possibly have to do with IRC? :) |
| 03:44 |
|
prologic |
haha |
| 03:44 |
|
prologic |
what do you think :P |
| 03:45 |
|
prologic |
I actually explain the reason for the name choices in the README(s) :P |
| 03:45 |
|
prologic |
I'm of course just making shit up anyway :P |
| 03:49 |
|
pdurbin |
Ah. "The connotation here is that IRC (Internet Relay Chat) is a place of chaos, strife and discord. IRC is a place where you argue and get into arguments for the sake of argument." Yeah, I like that. |
| 04:02 |
|
prologic |
:D |
| 08:27 |
|
|
aditsu joined #sourcefu |
| 13:59 |
|
dotplus |
I think that a big part of the problem with discussion of these communication protocols, technologies and their implementations is that it can be very difficult to evaluate each part separately. For example, the sociocultural aspects (such as: who uses/can use, who you can reach, what they talk about, acceptable topics/tone and other (sub)cultural norms) from technical aspects (ease of development |
| 13:59 |
|
dotplus |
(including vibrancy/maturity/etc of the dev ... |
| 13:59 |
|
dotplus |
... team(s)/project(s)), multiprotocol support, robustness/scalability/security options) of the clients or the servers from the UX aspects of the client(s) from the possibilities of the protocols themselves. |
| 14:00 |
|
dotplus |
When one says "I don't like Slack", what does that mean? |
| 14:01 |
|
dotplus |
the clients? (that the "native" clients are terribly resource intensive webkit(?)-based apps) |
| 14:06 |
|
dotplus |
the company? (that Slack Technologies, Inc. has at best a questionable stance on interoperability, privacy, AUP, and other legal issues) |
| 14:11 |
|
dotplus |
the social network aspects? (that instead of ~one (major) place to go for communication around open source technologies, freenode, I'm now constantly (not?) hearing about/joining new "Slacks" each of which has relevant channels) |
| 14:15 |
|
dotplus |
if only there were useful references (by academics? isn't that what they're _for_?) to sort out some of this mess, so the rest of us can have useful shared vocabulary in order to separate the orthogonalities and think clear about which aspects we care about in order to select/improve. |
| 14:18 |
|
dotplus |
argh, in that above, how could I leave out the question of sync/async communication? |
| 14:19 |
|
dotplus |
anyway, if anyone here is familiar with .edu research groups that are attempting to address this area, I would love a pointer. |
| 16:22 |
|
|
aditsu joined #sourcefu |
| 16:44 |
|
pdurbin |
dotplus: you're looking for an academic paper on why people like or dislike various chat solutions? |
| 17:25 |
|
dotplus |
errm, perhaps. descriptive material about which types of techs (protocols, networks and (client) software) promote/facilitate communication styles that are suitable for certain purposes. Even some sort of taxonomical work about the various different axes would be useful. I suspect that an academic paper is not really the most useful/suitable medium. |
| 17:25 |
|
dotplus |
I also suspect that such a thing does not really exist and I'd have to write it:( |
| 17:27 |
|
dotplus |
even starting to look into this is not straightforward. which dept? is it psychology? communications? sociolinguistics? IS? |
| 17:38 |
|
pdurbin |
buh, dunno, I think sivoais is our resident academic |
