2016-11-21 UTC
tantek, michel_v, XgF, renoirb, MOLONLABE2016, begriffs and mlncn joined the channel
tantek, dkm, userXYZ, wolftune, sbruce, KevinMarks, begriffs, KevinMarks_, snarfed and rpcope joined the channel
wolftune, cweiske and sl007 joined the channel
pindonga, voxpelli and jacus joined the channel
# 09:24 petermolnar !tell aaronpk I left a comment for that guy, dunno if he proceeded
# 09:24 Loqi Ok, I'll tell them that when I see them next
# 09:26 cweiske ben_thatmustbeme, how is your pubsub hub implementation going?
# 09:27 ben_thatmustbeme Not bad. Got a bit distracted and started working on inkstone again. I have a very basic framework right now
# 09:33 Zegnat I am available, one of my colleagues would like to tag along, and if sebsel comes a long that makes 3 dutchies.
pindonga, DanC_, Pierre-O, ricardokirkner and [sebsel] joined the channel
plindner, adactio, jcgregorio[m], rdesfo[m], M-hotzeplotz, M-rongladney, M-nd, TheGillies, M-mxuribe, Guest60189[m], 7F1AA0SO0, 14WAAK33R and Jeffrey[m] joined the channel
# 10:49 Zegnat I think cmal and jboy might be interested too, though not sure how to get their attention right now
# 10:50 Zegnat !tell cmal,jboy We are looking at a possible HWC on 2016-11-30 somewhere at Schiphol. Interested?
# 10:50 Loqi Ok, I'll tell them that when I see them next
# 10:50 Zegnat (Did multi tells work now? sknebel? aaronpk?)
# 10:59 Zegnat !tell yobj We are looking at a possible HWC on 2016-11-30 somewhere at Schiphol. Interested?
# 10:59 Loqi Ok, I'll tell them that when I see them next
nitot_, mlncn, hs0ucy, nitot, sl007, tantek, KartikPrabhu, lmorchard, wolftune and adactio joined the channel
John_Duh, snarfed, wolftune, John_Duh_, nitot, hs0ucy and KevinMarks joined the channel
nitot, snarfed, snarfed1, hs0ucy and wolftune joined the channel
wolftune joined the channel
# 16:36 sknebel hm, I saw some complaint about their registration process a few days back
wolftune, squeakytoy, jmelesky, tantek, KevinMarks, gRegorLove, KartikPrabhu and nikivi joined the channel
# 18:09 Loqi aaronpk: petermolnar left you a message 8 hours, 45 minutes ago: I left a comment for that guy, dunno if he proceeded
# 18:10 snarfed oh wow true. surprised he didn't have any monitoring.
tantek joined the channel
begriffs and raretrack joined the channel
# 18:34 tantek snarfed, do you have any suggestions for monitoring? e.g. what do you use for monitoring your site(s)?
benwerd joined the channel
# 18:48 snarfed monitoring is automated collection of a site or app's vital stats, e.g. is it up or down, how much traffic is it getting, what's its average latency, etc.
# 18:49 snarfed monitoring is automated collection of a site or app's vital stats, e.g. is it up or down, how much traffic is it getting, what's its average latency, etc.
hs0ucy joined the channel
# 18:59 KevinMarks People posting svg's with js in to svgur.com and mailing me like they solved a problem rather than causing one.
begriffs, wolftune and awolf joined the channel
snarfed1 and tantek joined the channel
# 19:26 Loqi rhiaro has 58 karma in this channel (248 overall)
begriffs, Pierre-O, cweiske and awolf joined the channel
# 19:49 ben_thatmustbeme KartikPrabhu: it looks like one of them is within that post type, the other is prev by all types
# 19:49 rhiaro cos I log a lot of stuff, it lets you skip back through articles if you don't want that to be interrupted by all my food posts
# 19:50 rhiaro Many things about my site are not obvious :P Are the icons loading for you?
# 19:50 ben_thatmustbeme it wasn't obvious to me until i saw the icon changed when i went back one and got a food post
# 19:50 KartikPrabhu rhiaro: yeah I can see the icon on the second "Prev" link, but couldn't immediately guess what it meat
mlncn joined the channel
# 19:51 rhiaro I need to do something about the icons anyway, they're pretty screwed up cross devices
Baltimoron and awolf joined the channel
# 20:02 cweiske pubsub: what prevents a third party from sending a forged "you have been unsubscribed" ("hub.mode=denied") request to my client?
gavinc joined the channel
# 20:06 rhiaro breaking news cweiske: pubsub is called websub now. Also I think there's a verification step for unsubscriptions, too
# 20:09 aaronpk might want to wait a coupel days til we publish the next draft
# 20:11 cweiske I've never seen this capability URL document before
# 20:11 Loqi [Jeni Tennison] Good Practices for Capability URLs
renoirb joined the channel
# 20:12 cweiske "3.1.3 Websites without Accounts" is an interesting idea
benwerd, snarfed, KevinMarks and wolftune joined the channel
# 20:45 cweiske now that not even "hubbub" is part of the name anymore, the double b is also out
# 20:49 cweiske the "ph" in the names is also for PHP, so changing the first letter would not work
[sebsel] joined the channel
pfefferl_, KartikPrabhu, snarfed1, benwerd_ and hs0ucy joined the channel
# 21:13 cweiske websub: is it valid to have the callback URL contain a "hub.mode=notification" parameter? that way I'd always have that parameter, and the hub only overwrites it value for "subscribe", "unsubscribe" and "denied"
# 21:14 cweiske "Existing parameters with names that overlap with those used by verification requests will not be overwritten"
gRegorLove joined the channel
# 21:17 aaronpk you can use the GET vs POST to check whether it's a notification or verification
Pierre-O joined the channel
# 21:25 aaronpk i thought you already had a working hub? are you building a new one?
# 21:27 cweiske whenever you implement push for the chat log pages :)
# 21:27 cweiske and apart from it it means I don't have to manually tell search.cweiske.de my new blog posts
tvn joined the channel
# 21:30 cweiske I just fear that this will increase traffic quite a lot
# 21:32 snarfed1 yeah you definitely want to decouple recrawling from websub pings :P
# 21:32 snarfed1 think "nudge," not "trigger"
pfefferl_ and tantek joined the channel
# 21:34 cweiske this would make a good use case for diffed fat pings
# 21:36 KevinMarks The chat is a list of things with timestamps, maps to a feed pretty well
# 21:36 tantek rhiaro: re: titles on links - does not help for touch devices
# 21:37 cweiske KevinMarks, yes, but that's not the issue. when the search engine gets pinged that the "today" page has been updated, and there are 1000 chat line links on it, all of them have to be crawled
# 21:38 tantek cweiske++ for the ph* naming convention - clever!
# 21:38 Loqi cweiske has 58 karma in this channel (66 overall)
# 21:38 cweiske so best would be if the hub would send only the new content in the fat ping
# 21:39 tantek finishes catches up on the logs, goes back to eating his salad
# 21:40 cweiske yes, which is why fat pings are hard to implement on the hub side
# 21:41 KevinMarks Atom wasn't a great universal api, but it did handle lists with timestamps
wolftune joined the channel
# 21:45 Loqi atom has 1 karma in this channel (2 overall)
# 21:46 tantek except for that whole requiring artificial precision of seconds thing
# 21:48 cweiske no. "3.3 Date Constructs" in the atom spec allows normal second resolution in datetimes
# 21:51 tantek no. requiring seconds precision when it is unavailable is wrong
# 21:51 tantek it is always wrong to mandate precision in a format
# 21:52 tantek "as accurate as possible" yes - but when it is not possible or not available, then it is HARMFUL to pretend to have accuracy you do not
# 21:52 tantek unfortunately Atom makes that harmful mistake by *requiring* seconds precision instead of being content with just minutes
# 21:57 cweiske please give an example when having "00" as seconds is harmful
# 21:59 tantek when you don't know what the exact seconds are
# 21:59 tantek e.g. most of my blog posts before IWC Brighton
benwerd joined the channel
# 22:01 cweiske wrong/false is not necessarily harmful. which is why I asked for an example when it's harmful
mlncn joined the channel
# 22:03 cweiske harmful is when the plane crashes into a building because the seconds were "00" instead of "12", but I fail to construct such a connection
# 22:05 tantek in general it's a bad policy to publish things as accurate/true when they are not
# 22:05 tantek the harms may be small in particular instances, but they are still there, thus specs MUST avoid making such artificial precision requirements
# 22:06 tantek cweiske: more here if you're really interested wikipedia false precision
# 22:06 tantek KartikPrabhu: indeed. reductio-ad-plane-crashum
# 22:06 KevinMarks Irc would be an interesting case, given out of order arrival and clock skew
# 22:10 cweiske the wikipedia page and all references pages do not talk about timing precision
pfefferl_, LanaCoyote, snarfed, benwerd and wolftune joined the channel
# 23:06 cweiske the moment you realize that half of our work is already done .. in a git branch you already forgot about
# 23:07 cweiske and rediscovered just before committing the new code
# 23:13 tantek how close was the second implementation to the first?
# 23:15 cweiske different. I then used php arrays for config, now I use the database
# 23:15 cweiske but link extraction is already done in the other branch
aegibsonme joined the channel