2020-03-12 UTC
# 00:01 [Michael_Beckwit i’m at a coffee shop as I type this, and it’s 7pm
ttal, beko, [tantek], nickodd, gRegorLove, geoffo, [Michael_Beckwit, [jeremycherfas], DanQ, [Jeff_Hawkins], cweiske, leg, [g33kcentric], KartikPrabhu, omz13, [LewisCowles] and jceb[m] joined the channel; nickodd left the channel
# 11:04 GWG Zegnat: Re distraction, I added more Parse This code. YouTube doesn't link to their RSS feeds so I added a function to generate them from a YouTube URL.
# 11:06 GWG Also parsing tv and movies from IMDb so I could cite them if I posted I am watching something and going to do music from discogs
[jgmac1106] joined the channel
# 11:13 [jgmac1106] I do the "google local guide" as it is actually turning into tangible benefits as businesses have started to seek out my reviews
[grantcodes], misterwhatever, [jeremycherfas], [calumryan] and [KevinMarks] joined the channel
# 12:04 Zegnat The YouTube RSS feeds discovery is on the indieweb wiki too, if you want to check if you got them all
KartikPrabhu joined the channel
# 12:34 GWG I don't like schema, but it is certainly better than nothing
# 12:37 GWG And why YouTube can't put in a rel alternate like everyone else...
# 12:41 GWG I also did something experimental with my mf2 return
# 12:42 GWG JSON-LD on many sites has a publisher property that is an organization object that I'm mapping to a p-publication h-card so I can get the organization URL and logo in my return
hs0ucy, loicm, KartikPrabhu, plut4rch, [grantcodes] and [Michael_Beckwit joined the channel
# 15:51 jacky granted, we'd lean on URLs and one's preferred name but this signaling helps if I wanted to continue to reply to one of the mentioned parties
[yo] joined the channel
# 15:58 [yo] it’s all about the reader experience!
[AlisonW], [email096], [jeremycherfas], gRegorLove, [tantek], [datashaman], [chrisaldrich], [LewisCowles] and [schmarty] joined the channel
# 18:11 jacky tests++ for finding out why I had to double import webmentions to get them to appear on my site
# 18:11 Loqi tests has 6 karma in this channel over the last year (8 in all channels)
# 18:11 jacky I'm tempted to benchmark things, which don't matter for personal stuff but like knowing me; once I can get a client that rivals the performance/quality of things like Twitter's mobile client; speed will be important to me
[davidmead], nickodd, geoffo, [Michael_Beckwit, [AlisonW], [jgmac1106], [KevinMarks] and [Rose] joined the channel
# 19:34 jacky a lot of people use it (haphazardly tho, with no explicit documentation - if you know, you know)
# 19:34 jacky I do expose `.well-known/nodeinfo` (nodeinfo) on my site
# 19:36 jacky most of the stuff in host-meta can be expressed in HTML
# 19:42 jacky I guess the reasoning for putting it elsewhere is to not deal with 'parsing HTML' which isn't _that_ much of a burden - it's like shipped with any computer
# 19:43 sknebel some of it also makes sense to be independently maintained
# 19:44 sknebel no need to patch your off-the-shelf software that shows your front page to include it
# 19:45 jacky wouldn't have to patch it if the software is extensible ;)
# 19:45 jacky or had a flag for enabling this bits of exposure
geoffo joined the channel; nickodd left the channel
# 19:53 [AlisonW] but so long as you fill in all the options programmatically there's no actual _harm_ in repeating the info in different places if it makes some systems easier to implement
geoffo and [yo] joined the channel
# 20:21 [tantek] kinda really dislike the "xyz.txt" thing, almost enough to label it an anti-pattern
# 20:22 [tantek] though it's mostly harmless so I haven't bothered really writing anything up about it. just distracts from actual useful standards efforts
# 20:23 [tantek] I mean come on, how can anyone hate on those cute graphics?
# 20:23 wink but they messed it up with / and not .well-known
# 20:23 wink they came late enough to actually been able to think about it
# 20:23 [tantek] nah, they identified that .well-known is crap for usability
# 20:23 wink sure, but who reads those humans.txt?
# 20:24 wink nerds, who don't care for usability
# 20:24 [tantek] nah, nerds care for usability, just often in different/weird ways, e.g. unix CLIs
[Ana_Rodrigues] joined the channel
# 20:25 wink the nerd barrier in this context is as low as understanding urls
# 20:25 [tantek] yup, excellent proof that "a few extra chars" matters
# 20:25 wink as in you wouldn't have checked it at google?
# 20:26 wink because of the typing?
# 20:27 [tantek] the ultimate evidence of that is all the short unix CLI commands
# 20:27 wink no, that's backwards logic
# 20:27 wink all the good ones were taken very early
# 20:27 wink of course they were short.
# 20:29 [tantek] hmm, I'd say that reasoning is faulty because of the example of Twitter, where people *ignored* short usernames for ~2 years
[datashaman] joined the channel
# 20:31 wink I'm pretty sure it was disallowed at the start
# 20:31 wink I think I remember when @h got his
# 20:32 wink also which evidence?
# 20:33 wink ok, you could've linked that earlier
[LewisCowles] joined the channel
# 20:33 [tantek] why dislike them for emotional reasons when you can use usability to debunk them fully?
# 20:34 wink you can also dislike that fqdns are backwards for users while they are are actually the other way round
# 20:34 wink I've never seen any studies if it had made more sense to have the hierachy with the root at the start
# 20:35 [LewisCowles] Java is a case-study in why packages with backwards domain names are stupid
# 20:36 wink ok, what came first? co.uk or the DNS? :)
# 20:36 wink so maybe introducing co.uk was the real culprit
# 20:37 wink and yes, I'm not defending Java here, they came later. but the order is actually correct
# 20:37 wink so you're disliking it for emotional reasons? :P
# 20:39 [LewisCowles] TBH I think tantek.com is better for the reason non-english ordering subject, detail makes more sense
# 20:41 wink yeah, least specific to most specific
# 20:42 wink I find that very much less specific
# 20:42 [LewisCowles] so if we managed to work out further details, like solar ages, we could pre-pend the year and have ordering work
# 21:07 wink if completely ignoring everyone who changed handles or left twitter or uses digits, I see 60% of 1 and 2 letter handles being taken by accounts created i nthe first 2 years
# 21:09 wink not saying they were all gone instantly but "ignored" sounds different if we assume that not all of them waited for 2 years but followed your example in the first year
# 21:23 [tantek] there were maybe 3 actual grabs until 2008 sometime, at which point a spammer/squatter grabbed the rest
# 21:24 [tantek] then those were released one at a time to most of the people that have them today
# 21:24 aaronpk the date the account shows it was created is the date of the original account, not the date of the last rename
# 21:26 wink I said that I'm willing to ignore that because I'm not starting to crawl archive.org now
# 21:26 wink or maybe I was just unluicky that my initials as a 2letter combo were gone a lot sooner. I can't imagine I didn't try it
KartikPrabhu, [manton], [schmarty], [datashaman]1, [Ana_Rodrigues], [LewisCowles]1, [LewisCowles], tsrt^ and Pikseladam joined the channel
# 23:10 [tantek] 2 letter combos were grabbed much sooner than most of the 1 letters