2014-11-14 UTC
# 00:01 tantek kylewm - the problem is in your use of "the" (no seriously). It is either *your* lightsaber, or *a* lightsaber.
# 00:02 tantek implied in the SW mythology about lightsabers is that you have *only* one, that you are expected to have *built* it *yourself*, and that you *depend* on it as an extension of your *self*.
# 00:02 kylewm KartikPrabhu: are you making lightsaber sounds?
# 00:03 bret sidenote... checking out brianloveswords's super minimal but super awesome selfdogfooding POSSE site http://bjb.io
# 00:03 tantek in RoJ, Luke is selfdogfooding his own lightsaber that he made. however in Ep1, Anakin is merely dogfooding C3P0, which he built to also help his mom. C3P0 is not part of Anakin. No aspect of identity/self.
# 00:04 tantek wonders if he needs to add that example to the wiki
# 00:08 tantek.com edited /selfdogfood (+484) "attempt to make it more clear how selfdogfooding is above and beyond just dogfooding (just using your code on some site somewhere)" (
view diff )
pauloppenheim and npdoty joined the channel
# 00:17 KartikPrabhu in the same vein... are people who use indieweb code such as Known but not actively code it not follow the indieweb selfdogfood principle?
gr0k joined the channel
# 00:20 tantek KartikPrabhu: partially correct. they are not selfdogfooding Known, unless they actively *create* part of it (code is just one aspect - see /creator )
# 00:20 tantek However, there may be other things on their personal site that they create, in which case they are selfdogfooding those creations
benjamin-agaric joined the channel
# 00:21 KartikPrabhu so selfdogfooding is for code creators not artists and writer who would use indieweb code
friedcell and xxcoeurxx joined the channel
# 00:21 gRegor` Ooh, glad to see build your own lightsaber getting attention :)
gr0k joined the channel
# 00:22 tantek for example, I am selfdogfooding /Falcon as sole creator, and I'm even selfdogfooding the tmhOAuth library I use for relmeauth as a *partial* *fractional* contributor
# 00:23 tantek correct, just "writing" is not selfdogfooding. unless perhaps you're a philosopher and your website is actually demonstrating the very philosophy you write about
# 00:24 tantek if you wrote a recipe for dogfood, and then cooked it yourself, and ate it regularly, you would be selfdogfooding that writing
# 00:24 KartikPrabhu not just philosophers but if a photographer uses known to publish their photos, it does not count as a complete indieweb principle satisfaction
# 00:24 tantek but most writing is not recipes for something else. most writing does not "build" anything.
# 00:24 tantek KartikPrabhu: right, perhaps more generally, merely publishing is not dogfooding
# 00:25 tantek you have to be creating some aspect of a *tool* which you then *use*
# 00:25 tantek as a writer, if you invent new words, and use them as part of your personal identity on your website, you could be selfdogfooding those words
# 00:25 KartikPrabhu yeah. I suppose as long as the goal is to own their own photos/writing I suppose it doesn't matter if they are technically selfdogfooding or not
# 00:26 tantek selfdogfooding is a filter we applies to creators
# 00:26 kylewm whoa. selfdogfooding writing dogfood recipes. tantek you have gone too far!
# 00:26 gRegor` Interesting. Until now I think I considered people publishing on their own site as selfdogfooding. Thanks for the clarification.
# 00:27 gRegor` Yo dawg, we heard you like dogfood . . .
# 00:27 kylewm well hang on now, andicascadesf was chided yesterday for not selfdogfooding her suggestion to blog every day for a month
# 00:28 tantek because "blog every day for a month" is essentially a script that human executes. it's code.
# 00:28 tantek she wanted others to run that code, but she wasn't running it herself (or so it seemed)
# 00:28 gRegor` So yeah, I think that's a stretch.
# 00:28 tantek because she wasn't running it herself, on her own personal site, she was not selfdogfooding it
# 00:28 KartikPrabhu kylewm: saying"lets all blog once a day" is different from "join me and blog every day"
# 00:29 tantek which is different from "hey everybody, *you* should blog once a day"
# 00:29 gRegor` Gotta head out, though. Will catch up later.
# 00:29 tantek and: "*I'm" blogging once a day. You should too"
# 00:29 kylewm KartikPrabhu: you're right, chided was too strong. I cannot think of a better synonym atm
# 00:30 KartikPrabhu me thinks the whole "blog every day" thing is what turned people off blogging in the first place. Not everyone can come up with stuff to write everyday
# 00:30 kylewm KartikPrabhu: "let's all blog every day" vs. "join me and blog every day", which one of those is what?
# 00:31 KartikPrabhu kylewm: I am not de-constructing the exact words. maybe I am building a strawman who knows :P
# 00:31 gRegor` writing a blog post at random times (as the fancy hits you) is just publishing, not selfdogfooding, but setting up a schedule makes it selfdogfooding?
# 00:32 tantek gRegor`: depends, if your "setting up a schedule" is programmatic, and you share that code, and do it, you could be said to be selfdogfooding *just that code*
# 00:33 gRegor` I don't think anyone that "blogs every day" is writing code for it.
# 00:33 tantek gRegor`: I think you're missing the key aspect that certain subsets of English statements can be interpreted as code that humans execute.
# 00:33 kylewm between dogfood-ception and the definition if "the"
# 00:33 gRegor` Wait, is this evil tantek? jk
# 00:34 gRegor` Ok, *actually* heading out now
# 00:35 tantek KartikPrabhu: could you take a crack at documenting the clarifications above where you said "yup. got it " as FAQs on /selfdogfood ?
# 00:35 tantek since clearly you had questions, and they eventually got answered.
veselosky joined the channel
# 00:45 kylewm I feel like half of this page is tongue-in-cheek
emmak, j12t, snarfed, scor, KartikPrabhu, mko, gr0k, npdoty and JasonO joined the channel
# 01:38 kylewm aaronpk: have you looked into server sent events more? it seems like having a traditional http server + a realtime server communicating over redis (i.e. the way you have done real time comments so far) is way more common than i knew
xxcoeurxx joined the channel
# 01:39 aaronpk Yeah I have a prototype of my realtime comments mechanism using SSE now
fmarier joined the channel
# 01:39 aaronpk I like that it removed one of the moving parts, that I don't have to have a separate app running
# 01:40 aaronpk Yeah because SSE is really simple. It's basically just a way for the server to keep a connection open and send partial responses
# 01:40 kylewm how will you handle the thread problem though?
# 01:41 aaronpk We used to do this with long polling, except with long poking the client would have to make a new connection each time you sent an event
# 01:41 kylewm from the documentaiton, I'm having a little trouble differentiating what is necessary and what's necessary at scale
j12t joined the channel
# 01:42 kylewm it's not load that i'm worried about though...it's holding open those threads
# 01:43 kylewm i guess it would be the same issue with long polling?
# 01:43 kylewm and possibly i just don't know how web servers work
# 01:45 aaronpk So Apache will make a new OS thread for each request. I think with nginx that is not the case
# 01:45 kylewm but like right now my uWSGI application has 4 processes and 2 threads each... so if 8 people have the site open at the same time, it would be completely locked down
# 01:45 aaronpk Are you sure? cause node apps don't have that problem
cmhobbs joined the channel
# 01:47 kylewm and I'm also thinking that nginx/apache shouldn't matter, since they are just proxying to the application...
# 01:47 aaronpk Apache can either run the php code itself or proxy to a back end php server
gr0k joined the channel
# 01:49 aaronpk I have been switching to nginx but still have apache in places
# 01:50 kylewm if you are using php-fpm, I think it's the same kind of thing.. some number of worker processes and threads
# 01:56 kylewm he recommends Flask + redis + a realtime server
# 01:57 kylewm and also says bad stuff about Tornado (the Python equiv of node.js)
# 01:58 kylewm but I am like "you just recommended a realtime server"
# 01:59 aaronpk I think he's basically saying the thing that is keeping persistent connections open should not be your main web app
# 02:01 aaronpk (where webmention.io is providing the component that "maintains the standing connections and subscribes to redis")
# 02:02 kylewm I mean I think it's exactly how you've implemented them too
# 02:02 kylewm where node.js subscribes to redis and holds up the persistent connections
# 02:03 aaronpk so my new implementation of SSE using the same PHP app is explicitly moving away from the direction he's describing in that article
# 02:04 kylewm does it make sense why i'm worried that will make you very easy to DoS?
# 02:04 aaronpk I think there's a way to do it with nginx+php-fpm that won't be a problem
# 02:08 aaronpk can run using the same code base and doesn't require running a separate server process
# 02:11 aaronpk my current architecture requires running this node.js app (or could easily be written in ruby or python) as a totally separate thing from whatever web server you're running (nginx+php, nginx+ruby, python whatever)
# 02:14 aaronpk that means an additional process on the web server as well as listening on a different port. and if you're SSL, then you have to do SSL termination in both places.
benjamin-agaric, xxcoeurxx, j12t, gr0k, mdik_, fmarier, julian``, cmhobbs, ben_thatmust, brianloveswords, brianlov_, hmans and smcgregor joined the channel
richardlitt and KartikPrabhu joined the channel
# 04:58 bret prioritizing the completion of a specific task to as close as its original calling time tends to result in increased overall concurrency, despite only a single thing happening at any one given time
# 04:59 bret (one thing at a time due to the single threaded event loop)
tylergillies__ joined the channel
# 05:23 bret tylergillies__: no I haven't! but it looks neato
# 05:24 bret lets cross our fingers something cool comes out of that effort.
# 05:24 tylergillies__ the next big innnovation is where you're going to be able to host a website on the general 'internet' and not need an entire server
# 05:25 bret yeah that's kind of a big dream for a lot of people it seems.... myself included
# 05:25 bret (although I've settled on things like gh-pages and neocites at the moment)
# 05:25 tylergillies__ gh-pages is nice
# 05:25 bret the large file problem is still totally unsolved ime
# 05:27 bret nice. im a big fan of the fork n go/one click deploy model
# 05:28 tylergillies__ yeah, it used to be on wordpress but we kept getting hacked
# 05:28 tylergillies__ so i copy pasted the html to jekyll and went from there
# 05:28 bret I'm putting together a indieweb gh-pages boiler plate site right now
# 05:29 bret still trying to figure out how to use the Jekyll data folder as a centralized data store for things like people data and stuff
# 05:30 bret if http://ipfs.io/ can accept a git push with static site content, all of this could be hosted on something like that instead of gh-pages
# 05:33 tylergillies__ bret: i sent you a few bucks worth of bitcoin :)
# 05:34 tylergillies__ camilstore is pretty much a store all the things model right?
# 05:34 bret camlistore is an imutible datastore.... so everything you put in is saved to disk and deltas are layered on top as the data changes... everything you put in can theoretically come back out at some point
# 05:35 bret the idea is versioning for large files
# 05:35 bret implications are huge, but it assumes unlimited storage.... which is fairly realistic for google employees
KartikPrabhu joined the channel
# 05:36 bret at home, maintain large raids tends to be a pain in the kneck
# 05:36 bret it also handles things like http file tranmissions and file chunking
# 05:37 bret and has modeling layers on top of the data store so think gmail labels instead of folders
# 05:37 bret pretty neat, but I haven't been able to find a way to really make it work for me yet. its a huge project and is only a few years in
# 05:38 bret ime its confusing and slow to use though
# 05:44 bret i guess maybe we should also offer a syncthing option
# 05:44 bret but I dont even know how that does peering
# 05:46 tylergillies__ BTSync is pretty cool
tantek joined the channel
# 05:50 bret new personal policy... not writing about things I dont use
# 05:51 bret actually no, whats the point. ill write about whatever i feel like
# 05:52 tylergillies__ kylewm++
# 05:52 tylergillies__ pulse link i forgot that existed. now running
# 05:54 tylergillies__ That page is made in clojurescript
# 05:55 tylergillies__ im using core.async which is essentially golang goroutines in javascript
# 05:55 tylergillies__ i have a state function that fires off an update on all the dom elements on button click
# 05:57 tylergillies__ JVM is the best platform in the world
# 05:57 tylergillies__ literally billions of dollars went into making it battle ready
# 05:58 bret just noticed, the bitcoins i got from the bt faucet 2 years ago magically turned into 8 USD
# 05:59 tylergillies__ i found a whole bitcoin in one of my random backups
# 05:59 tylergillies__ before the crash so it was like 700 bucks
# 06:06 bear looks like I missed the "if you run a static site, how do you do ..." discussion above
# 06:09 tantek welcome tylergillies - add yourself to indiewebcamp.com/irc-people !
# 06:15 kylewm bear: do you know about server push stuff in python? having a hard time figuring out how to do it without going all in on gevent or Tornado
KartikPrabhu joined the channel
# 06:16 bear is the python app on the server that is going to do the push to a web client? another client?
# 06:16 tylergillies__ tantek: is there a dns based indie auth protocol? My main site has a pretty strict url structure and redirects are setup using docker nginx template regex lines so I'd rather not mess with it
# 06:17 kylewm bear: just trying to recreate aaronpk's "realtime comments", but without running two separate server processes
# 06:17 kylewm so the server will be pushing a (very) occasional event up to the client
# 06:17 bear I would use python websockets behind nginx for that
# 06:18 tantek tylergillies__: that looks like a group forum - not a personal site
# 06:18 tylergillies__ tantek: its the only site i have on the internets heh
# 06:18 bear tornado, gevent and all of those really just implement long polling
# 06:18 tantek do you not have a personal site tylergillies__ ?
# 06:18 tylergillies__ nope :(
# 06:18 tylergillies__ i have a domain
# 06:19 kylewm tylergillies__: I thought you were running Known
# 06:19 tylergillies__ no website though
# 06:19 tylergillies__ but that computer got formatted
# 06:19 tylergillies__ so now its not
# 06:20 bear kylewm - the choice of gevent or tornado (or even wsgi) will depend on what you have running now server side
# 06:20 kylewm bear: OK, I am discouraged bc gevent has not been ported Py3 yet. but perhaps ws4py and asyncio can work
# 06:21 tylergillies__ im surprised theres no indie auth protocol where you just add a TXT record or something
# 06:21 tylergillies__ tantek++
# 06:21 tylergillies__ good call
# 06:21 bear kylewm - for something simple I would run CherryPy under Python v3
# 06:22 tantek tylergillies__: configuring DNS is painful enough for most people, nevermind figuring out how each DNS provider has their own UI for editing TXT records etc.
# 06:22 kylewm bear: ahh, this is the first Im hearing about CherryPy. need to read up. thanks!!
# 06:22 tylergillies__ its not a file
# 06:22 tylergillies__ TXT is a dns record type
# 06:22 tantek KartikPrabhu: pretty sure he means a TXT *record* associated with your DNS
# 06:22 tylergillies__ my pgp key is in my dns
# 06:23 kylewm is glad he still has Loqi muted, with all the what is'ing going on
# 06:23 tantek tylergillies__: ^^^ that's why there's no indie auth protocol where you just add a TXT record or something
# 06:23 tylergillies__ tantek: righto
# 06:24 bear TXT record in DNS is the same IMO as having a well known url for discovery - it just avoids the web server but that is also it's weakness because it's not protected by SSL
# 06:24 tylergillies__ SSL bit is a good point
# 06:24 tantek so both a usability issue, *and* a security issue
# 06:25 bear now having a TXT record to aid in discovery may be useful, but IMO that is much better served by having microformat h-card at your root domain
# 06:26 tantek thanks KartikPrabhu - I think it's quite important to capture such nuances / details / clarifications about /selfdogfood
KartikPrabhu joined the channel
# 07:02 tylergillies__ indieauth.com isn't verifying my site :(
# 07:03 tylergillies__ hey it worked
# 07:05 tylergillies__ meh failed on second part
cweiske joined the channel
stream7, loic_m, Haxxa, eschnou, stream7_, alanpearce, friedcell, pfefferle and petermolnar joined the channel
# 08:45 tommorris it's the thing I've been wishing Amazon would do for a while.
pfefferle_, friedcell, Sebastien-L, gavinc_, friedcell1 and alanpearce joined the channel
friedcell, eschnou, xxcoeurxx, sammachin and Pierre-O joined the channel
cweiske joined the channel
scor, Loqi and aaronpk joined the channel
# 15:10 gRegor` Morning, indiewebcamp
Deledrius and snarfed joined the channel
gr0k joined the channel
# 15:24 Loqi Ok, I'll tell them that when I see them next
# 15:26 gRegor` Morning, ben_thatmust
# 15:34 aaronpk that would be a good one since they have a great api for that
brianloveswords joined the channel
# 15:37 snarfed aaronpk: sadly, in this case at least, the api isn't so great
chalettu joined the channel
# 15:45 petermolnar snarfed I digged a little around and getting favs and comments does not seem to be an issue once you have the list if image ids and getting the ids seem to be doable as well
danlyke joined the channel
verdi_ and jonathanfrei joined the channel
chrissaad, friedcell and brianloveswords joined the channel
caseorganic and danlyke joined the channel
snarfed and Pea1 joined the channel
# 16:59 bret gRegor`: stop it. that ALA article is triggering my OCD ;0
brianloveswords joined the channel
# 17:00 gRegor` I have yet to dive into <figure>, so no idea
# 17:00 gRegor` I would venture 'yes', loosely
# 17:00 bret in the academic paper sense, its like some kind of image, or graph with a caption and label
# 17:01 bret aaronpk: bummer when new things dont work on all the browsers :(
# 17:02 aaronpk oh yeah. I was mostly just commenting on the fact that there's an entire blog post with pictures hosted as a gist.
caseorganic joined the channel
# 17:05 bret i seen people do full on articles in gists before, but not with photos
indie-visitor joined the channel
# 17:19 Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
gr0k and wolftune joined the channel
# 17:43 kylewm Is the plan for IWC Online to have 1/2 day of discussions and 1/2 hack day?
j12t and gr0k joined the channel
chrissaad1 joined the channel
# 18:07 gRegor` Haha "22:00 Camp Closed - clean up your room"
# 18:08 aaronpk I should totally get pizza delivered to my room for this
# 18:08 gRegor` OMG, @indiewebcat
# 18:12 GWG I'm getting a dryer delivered in the middle of it
friedcell joined the channel
friedcell, pauloppenheim and snarfed joined the channel
friedcell, wont_on and chrissaad joined the channel
# 19:09 snarfed hey all! …just to confirm, we don't yet have explicit mf2 for checkins, right?
# 19:09 snarfed of the five examples i looked at, most use p-location and/or h-card for the venue, and occasionally h-geo, but nothing like h-as-checkin or similar
# 19:17 kylewm I think he is the most prolific poster of checkins
# 19:19 kylewm didn't glennjones add automatic gps based checkins?
caseorganic and friedcell joined the channel
# 19:20 aaronpk i'm not a fan of automatic checkins. automatic prompts sure, but not actual posts.
# 19:21 kylewm glenn's do not appear to be associated with a venue
# 19:22 kylewm anyway, I think "p-location h-card" is correct if it's a venue, "p-location h-geo" if it's coordinates
# 19:22 aaronpk that still doesn't indicate it's an actual checkin though. because a photo at a venue would also be "p-location h-card"
7GHAAMMZO joined the channel
# 19:24 snarfed kylewm: kinda. maybe more like we just don't have a checkin post type
# 19:24 snarfed i don't know that we need one. just thinking out loud
# 19:24 aaronpk you mean like it's also ambiguous whether something is a note+photo or a photo+caption?
tantek joined the channel
# 19:25 aaronpk I feel like it doesn't really matter for notes/photos, but there is definitely a distinction with checkins
# 19:25 tantek and they completely forgot/ignored the "cite=" *attribute* on <blockquote> sigh.
herbsmn joined the channel
# 19:26 snarfed tantek: yup, read that a few times. definitely helpful but didn't provide an explicit mf2 class…which is fine
# 19:27 tantek hmm - look at that - no "how to" section. oops.
# 19:27 gRegor` Interesting, tantek. realized after sharing the link it's also from January 2013
# 19:27 snarfed right now i'm going with .h-event > .h-as-checkin > .h-card.p-location
# 19:28 snarfed (er, sorry, top level is .h-entry, not .h-event_)
# 19:28 tantek or put it on the same element as the .h-entry
# 19:29 snarfed (same reason i have weak/hidden reply contexts :/)
# 19:29 snarfed if we don't like h-as- classes, i can just drop the h-as-
# 19:29 snarfed i want an explicit checkin class, but i don't care much whether it's mf2
# 19:29 tantek problem is not the h-as-, but rather the extra level of object
# 19:29 aaronpk tantek: I don't think that's accurate. (p-location venue implies checkin)
# 19:30 Loqi gives snarfed an explicit checkin class
# 19:30 snarfed aaronpk: right. tantek: most actual users of .h-card.p-lcoation right now probably wouldn't expect that that alone implies a checkin
# 19:30 aaronpk i've been publishing notes with venues lots of the time, and they certainly aren't checkins
# 19:30 tantek aaronpk - I'd say they're implicit checkins at a minimum
verdi_ joined the channel
# 19:31 snarfed heh ok. debating implicit vs explicit is where i back away slowly :P
# 19:31 aaronpk trying to find a counterexample but it may take some digging. i'm looking for the example where I post a photo the day after I take it
# 19:32 tantek snarfed - no it's ok. aaronpk's example is a photo due to the u-photo
# 19:32 tantek that test is before the " (p-location venue implies checkin)"
# 19:32 tantek snarfed: not an accident. that was a deliberate ordering on my part.
Sebastien-L joined the channel
# 19:32 aaronpk so I was certainly not checking in to ground kontrol when I posted that on the 4th
# 19:32 snarfed tantek: i mean, people could still post with p-location and not intend it as an explicit checkin
# 19:33 aaronpk (also I've definitely attached a photo to checkins)
# 19:33 tantek kylewm: they can - however, a u-photo right now primarily implies a *photo* post, not a checkin
# 19:33 tantek the implied post type algorithm is evolving based on what people publish and what they (seem) to imply by it
# 19:35 tantek regardless, snarfed, it's a good question - "how should a check-in be marked up?"
# 19:35 tantek I have the additional problem with many of my checkins that they are "published" after the time of the actual checkin (nevermind the photo aspect)
# 19:36 tantek If I were publishing checkins I'd want a *separate* datetime property for the time of checkin
# 19:37 tantek (optional of course, with the dt-published being the default implied time of the checkin)
# 19:46 kylewm seems like he's talking about indepdent web publishing, not IndieWeb stuff like webmention
# 19:47 kylewm @DeadSuperHero is saying Diaspora is the way to get federation into people's personal sites
j12t_ joined the channel
# 19:50 tantek kylewm: except that Diaspora doesn't federate with any other implementations, so it's not really federating.
glennjones joined the channel
# 19:52 kylewm I think it's a new project by the same person who did friendica
caseorganic joined the channel
# 19:53 gRegor` what is postsquare?
# 19:53 tantek kylewm: are you saying RedMatrix is running pump.io?
# 19:54 tantek reads like another single-open-source-project monoculture
# 19:54 kylewm tantek: no I'm saying statusnet is related to pump.io in the same way friendica is related to redmatrix
# 19:55 kylewm which I believe is the "interfederation" effort
# 19:55 tantek does anyone not see the rampant irony in every "federation" project requiring you to run *their* software?
# 19:56 kylewm friendica actually talks to lots of different networks
# 19:56 tantek kylewm - that's a big claim. citations needed.
# 19:58 kylewm tantek: friendica.com/features##Built-in+support+for+OStatus+federation
# 19:59 kylewm they claim first-class federation with statusnet and with Diaspora
# 19:59 tantek does anyone run friendica on their personal site?
# 20:00 tantek kylewm: lots of claims of federation - especially on mailing lists and readmes. Can you point to any *evidence* of actual federation happening? E.g. a post from one site/system running some software making it to another site/system running some other software, and a reply being sent back?
gavinc_ and lukebrooker joined the channel
# 20:03 kylewm "first class federation" I would define as you haev ONE account that talks to lots of different servers running different software
snarfed joined the channel
# 20:03 kylewm and I'm not trying to sell you on Friendica, just trying to answer questions about it
# 20:05 tantek kylewm: that's a reasonable start of a definition
# 20:05 tantek having seen years of claims of federation but very little actual evidence / usage, I remain skeptical of any cross-system federation claims
# 20:07 Loqi federation in the context of the indieweb refers to services and features on indieweb sites that work directly with other indieweb sites peer-to-peer, without being bottlenecked by any kind of centralized service or silo http://indiewebcamp.com/federation
# 20:08 tantek so that's why they are asking for it to be an option on Indiehosters
# 20:11 kylewm showing posts from Diaspora in his redmatrix instance
# 20:12 tantek why is it behind a login and the POSSE copy not?
# 20:13 tantek so pardon me if I don't believe there's some invisible original private post from which it originated.
# 20:13 kylewm I reeeeally don't have a dog in this fight; I just know that friendica was the one effort that didn't seem to want purity
# 20:13 tantek I don't have a dog in this fight either - I'm skeptical of all the federation claims.
friedcell joined the channel
# 20:17 kylewm and I don't think you're wrong to be skeptical
# 20:18 kylewm you said "does anyone not see the rampant irony in every "federation" project requiring you to use *their* software", and I was just saying that they didn't do that
gavinc, brianloveswords, friedcell and thedod joined the channel
# 20:40 thedod trying to deploy kylewm's redwind on a pretty much broken machine (compiled python3. twice). Getting there :)
glennjones joined the channel
friedcell and tantek joined the channel
# 20:53 tantek kylewm - that's a good point. even if it's only claimed, saying you federate with other projects is incrementally better than "install our software to federate".
# 20:54 tantek note to self, filter on "To better focus our efforts and to align them with the needs of our global member base" -> likely hollow rationalization for what comes next
# 20:56 kylewm but that's a shame, A Better Queue was a really nice service (basically just a nice looking mashup with Rotten Tomatoes)
# 21:23 aaronpk do you think people would say that bittorrent sync is federated? it's all peer-to-peer! just run this software! ;)
# 21:24 tantek perhaps we need to define levels of federation
eschnou joined the channel
# 21:34 bear there can be different things you can transfer in a federated system, but federation itself has only one meaning - the ability to send messages between different networks
# 21:34 herbsmn I'm pretty sure the government invented federation. I mean, it's right in the name! FEDeration!
# 21:38 gRegor` Wonder if we need a "near miss" section of site-deaths
# 21:41 tantek gRegor`: that's probably better captured on the mlkshk page
# 21:46 Loqi Ok, I'll tell them that when I see them next
# 21:49 snarfed tantek: probably not implementing, but i may compose them manually occasionally
# 21:51 thedod Hello again. An "etiquete" question: can I install my redwind as /indie (and add some microformat hints at /)? Or should my indiewebcamp presence be proudly displayed at /?
# 21:52 tantek you can install and run whatever software you want in whatever URL structure you want ;)
# 21:52 tantek it is likely that people will refer to your personal domain as your indieweb presence
# 21:52 thedod now the question is whether I'll bump into tech problems, but that's minor :)
snarfed1 joined the channel
# 21:57 kylewm thedod: it will probably take some work to get it to run outside of /
# 21:57 thedod Maybe I've toyed with having a tahoe-lafs cloud as my / for long enough
# 21:59 thedod it lets you store data on a 1 or more hosts you don't necessarily trust, so that they can't tell what's there, and also they're redundant (even if they inject false data, you can recover)
# 21:59 thedod today there are many services claiming they're "like dropbox but really secure"
# 22:00 thedod this is a decade old project for banks who want backup. not a toy
# 22:03 thedod now it's handy to have as a filesystem (backup, sharing, etc.), and I loved the trick that they did a plugin to tiddlywiki that lets you update to the cloud (if you run the client and do this on localhost), but maybe it jumped the shark
# 22:04 thedod 5am here, night is young. let's hope the nginx is willing
mlncn and tantek joined the channel
# 22:37 thedod kylewm, first time I deploy with uwsgi and I wonder: config.py gets stuff from env. where do I make these env declarations in a uwsgi+nginx environment?
# 22:38 thedod I also need to declare the virtualenv to run the "right" python
# 22:39 thedod (compiled it 4 times today it would *better* be the right one)
tantek, krendil, snarfed and brianloveswords joined the channel