2014-03-21 UTC
# 00:00 tantek or rather, if the format recommends escaping, it's a format failure.
# 00:01 tommorris aaronpk: there’s nothing particularly wrong with putting Atom elements inside RSS 2.0 feeds, except for the whole RSS 2.0 thing. ;)
# 00:04 tantek hey aaronpk - next IndieWebCamp is actually NYC so I'm going to fix the top level redirects to point to that (rather than /2014 PDX)
# 00:06 tommorris my feed is broken at the moment. I’m so tempted to just replace it with the output of an hAtom converter.
# 00:08 tantek do it! Would be interesting to see the result
# 00:08 tantek that would be a nice thing to run at the community level
# 00:12 gRegor` Next I'm going to set up a slightly modified version for apprentices.
# 00:12 tommorris tantek: the only thing that is putting me off is I have some custom feed stuff for App.net
pfenwick, krendil, KevinMarks, melvster, KartikPrabhu, ricmac_, npdoty, fmarier, scor, pauloppenheim and ricmac joined the channel
# 01:22 kylewm I used the hashtag #siloproblems on Twitter, and it would appear the other hits are people dealing with actual grain silos, and people who work at a fish-n-chips place in NY :D
# 01:23 snarfed i have this far-off idea to eventually bundle activitystreams-unofficial, portablecontacts-unofficial, etc all together, unify the APIs, and rename the whole library "granary" in line with the silo metaphore
# 01:24 snarfed yet another buy-the-domain-and-maybe-never-get-to-the-actual-project
# 01:24 kylewm love it, and I just learned the correct spelling of that word
pfenwick joined the channel
# 01:27 kylewm ohh, hadn't looked at portablecontacts before
gRegor` joined the channel
# 01:30 kylewm I like the ~ though, reminds me of the good ol' days :)
# 01:32 kylewm I guess the idea that (e.g., if you lost your database), you could conceivably re-create the links if they are based on something
# 01:32 KartikPrabhu oh i see. is the problem that if my database table changes then all ids will be mixed up and mess up my urls?
# 01:33 kylewm the database won't renumber your rows, but if you exported and reimported somewhere else, maybe they could?
# 01:35 Loqi Ok, I'll tell him that when I see him next
# 01:36 Loqi Ok, I'll tell them that when I see them next
# 01:39 kylewm I followed tantek's algorithm more or less to the letter (except I think I am using UTC times where he uses Pacific)
# 01:41 KartikPrabhu oh no found the algo. I was asking datetime can't be enough but there is some unique identifier
# 01:42 kylewm oh, yep datetime.toordinal() - date(1970, 1, 1).toordinal()
adambrault joined the channel
# 01:43 KartikPrabhu yeah but usually there can be multiple posts a day specially for notes
# 01:43 kylewm so there is a fourth # tacked on at the end that is a counter within the day
# 01:44 kylewm which feels a little bit like the same problem algorithmic warns against, but i couldn't think of a way around it
# 01:44 kylewm that would make it more difficult to reconstruct in the future
# 01:44 kylewm wondered if ppl "hide" notes instead of deleting them
# 01:45 KartikPrabhu but i like the algorithmic idea in spirit. interesting. will wait for tantek to arrive
# 01:45 kylewm cool, I'm heading out but will be interested to know what you find out, good night!
pfenwick and fmarier joined the channel
# 03:11 bret Made a little bit of progress. Hating the node express web framework the more I use it. iwc-log-feed.herokuapp.com
gRegor` and jedahan joined the channel
# 03:59 bear our node devs like Hapi (after having done express for 2 yrs)
KevinMarks joined the channel
# 04:37 bret Trying to get their deva to write better exMples and tutorials
tilgovi and jedahan joined the channel
# 04:48 bear yea, they have a contest active for that
netweb and snarfed joined the channel
kylewm joined the channel
# 05:21 bret maybe, i need a lot of hand holding still. like, full on project walk throughs
krendil joined the channel
# 05:21 bret im starting to think node isnt the greatest place to start
KevinMarks, kylewm, cweiske and melvster joined the channel
# 06:51 cweiske it'd be nice if the wiki would store my URL, so that I don't have to type it on the next login
tantek joined the channel
# 06:54 Loqi tantek: KartikPrabhu left you a message 5 hours, 19 minutes ago: I was implementing a shorter URL scheme using db ids. but a bit confused about it due to http://indiewebcamp.com/algorithmic ... any elaboration would be helpful! :)
npdoty joined the channel
# 07:10 tantek !tell KartikPrabhu using db ids is NOT algorithmic and introduces fragility.
# 07:10 Loqi Ok, I'll tell them that when I see them next
# 07:11 KartikPrabhu tantek: yeah I sort of understood that later. but ins't using n-th note on this day fragile too, if you delete that note?
# 07:11 Loqi KartikPrabhu: tantek left you a message 56 seconds ago: using db ids is NOT algorithmic and introduces fragility.
# 07:12 tantek if you delete a note then it's fine for it to 404, unless it's a comment in which case you want it to 410 to follow comment CRUD protocol
# 07:13 KartikPrabhu yes but if you base your URL shortening on "date + nth note on this date" (assuming that is what Whistle is doing) then it will mess up the other URLs
# 07:13 pauloppenheim KartikPrabhu: if you want a tag more specific than day, would timestamp work?
# 07:14 tantek KartikPrabhu - per the need to 410 gone for deleted comments, even your deleted notes should have tombstones
# 07:14 KartikPrabhu pauloppenheim yes that is one of the thigns I am considering. but doing it with slugs would be best IMO
# 07:15 tantek also doing it with slugs is a bad idea, because your slugs should be editable without breaking the permalink
# 07:15 tantek no, depending on them for uniqueness is a bad idea
# 07:16 tantek kartikprabhu - all these design decisions and more are explained here: tantek.com/w/Whistle
# 07:16 tantek pauloppenheim: or you might post more than one thing per second
# 07:16 pauloppenheim i'm personally not convinced by the idea of "permalink", i'm not sure anything on the web will be "permanent"
# 07:16 KartikPrabhu yes diud read that hence all the confusion. That goes against everythign I have built so far :(
# 07:17 tantek there's lots there that goes against A LOT of things that were built before
# 07:17 KartikPrabhu true. but now I am conflicted whether to correct using slugs for unique URLs or to keep going with it
# 07:17 tantek KartikPrabhu - you asked about the slug, but you must have missed this then if you did read that page: "The slug is purely for display and search, the unique portion of the URL is actually:"
# 07:18 tantek anyway, the slug is semi-human-readable, thus it should be editable without breaking anything
# 07:18 pauloppenheim the issue is selecting immutability - you can make anything not change, so long as it remains actually immutable
# 07:20 pauloppenheim i'm a bad person to talk to about this right now, i'm actually dealing with it professionally right now
# 07:20 tantek pauloppenheim: not helpful for a permalink, since content does change :P would be helpful for a checksum, but that's different.
# 07:20 KartikPrabhu on a more positive note. Just showed off POSSEing notes to twitter to a friend and the reaction went "oooo nice!!! O_O le me is impressed"
# 07:21 tantek re: "sounds like designing" - that's why I documented my design with reasoning, so you could just copy it if you wanted to.
# 07:21 KartikPrabhu tantek: yes. and I am very glad you document all of this. Much to learn!
# 07:22 tantek I was tired of people building fragile systems, and then copying their designs in a cargo cult manner.
# 07:22 tantek So I figured I'd fix it, and show my work, hoping that some folks might find that (and check my work!) instead of just doing things the way everyone else does things.
# 07:22 tantek database ID crap, slugs as fragile permalinks etc.
# 07:25 tantek there's a lot of crappy designs in frameworks. A LOT.
# 07:25 tantek pauloppenheim: creation dates are not fragile.
# 07:26 tantek pauloppenheim: they are memorable approximately, and relative to other things, so as long as you can remember things nearby, you can easily find them
# 07:26 pauloppenheim i've actually had problems with this from using a date indexed system for over a decade
# 07:26 tantek KartikPrabhu: got to start with well designed URLs for a website
# 07:28 pauloppenheim it may be worth giving up on the idea if your content is still searchable somehow
# 07:28 tantek also for searchability - hence the permashortcitations
# 07:30 tantek pauloppenheim: I'm having difficulty connecting any of your criticisms to real world examples.
# 07:30 pauloppenheim i took a photo of a notecard of the 10 places i want to visit in the next 20 years, sometime in the past 5 years
# 07:31 pauloppenheim i can sit and try to specify the date better, but if it had only been tagged "travel", i would be able to find it in a snap
# 07:32 pauloppenheim it is located at a canonical path on my hard drive, but that path has no indication of the topic.
LauraJ joined the channel
# 07:45 pauloppenheim i wish my scenarios wasn't a problem, because then I would have one less problem :)
# 07:46 Loqi Got it! There are now 10 spammers blacklisted
# 07:46 tantek pauloppenheim: that's poor reasoning. yes you should tag stuff. does the obviate the need for simpler permalinks? no. do permalinks solve everything? no. they never claimed to.
# 07:47 pauloppenheim cweiske: right, but the wikipedia solution works for that constraint as well
# 07:47 tantek pauloppenheim: wow really? simple and short way of sharing a pointer to some info on the web.
# 07:48 tantek pauloppenheim: not true. stuff (including URLs) get deleted on wikipedia all the time by deletionists.
# 07:49 tantek KartikPrabhu: no, I think there's a whitelist
# 07:49 tantek but if you want to simple do the commands with a !tell aaronpk,tantek in front of it - one of us will notice and do it I'm sure ;)
# 07:50 tantek pauloppenheim: "does not require url design" - only if you don't care about the UX of URLs
# 07:51 tantek WP pages get moved then original URLs get re-used
# 07:51 tantek WP pages for one thing / person get used for another
# 07:51 tantek so no, WP URLs are crap for permalinks - sorry - and that was the case you were making
# 07:52 tantek haha that's funny - clearly you haven't see what happens with moves and redirects and overwrites thereof
# 07:53 pauloppenheim i'm confused, because i have, and that is exactly the reason i am advocating its usage
# 07:53 tantek once you've moved a page, and it leaves a redirect in place, if that redirect is replaced with content, then POOF all previous history is lost
# 07:54 tantek anyway, off to bed, if you think Mediawiki is a good solution for permalinks, then run it on your own domain!
# 07:54 tantek til then - your advocacy is more talk than real
# 07:54 tantek " if that redirect is replaced with content, then POOF all previous history is lost"
# 07:55 tantek anyway, like I said, go for it, run Mediawiki on your own domain
# 07:56 tantek until then, you're comparing vapor with shipping sites
# 07:57 tantek "once you've moved a page, and it leaves a redirect in place, if that redirect is replaced with content, then POOF all previous history is lost"
# 07:59 pauloppenheim i mean, hopefully the URL has meaning, and the new thing that goes there has the same meaning as the old thing
# 07:59 tantek I've seen others (couple of years ago, not sure which now) that were not
# 07:59 tantek so yeah, if people have good behavior with change comments, that helps
# 07:59 tantek but that's a people thing, not a system thing
# 08:00 tantek no the data is only there in the change comments, which are based on human behavior
# 08:00 pauloppenheim but yeah, not necessarily awesome, but possibly more human-valued than a raw date, but maybe not as much as a slug
# 08:02 tantek it is in the same way that random referrer query strings are irrelevant
# 08:02 tantek nothing says the entire URL has to be the "perma" part of the permalink
# 08:03 tantek until then I think you're arguing about mostly a hypothetical way of maybe using some of mediawiki's mechanisms but reimplemented
# 08:03 tantek so - have no idea if anything you're suggesting will actually work as you claim
# 08:04 tantek the design is by no means complete, but I'm fairly convinced of the line of thinking.
# 08:04 tantek might still run into obstacles while implementing, usual caveat
# 08:04 tantek figured I'd at least braindump the design so far rather than waiting for a sense of completeness/pefection
# 08:05 tantek anyway, see what you can get working minimally and quickly
# 08:05 tantek or feel free to park your design thoughts somewhere too and move on to something simpler and quicker. both are reasonable approaches.
eschnou joined the channel
friedcell, jonnybarnes, yaf, bnvk, glennjones, tantek, michielbdejong, pfenwick, LauraJ, eschnou and Acidnerd1 joined the channel
melvster, Sebastien-L, scor, LauraJ, jedahan, tantek, glennjones, fffuchs, chloeweil, squeakytoy and Guest9599 joined the channel
LauraJ joined the channel
KartikPrabhu joined the channel
yaf joined the channel
# 13:54 KartikPrabhu looks like tantek just got invited to a twitter indieweb/RSS discussion
kylewm, Sebastien-L and LauraJ joined the channel
kylewm_ joined the channel
# 14:26 kylewm I'd love to see an example of a "tombstone" for a deleted post, if anyone has one handy
_6a68 and snarfed joined the channel
paulcp joined the channel
KartikPrabhu and gRegor` joined the channel
adambrault joined the channel
# 15:13 KevinMarks Irony of Dave doing this on twitter, where he blocks me is amusing.
ttepasse and bnvk joined the channel
# 15:45 kylewm Is it worth actually tracking deleted post to differentiate between "never here" and "not here anymore"?
LauraJ joined the channel
# 15:54 KartikPrabhu kylewm: shouldn't "never here" be a 404 and "not here anymore" be a 410?
# 15:56 kylewm KartikPrabhu, that's what I'm wondering. It looks like idno returns 410 for both cases
bnvk joined the channel
# 16:03 KartikPrabhu kylewm: yeah you are right. At first I thought the pages might be the same but different status codes, but no they are both 410
jedahan joined the channel
# 16:04 KartikPrabhu !tell benwerd: any reason idno seems to return a 410 status code even if the page never existed, instead of 404?
# 16:04 Loqi Ok, I'll tell them that when I see them next
KevinMarks_, indie-visitor and caseorganic joined the channel
# 16:19 bnvk !tell aaronpk how often does the wiki kill my session login?
# 16:19 Loqi Ok, I'll tell them that when I see them next
caseorga_ and GregFalken joined the channel
# 16:26 Loqi aaronpk: bnvk left you a message 7 minutes ago: how often does the wiki kill my session login?
# 16:26 aaronpk bnvk: I tried to set the cookie length to like 60 days, but that might only take effect after the next time you log in
# 16:26 aaronpk it seems to not be killing my session frequently the last couple days
# 16:26 bnvk awesome, yah, it killed my session while I ate lunch and lost something I was typing :)
# 16:26 aaronpk yeah i'm still logged in this morning after not touching the site since like 18 hours ago
# 16:27 aaronpk ah yeah, I never trust web forms to actually keep what i'm writing over a long period of time in general
# 16:27 aaronpk either short edits committing often, or I write in a local text file
# 16:28 bnvk usually I do as well, but this was only 2 sentences
# 16:28 aaronpk whoa @davewiner went on a twitter rampage this morning
benprew and paulcp joined the channel
# 16:33 KartikPrabhu aaronpk: yes that would be one reason for the whole deleted/never here discussion
LauraJ and tantek joined the channel
# 16:35 aaronpk ah cool just wanted to make sure you saw that page :)
ttepasse and LauraJ joined the channel
cweiske joined the channel
snarfed and paulcp joined the channel
eschnou and iangreenleaf joined the channel
# 17:03 snarfed i've always said that's the key thing indieweb is missing
# 17:04 bnvk hehe, you joke, yet I have plans... mwuhahahaha
# 17:05 bnvk well, actually, it would work really nice with this idea floating around the crypto currency communities called DAO (distributed autonomous organization), imagine auto payments getting handled via a blockchain upon install of something
# 17:07 kylewm snarfed: is it possible to use activitystreams-unofficial as a library, without google app engine?
# 17:08 snarfed it *should* be usable w/o app engine…but may be rough
# 17:08 snarfed i definitely want it to work though, so i'm happy to help
# 17:09 snarfed it may be as easy as pulling out the appengine imports and cleaning up any dangling usages, which should be minor
# 17:09 snarfed i haven't updated it in a while though. i'll do that now.
# 17:10 snarfed bnvk: interesting! DAO sounds like "direct democracy" or "participatory economics"
# 17:10 aaronpk is not interested in blockchain-based currencies. wake me up when there is something different.
# 17:11 snarfed currency is the kind of thing where i want to be a very, very, very *late* adopter. not an early adopter.
# 17:12 aaronpk has actually used a real "indie" currency daily for the past 8 years
paulcp_ joined the channel
# 17:12 aaronpk i say "real" because I am able to get use out of it and none of you have to even know it exists
# 17:13 bnvk aaronpk: what is your particular qualm with blockchain based technology as a whole?
# 17:13 bnvk aaronpk: I figured you just didn't like the runaway money hungry investor bitcoin mentality, not blockchains as a whole
# 17:15 kylewm snarfed, ok I didn't need to test at all, just look closer at the regex, oops
# 17:16 aaronpk doesn't really want to get into it. it's a relatively unpopular opinion these days, and I have something that works for me, so meh
# 17:16 snarfed updates twitter-activitystreams.appspot.com anyway
# 17:16 snarfed kylewm: happy to accept PRs if the regex needs improving
# 17:16 kylewm it expects (sensibly) the citation to be at the end of the source .. but when you include media in a tweet it sticks a pic.twitter.com link at the end
# 17:16 gRegor` Is the indie currency you speak of, "USD", aaronpk? :)
barnabywalters joined the channel
# 17:17 snarfed i originally accepted permashortcitations anywhere, but there were too many false positives
# 17:17 bnvk aaronpk: fair, I wasn't wanting to rile you up, was just curious as I remember how interested you were in making your own currency / methods of exchanging money
# 17:17 gRegor` Ooh. Nice, aaronpk. :)
# 17:17 gRegor` I like the idea of coffeecoin.
# 17:17 snarfed kylewm: just fyi, PSCs never got much adoption, and it's not clear that they will
# 17:18 bnvk gRegor`: wait, did you see my blog post?
# 17:18 aaronpk which is only vaguely a currency, but then bitcoin is also just a shared ledger with a much larger group
# 17:18 gRegor` bnvk: Nope. Link?
# 17:18 snarfed permashortlinks may have a bit more potential, e.g. (kyl.im/n/123) as opposed to (kyl.im n/123)
# 17:18 bnvk oh, ref: aaronpk mention of coffee, hehe
# 17:18 gRegor` Oh, I thought maybe you'd written a post about "coffeecoin"
# 17:19 kylewm snarfed, interesting, yes that makes sense. trade people asking "what are those cryptic codes at the end of yoru tweets" for "why do you always link to stuff at the end of your tweets"?
# 17:20 kylewm snarfed, thanks for the food for thought (and I still want to play with activitystreams, but GAE was giving me fits this morning so I gave up)
# 17:20 snarfed kylewm: yeah. i still don't think we've found a way to do original post links in twitter that other twitter users are ok with
# 17:20 gRegor` PSC? Perma short code?
# 17:20 snarfed np kylewm! feel free to ping me on gae problems too, i'm pretty familiar with it. :P happy to help
# 17:21 snarfed i wouldn't encourage using them, but that's just my opinion
# 17:22 gRegor` Yeah, it doesn't seem likely we'll ever need to prove our indieweb post is the original post, right?
# 17:22 snarfed oh no, they're not really for "proof" as much as just linking back without implying that there's more content in the original post when there actually isn't
# 17:23 gRegor` And original post discovery, right?
# 17:23 gRegor` It would be nice if Twitter supported it as metadata
# 17:24 kylewm I looked through the API quite a bit, was hopeful that they'd let me sneak something into "entities" but no dice :)
# 17:27 kylewm what if you had a webmention-like endpoint for syndication-reverse-lookup?
# 17:27 kylewm i guess it would add yet another step to original-post-discovery... need to give it some more thought
# 17:28 aaronpk but you'd have to know the domain to go look at before you could do that
# 17:29 snarfed aaronpk, kylewm: true. easy to use the web site in the profile, though, rel-me/indieauth style
# 17:29 snarfed bridgy already now pulls and uses that when it's available. needed it for publish.
barnabywalters joined the channel
# 17:32 aaronpk like how do you get to the profile given a permalink to a tweet?
snarfed1 joined the channel
# 17:34 tantek pretty sure choice of symbols is already covered in tantek.com/w/NewBase60
snarfed1 joined the channel
# 17:35 snarfed1 btw, little-known feature: you can use twitter-activitystreams.appspot.com (and facebook-activitystreams.appspot.com, etc) to convert any silo post to mf2, either html or json
# 17:35 tantek KartikPrabhu: what war on RSS? the Atom/RSS war? That's over. They both lost.
# 17:36 snarfed1 both web ui and api available. just authenticate and then set the format drop-down to json-mf2 or html
# 17:36 KartikPrabhu tantek: no i realise I can't use both "human readbility" but I could replace underscore with hyphen? I was referring to a Dave Winer tweet :P
# 17:37 aaronpk tantek: actually it looks like hyphen appears first in the list of "HTML ID and NAME tokens", and the derivation section fails to mention why underscore over hyphen
# 17:39 kylewm tantek, I was going to but later realized it was doing it for every outgoing link
# 17:40 kylewm OK, just at first I thought it was explicitly marking t.co (twitter) links
# 17:40 kylewm (like how instagram photos don't show up in twitter streams)
# 17:42 tantek Twitter is insisting on others' adopting their proprietary TwitterCards metacrap
# 17:42 tantek Instagram just said no to that, which is perfectly reasonable
# 17:42 aaronpk yeah, just pointing out that it wasn't like twitter was directly attacking instagram on that
# 17:42 gRegor` Hm, I always assumed it was Twitter since instagram links worked one day but not the next
# 17:42 tantek Instagram experiment with Twitter Cards and then dropped them. I don't blame them.
# 17:42 aaronpk instagram just removed the proprietary twitter card markup
# 17:42 gRegor` I thought it was exactly that, since they launched their own pic service
# 17:45 KevinMarks After twitter revoked instagram friend lookup on twitter, iirc
# 17:45 tantek BTW - I was able to get instagram to adopt rel=me - so if there's more standard markup we want them to add - we can do so too
# 17:45 tantek KevinMarks - yup - love those bizdev mindsets
# 17:46 tantek ShaneHudson - how does that slide support anything indie*web*?
# 17:46 KevinMarks The really stupid one was twitter revoking Google's firehose access. Now they're wondering why growth is falling off with tweets hard to find
# 17:47 Loqi Ok, I'll tell them that when I see them next
barnabywalters joined the channel
benwerd joined the channel
# 17:52 Loqi benwerd: KartikPrabhu left you a message 1 hour, 48 minutes ago: any reason idno seems to return a 410 status code even if the page never existed, instead of 404?
squeakytoy joined the channel
# 17:56 benwerd (The SEO reply to @EvilTantek is the best thing I've seen this week)
# 17:56 snarfed1 KartikPrabhu: as tantek would say, "step back slowly"
# 17:56 snarfed1 hey benwerd! long time no see
# 17:56 snarfed1 glad austin didn't kill you :P
# 17:57 KevinMarks This has led to reports of twitter abolishing @ and #, though Vivian schiller denied that
# 17:57 tantek SEO folks enjoy jumping through Google's hoops no matter how high
# 17:57 benwerd snarfed1: I literally had to have an FBI background check while I was there. Rigorous.
# 17:57 tantek Google's good at making them SEO folks do tricks.
# 17:57 snarfed1 ugh snarfed1. brb.
# 17:57 tantek KevinMarks - people misinterpreting things on Twitter? I'm shocked. ;)
snarfed joined the channel
# 17:58 tantek hiding the "@username " at the start of replies makes sense in a threaded display context.
indie-visitor joined the channel
bnvk joined the channel
# 18:02 Loqi bnvk: tantek left you a message 15 minutes ago: what "sounds sensible" to you?
# 18:04 bnvk tantek: where did I say "sounds sensible" ?
tilgovi and barnabywalters joined the channel
# 18:08 gRegor` Hahaha to @eviltantek and @silowebcamp
# 18:11 snarfed man. rss/atom opinions are such a great example of how even in a community this small, there are vocal minorities and silent majorities
# 18:11 gRegor` "I think it [the wah wah peddle] had its moment a long time ago. Kinda like Netscape or something." - Moby
# 18:13 gRegor` I don't really get the anti-RSS sentiment.
# 18:14 tantek KartikPrabhu - it's not hate, it's more sadness & disappointment.
# 18:15 snarfed gRegor`: i think i do. plain html wasn't enough on its own, but with microformats, feed readers have all the info they need from a blog's html itself. they shouldn't need a separate feed file.
# 18:15 gRegor` Oh, I know, KartikPrabhu :)
# 18:15 snarfed …which i agree with…but i think we've overstated our case, and people have interpreted it as "we hate rss/atom," as opposed to, "we think mf2 is a good successor"
# 18:15 tantek KartikPrabhu - yeah, it's just a sad situation though.
# 18:16 tantek snarfed - nope, not "people have interpreted" - just a strawman
# 18:16 snarfed tantek: do you agree with that summary? you probably feel the strongest, so i'm curious if i have it right.
# 18:16 bnvk I wonder if it was more that the virality and people centric (avatars, names, etc... that jack into our big social brains) inherent to social media, yet devoid of RSS, that really killed it, not all this speculation
# 18:16 snarfed eh maybe. the language on /feed is (or maybe was) probably stronger than most of us would agree with
# 18:16 tantek snarfed - nah, the real tragedy is the time wasted on the RSS/Atom wars.
# 18:17 tantek bnvk - bingo - RSS was/is being abandoned. We just documented it.
# 18:17 gRegor` Oh, so this is the "RSS is dead" argument?
# 18:18 gRegor` A whole lotta sites out there have RSS feeds, for them being so dead.
kylewm joined the channel
# 18:18 gRegor` KartikPrabhu: Far more than consume mf2
# 18:19 tantek like I said, we're just documenting the demise of RSS, not causing it
# 18:19 gRegor` I'm all for improving (mf2 feeds), but I think the "RSS is dead" is silly
# 18:19 tantek bnvk - you're own RSS feed dying proves the point
# 18:19 bnvk that just broke, I fixed it a few months ago, but yah... the fact that it was just brought to my attention
# 18:19 KartikPrabhu oh no one is claiming that mf2 is the bestest. but it does some of the issues with RSS
# 18:20 gRegor` I didn't think "we" (or anyone) are causing it. I just have seen articles for years about RSS dying, yet it's still here (and haven't seen much evidence of it dying).
# 18:20 tantek how long did it take for someone, *anyone* to mention it?
# 18:20 tantek doesn't that seem to indicate no one is actually consuming it or cares?
# 18:20 tantek so all this "people consuming RSS" rumors are basically overplayed
# 18:20 gRegor` I set up my RSS feed years ago and it's been working fine. *shrug*
# 18:21 tantek similarly I setup my Atom feed years ago and it's been (mostly) working
# 18:21 tantek so someone/something polls it a lot, but who knows who actually cares?
# 18:21 bnvk yes, but it's an interesting flaw perhaps in feed readers that they didn't make the "report broken feed" to owner an easy users project
# 18:21 gRegor` Similarly, every WordPress install has a feed that works out of the box, sans messing it up manually or with a plugin. (latter is hypothetical, I've not seen it personally)
# 18:22 gRegor` I agree on the DRY point, of course
# 18:22 KevinMarks Also that your xhtml embedding blows up feed readers that assume crappy escaping
# 18:22 tantek bnvk - that's what you see as "an interesting flaw … in feed readers" ?!?!?
# 18:22 tantek not the lack of integration with posting tools etc.?
# 18:23 bnvk tantek: you're saying user feedback should be on the onus of the platform doing the publishing detecting the post is broken?
# 18:24 tantek bnvk - I'm saying you're lost in the weeds of a 2000-era mindset. "feeds" "syndication" etc.
# 18:24 snarfed i kinda feel like we're agreeing violently, and the only argument is over degree
# 18:24 snarfed saying *no one* consumes rss/atom is an overstatement, but we'd probably all agree that *fewer and fewer people* consume it
# 18:24 snarfed and "fewer and fewer" still means we should migrate to mf2, etc
# 18:24 tantek the only way to "win" the RSS/Atom war is not to play
# 18:24 tantek hence not worth engaging on this nonsense on Twitter
# 18:25 tantek where the few folks advocating RSS do so, by posting content somewhere WITHOUT an RSS feed.
# 18:25 snarfed right. and those of us who still use and love our feed readers can keep using them, right now, while we work toward replacing rss/atom
jedahan_ and caseorganic joined the channel
# 18:25 tantek snarfed - I have no imminent plans to turn off my Atom feed for example.
# 18:25 snarfed glad to hear it! since i still happily consume it :P
# 18:25 KevinMarks Reading the issues and tests for universal feed parser is a useful corrective to thinking rss works
# 18:26 snarfed but yeah. anyway. we should find some other big piece of indieweb news to try to distract people :P
# 18:26 snarfed (kidding. kidding. don't. nsa, i'm a good guy, i swear.)
# 18:27 bnvk tantek: I don't think i'm lost or stuck in a 2000-era mindset, thank you ;)
# 18:27 KevinMarks I just discovered that Google now can listen to the TV and tell me what is on.
# 18:28 tantek snarfed, Bridgy really helped a lot of people "get" POSSE / backfee
# 18:28 bnvk I realize that, yes, I'm working on an app built on a protocol with a 2.5 billion person adoption ;)
# 18:28 Loqi tantek meant to say: snarfed, Bridgy really helped a lot of people "get" POSSE / backfeed
# 18:28 tantek similarly, integrated reading/writing UIs will help another set of folks "get" it
# 18:29 tantek in terms of why an indieweb site matters, much more than separate feeds/readers
# 18:29 bnvk with that said, I'm gonna give it the ole college try for another year or two, if email is still as dismal I'll agree with ya full :)
# 18:29 tantek bnvk - given how distributed/federated email is, I want to believe it is "just" a UX problem - but it's not. comes back to the difficulty of hacking your own email server etc.
# 18:30 bnvk like I said, give it a year or two ;)
# 18:30 tantek but yeah, if you can solve that problem, you will have achieved something incredible.
# 18:30 tantek (to be clear, I *am* rooting for you to succeed)
# 18:31 aaronpk related to integrated read/write UIs... I had really good luck so far writing a microformats2 adapter into the selfoss reader
# 18:31 aaronpk the next step of course is to put the "write" UI into it using micropub
# 18:32 aaronpk selfoss seems to be the farthest along reader that I could do that with, so I'm tempted to find/make some time for that and dive into it
# 18:32 aaronpk I don't really want to write a whole new reader from scratch
# 18:32 bnvk Selfoss is a dreary lil town outside of Reykajvik with a river running through, and terrible restaurants!
# 18:32 KevinMarks I need to unfuck noterlive.com and get back on the micropub coding
# 18:35 aaronpk that's my fork, I haven't submitted a PR yet cause it's still kind of messsy
# 18:36 bnvk for anyone else? or do the gods hate me
# 18:37 gRegor` Bad Gateway! Go stand in the corner and think about what you've done.
# 18:38 aaronpk not sure what happened, php-fpm stopped responding
# 18:49 KevinMarks I love that the whole of Turkey is discovering dns and VPN tools
# 18:54 benwerd A little odd that it's Google's DNS that's being promoted, mind you.
# 18:55 benwerd Btw, anyone encountered Twister? It's (I'm not joking here) an encrypted decentralized Twitter partially based on the Bitcoin protocol
# 18:55 tantek benwerd, this is how it starts. govts screw up so badly that people being trusting and advocating multinationals instead.
# 18:55 tantek multinationals aren't going to take over, they're going to be handed authority by the people
# 18:55 tantek cue all scifi dystopia scenarios where multinationals run things and natl govts are odd little leftovers.
# 18:56 KevinMarks I heard about it. Bitcoin as pki, bittorrent as distribution?
# 18:56 benwerd tantek: Which isn't a million miles away from the situation we currently enjoy, unfortunately, particularly behind the scenes. Completely agree.
# 18:57 benwerd KevinMarks: Right. Couldn't install it, because of really onerous dependencies. But interesting.
# 18:57 tantek benwerd agreed, point is, this is how it transitions to being in front of the scenes.
# 19:01 tantek KevinMarks, from that medium post, the last image has a poster on a metal wall that says: DNS NASIL DEĞİŞTİRİLİR? which translates to: "How to change DNS?"
# 19:02 tantek and note the subheadings: ANDROID and PC/FIREFOX :)
# 19:03 benwerd I'd love to know who's printing & distributing the posters
# 19:03 tantek Nevermind people "not understanding URLs", people are being taught about DNS!
# 19:04 tantek so can the people adapt faster than the authorities can shutdown/intercept the adaptations?
# 19:11 tantek KevinMarks - I didn't get it - what was the point of that speech?
# 19:14 KevinMarks Jarvis Cocker may be a bit of an acquired taste. I am a big fan of his work, which is oblique and pop and eclectic at the same time. His point about "get on with it" felt indieweb to me
# 19:15 KevinMarks Hmm that liberation link gave me a JS popup claiming I have viruses on my Android. Shady.
paulcp and grantmacken joined the channel
# 19:20 benwerd oh ffs. invited to a private, highly sensitive lung transplant discussion group, with serious data implications.
# 19:22 kylewm snarfed, my site seems to have made a mess of my reply to your reply
# 19:28 gRegor` That huffpost link a bit ago reminded me how much I dislike loading their bloated pages, so I just set up an instapaper bookmarklet: http://pastebin.com/01fgqFFL
chloeweil and paulcp joined the channel
# 19:46 kylewm huh, so is class="p-in-reply-to h-cite" the right way to markup a reply context? and if so, what happens if there is a class="u-url" somewhere in the content of the comment? e.g., http://kylewm.com/reply/2014/03/21/1 snarfed's comment has a u-in-reply-to inside it
_skinny, paulcp, JasonO, eschnou and caseorganic joined the channel
# 20:11 KartikPrabhu kylewm: afaik reply-context markup goes "u-in-reply-to h-cite" then if the h-cite has any "u-url" then that is the url of the reply-context. If it has another "u-in-reply-to" then that is scoped to the h-cite itself
caseorga_ joined the channel
# 20:36 kylewm it might be ok... maybe the worst that can happen is a duplicate u-url to the source comment that
# 20:37 kylewm but I'm just including the e-content object from the other person's comment
# 20:37 KartikPrabhu yeah I still have to do proper reply contexts. Maybe i'll run into these issues then
# 20:42 kylewm in the meantime if you want to reply to me and stick malicious <script> tags in the comment body, you are welcome to ;)
# 20:42 snarfed kylewm: back now. any conclusion on your reply mf2 q's? need anything from me?
tantek joined the channel
# 20:50 kylewm snarfed, I changed my markup a little bit and it updated and looked better. and thanks!
# 20:51 tantek interesting. is that a multireply or a thread?
yaf and KartikPrabhu1 joined the channel
# 20:55 KartikPrabhu1 so multireply is one reply after another, where as thread is many replies to the same thing?
# 20:56 tantek is updating the multireply page right now as we speak for exactly this reason (that there's such a question)
calmsea joined the channel
# 21:05 KartikPrabhu love how idieweb implementations are going beyond just immitating silos
# 21:06 tantek KartikPrabhu: Twitter's display of a chain of replies and/or many replies to the same thing are so similar that it's hard to tell just by looking
# 21:06 tantek KartikPrabhu - indeed! going beyond what silos do is one of the huge benefits of having and creating your own IndieWeb site!
# 21:06 tantek snarfed, earlier re: permashortcitations vs. permashortlinks, I tend to agree with you
# 21:07 tantek as our "original posts" becomes more and more useful than any silo POSSE copy, it will make more and more sense to just use permashortlinks in POSSE copies - since users clicking on those will see more information (e.g. more indieweb/cross-site comments / likes)
ttepasse joined the channel
tilgovi joined the channel
# 21:20 tantek snarfed, do you think it would be helpful to describe permashortcitations as a transitional approach?
# 21:20 tantek I'm trying to figure out a way of capturing some of the consensus coming out of the discussions here.
# 21:21 snarfed or at least, to original post links, which may be permashortlinks (but don't have to be) …?
# 21:21 snarfed i might describe PSCs as a worthwhile experiment which didn't catch on
# 21:21 snarfed since we were doing original post links and PSLs before them
# 21:22 tantek you don't see any value even in the case where the original provides apparent end user benefit over the silo copy?
# 21:23 tantek the "I clicked on a link and saw nothing new" problem
# 21:23 snarfed sorry, trying to parse. i don't see any value in…PSLs over normal permalinks?
# 21:24 tantek in the case where the original has nothing substantial beyond the POSSE copy
# 21:24 snarfed oh. i agree that was the initial motivation, but in practice, i think we heard that PSCs annoyed people just as much as PSLs
# 21:24 snarfed PSCs were annoying because ppl didn't understand them
# 21:26 tantek have you not found any other PSCs via Bridgy?
# 21:26 snarfed eh. we may be overthinking it. there's definitely been usage of all three, but maybe not enough to conclude anything clear yet. that's fine.
# 21:26 snarfed good question! honestly i don't know. hard to data mine for, but i could try.
# 21:28 snarfed personally i only remember seeing PSCs from you, andy sylvester, and kyle
# 21:31 tantek snarfed - do you have twitter accounts for sylvester and kyle?
# 21:31 tantek kylewm, KartikPrabhu, were the "p-in-reply-to h-cite" questions answered?
# 21:32 KartikPrabhu tantek: somehow kylewm seemed to have multiple url props for his reply, which shouldn't happen really I think
# 21:33 tantek wants to compare presentation meaning vs. markup
# 21:34 KartikPrabhu repost: ylewm: huh, so is class="p-in-reply-to h-cite" the right way to markup a reply context? and if so, what happens if there is a class="u-url" somewhere in the content of the comment? e.g., http://kylewm.com/reply/2014/03/21/1 snarfed's comment has a u-in-reply-to inside it
# 21:38 tantek KartikPrabhu: so the question you ask re: u-url has a literal answer and I think an actual answer
# 21:38 tantek literal answer - there is no u-url inside the content of the comment from snarfed
# 21:39 tantek actual answer - if you're going to copy the e-content from someone's comment post to your own site, you need to be careful of all sorts of things people can do (un)intentionally in their markup that could screw up (visitors to) your site.
# 21:39 tantek imagine if he'd put a <script> element inside his e-content - would you happily include that in your reply-context?!?
# 21:39 tantek in fact that might make for a good indiewebify webmention receiving test
# 21:40 KartikPrabhu i think mf2 parsers are supposed to filter those in the value property, but for the html proerpty i think the site author should filter appropriately
# 21:40 snarfed tantek: in the scrollback, kylewm jokingly invited people to do exactly that (send <script>s to him in webmention e-contents)
# 21:40 tantek have indiewebify send you a webmention comment that includes a <div class="e-content"><script>alert("Your server should filter e-content before reserving it")</script></div>
# 21:41 snarfed aaronpk and i and others have discussed this before. obviously you want to sanitize, but there's also clear value in including *some* tags. links, images, etc
# 21:41 tantek it doesn't appear to be sanitizing the e-content from snarfed's site
# 21:41 aaronpk snarfed: I disagree, I would rather re-link everything myself
# 21:41 tantek including looking out for any naked uf2 properties
# 21:41 snarfed no, i didn't mean to imply kylewm is sanitizing,.
# 21:42 snarfed aaronpk: sure, that works too. but it's more than just links. imgs, <code>s, etc
# 21:42 snarfed an easy first implementation is just a simple tag whitelist
# 21:42 aaronpk I want to accept only a plaintext note, and if you have a bunch of formatting in it then it's probably an article and i'm just going to link to it with the article name anyway
# 21:42 tantek I feel like we've documented discussion of this before
# 21:43 snarfed aaronpk: eh i dunno. i definitely see value in some formatting in comments, mainly styles, links, and imgs
# 21:43 snarfed hell, people use them often in comments in silos. pictures and links in twitter/fb replies, bold/italics etc in g+ comments
# 21:43 aaronpk links can be re-linked, I don't really want someone to put an image in my comment thread, the only one that reall can't be reconstructed is <code>
# 21:43 gRegor` I see value in some formatting, too, but sanitizing HTML is a nightmare
# 21:44 snarfed sure. and to be clear, everyone gets to make their own call. but i know i'd allow imgs in comments on my site, and i'm not alone
# 21:44 tantek y'all need to document your existing approaches/implementations
benwerd joined the channel
# 21:44 tantek a bit too much "you need to" handwaving, KevinMarks :P
# 21:44 snarfed KevinMarks: agreed. for those of us who want to allow some tags, i'd suggest we look at existing comment tag whitelists like wordpress's.
# 21:45 aaronpk snarfed: wait til you get a reply from me that contains a 10x2000 px vertical image :)
# 21:45 tantek KevinMarks, depends on what you want to have happen on your site!
# 21:45 snarfed aaronpk: heh. that's the beauty of hand moderation. everything has edge cases, but that doesn't mean you ditch the common case
# 21:45 gRegor` Agree on whitelisting, but even if you just allow <img> I think you can insert javascript that way. I thought I saw that, at least.
# 21:46 gRegor` Granted, my approach is that webmentions won't be displayed until I approve them, or if I whitelist the domain they're coming from (trusted parties)
# 21:47 tantek unless someone has a better / more findable naming suggestion
# 21:48 tantek there's user experience trade-offs that none of those documents mention
# 21:49 tantek I could even see variable sanitizing based on WHO the comment is coming from
# 21:49 tantek e.g. maybe I allow images from trusted friends
# 21:49 tantek this "one sanitizing approach fits all" approach is really shortsighted
# 21:49 tantek rather, plumbing-focused instead of UX-focused
# 21:50 KevinMarks_ my point in posting those two links are that this is a deep hole to explore
# 21:50 tantek which of course is expected from the domains those links are at
# 21:51 aaronpk this is what happens when your ssh session to your IRC terminal is laggy
# 21:51 tantek aaronpk, you weren't blarging at the Secret PESOS?
KartikPrabhu joined the channel
# 21:52 tantek or you could POSSE to Secret by putting the permalink into a QRcode as the bg image
# 21:53 benwerd it's friday afternoon and I was fixing databases until 3am. this is all I got.
# 21:54 tantek anyone reverse-engineered the Secret permalink ID format? e.g. wcoubyckjckgvstnwkkrcmrbbr
# 21:55 gRegor` So shouldn't be an issue, presuming you don't display an iframe when showing comments from other sites. :)
# 21:55 tantek or just ping the internet archive with all secret permalinks
# 21:57 KevinMarks_ Didn't Brad Fitzpatrick make his FB profile picture a QR code of his own domain?
# 21:58 gRegor` To identify users, tantek?
# 21:58 tantek gRegor - nah, just to get a raw dump of everything
# 21:58 tantek KevinMarks - yeah that was his rel-me hack back in the day
# 22:01 aaronpk bret: YAY caseorga_ would like to cut something with it!
# 22:07 gRegor` I was reading about how it works the other day. Interesting.
# 22:07 bret aaronpk: rad! it runs with some funky software but we can probrably figure it out
# 22:08 tantek you need to self-heart to share your own with all your friends
# 22:08 benwerd aaronpk: I'm pretty certain it sometimes labels friends as being generic area posts, as friends of friends, etc
# 22:08 tantek it only sends a subset of the secrets from those you follow
# 22:08 aaronpk benwerd: I think if someone has fewer than 8 contacts then it intentionally shows up as area posts
# 22:09 tantek it sends you ALL of the secrets that your friends heart
# 22:10 KevinMarks_ except 'friends' is iOS addressbook, which for me is pretty random as I haven't used Apple addressbook since 2007
# 22:10 aaronpk heh, i have the opposite problem. i let facebook sync with my ios addressbook so all my facebook friends are potentially secret friends
# 22:10 benwerd it's going to be pretty fun when they port to android and suck up everyone's google contacts list
# 22:14 KevinMarks_ this numberwang version needs to say "thats wangernum and rotate the board"
benprew joined the channel
# 22:17 gRegor` On the wiki guest list, is "Missed You" for people that said they were coming but didn't show up? Otherwise I'm not sure the difference between "Regrets" and "Missed You" http://indiewebcamp.com/2014/Guest_List
# 22:17 snarfed gRegor`: i think people put themselves in regrets, and put others in missed you
# 22:22 bret how do people parse h-entry without losing html links?
# 22:22 bret do you sanitise the html? or re-link the value?
# 22:22 snarfed bret: heh. we had a big discussion about that just an hr or so ago. check logs.
# 22:24 bret html sanitizing.... thats a thing I could probrably find a library for
paulcp joined the channel
# 22:29 bret hey aaronpk, why do the logs have a bunch of 'U+0002' in them?
# 22:29 snarfed thanks for starting the stub, tantek! updating now
# 22:30 bret its pissing off these feed readers >:[ now, to figure out how to kill it
# 22:32 KevinMarks_ I remember Google reader had a stupid feed sanitzer that would break blockquotes
# 22:33 bret this will be good experience to build an h-feed to atom proxy next
# 22:34 bret then I can experiment with publishing h-feed only, with atom provided by the proxy
# 22:35 gRegor` Any other wiki template requests? I'm on a roll. :)
# 22:35 tantek yes, in the Guest Lists, "Regrets" is for when people explicitly cancel (e.g. even if they just message this channel saying they're sorry they can't go
# 22:35 tantek It's a polite way of keeping track of people who were too rude/inconsiderate to explicitly cancel
# 22:36 tantek bret - subtle shame box - for those in the know
# 22:36 tantek so that there was no obvious shame in anyone's name coming up in a Google result
# 22:36 tantek but it is helpful to keep track for the community
# 22:36 tantek so that we learn and can encourage people to behave better in the future (e.g. in in-person conversations)
# 22:38 kylewm sorry I missed the conversation about sanitizing, but it answered all my questions except maybe one -- would you recommend just always removing class="..." /style="..."
# 22:38 tantek ok I think I've caught up with wikifying stubs for today's conversations
# 22:39 tantek another alternative: drop "class", but keep "style" from "trusted" commenters
# 22:39 tantek just as "plaintext" removes plenty of nice styling
# 22:40 tantek KevinMarks, I think someone has to be brave and attempt a clever way to keep class names without risking the page content before we can recommend anything like that.
# 22:41 tantek if you get comment embedding working on your site with class names included, that could be a good starting point for analysis.
# 22:41 kylewm I suppose it would be nice to keep microformats, if you can sandbox them within the comment body?
# 22:41 kylewm the problem I had is they were leaking into the h-cite they were part of
# 22:41 tantek the problem is with naked uf property class names in the e-content that are outside of any embedded h-*
# 22:41 tantek right, that get "hoisted" (as KevinMarks would say) up to containing h-*, e.g. h-cite
# 22:47 tantek Kevinmarks - you do realize an "h-quoted" inside an "h-cite" seems silly right?
benprew joined the channel
# 22:53 tantek KevinMarks, kylewm it sounds like we're in new(ish) territory here
# 22:54 tantek far beyond TCMOS or any other similar recommendations (which all only handle plain text embedded citations)
ttepasse joined the channel
pfenwick joined the channel
# 23:03 kylewm tantek, it's really interesting stuff... it seems a shame to lose the metadata
# 23:03 kylewm though i can't think of a use case for keeping it right now
caseorganic joined the channel
# 23:03 snarfed is composing a reply to benwerd, and putting a URL directly in the contents, just because he knows benwerd strips html tags but auto-linkifies URLs
# 23:04 benwerd snarfed: I need to switch to something more intelligent
# 23:04 snarfed benwerd: eh. maybe, eventually, if it itches enough. ruthlessly prioritize, i say!
caseorganic joined the channel
# 23:09 tantek benwerd, or at least capture your itches so you can decide later whether they really itch that much or not.
# 23:10 tantek also interesting to the rest of us to see what's itching you ;)
# 23:10 tantek needs to go finish a blog post on URL experiments
basal joined the channel
# 23:22 tantek it's great to work with so many builders and creators. thank you all for being here.
# 23:23 gRegor` And some hopefully helpful links to the templates / prompts in HTML comments when adding yourself to an empty list.
# 23:25 snarfed welcome tantek. great to have such a good community organizer too. thanks for rallying us all!
bnvk joined the channel
# 23:29 bret thanks t, and thanks for your hard work as well :)
# 23:34 gRegor` Haha, that would be funny.
# 23:35 bret its cool to see his taking positie stance though
# 23:35 gRegor` (Yeah, I didn't think anyone actually would)
# 23:35 gRegor` So I'm not in the know, is this Dave/IWC an ongoing thing, or did he just mention IWC today kind of out of the blue?
# 23:35 bret he noticed rss criticism on the feed page
# 23:37 bret debate with dave doesnt seem terribly productive, but it would be really cool to get him to at least understand waht we are trying to do here
# 23:37 benwerd Complicated by a bunch of our (justified) criticisms towards an unrelated piece he wrote the other day, but I think he'd be an interesting perspective.
# 23:39 bret benwerd: that was tough to read/watch
# 23:39 kylewm benwerd, he said something to you like "I don't believe any of things you think I believe" did he expound on that at all?
caseorganic joined the channel
# 23:40 KevinMarks_ He said that in context he was saying it was a risk, but one you should take. It didn't read that way
# 23:40 benwerd It really didn't, and that was enough to make it concerning
# 23:40 bret but a totally rss feed hardened perspective would be nice, in direct context to all the stuff happening here
# 23:44 gRegor` I am working on something similar, but I think I'm going to show the text of replies, so it appears more like the "Notifications" tab on Twitter
KartikPrabhu joined the channel
tantek joined the channel