2015-05-12 UTC
# 00:06 kylewm unfortunately I'm not logging enough to see why Tumblr isn't quite working for you
# 00:07 kylewm it worked for me and returned a location and everything
# 00:09 kylewm which works but returns a redirect instead of a 201 that quill expects
# 00:10 kylewm I added some code to force heroku to always serve from https
# 00:10 bengo @kylewm What was it? I can confirm if it's horrid :P
# 00:11 bengo did that recently too for indieweb-example.herokuapp.com
# 00:11 kylewm turns out it was "from flask_sslify import SSLify"
# 00:12 bengo if the incoming protocol isn't https, redirect to HTTPS
# 00:12 bengo optionally also checking the X-Forwarded-Proto header
# 00:12 bengo which is what heroku passes along if it terminates the SSL and sends you non-ssl
# 00:13 bengo well... it will, but it will follow it with a GET
# 00:13 bengo not a POST to the Location header of your response
# 00:14 KevinMarks__ so when getting the micropub url, follow redirects and cache the redirected one?
# 00:16 bengo The server should probably be using a 301 or 307 instead
# 00:17 bengo "Many web browsers implemented this code in a manner that violated this standard, changing the request type of the new request to GET, regardless of the type employed in the original request (e.g. POST).[1] For this reason, HTTP/1.1 (RFC 2616) added the new status codes 303 and 307 to disambiguate between the two behaviours, with 303 mandating the change of request type to GET, and 307 preserving the request type as
# 00:18 KevinMarks__ my point is that if my tumblr points to the http version, quill will fetch that, get a 302 and then not cache the https one
# 00:18 kylewm well it looks like SSLify does support an option "permanent" to use 301 instead of 302
# 00:18 kylewm but i kinda doubt Quill would do anything special with redirects of either kind?
# 00:19 bengo 308 is probably the most explicit thing... It's means permanent and explicitly-repeat-POST
yakker and Jihaisse1 joined the channel
# 00:19 kylewm Tumblr and Blogspot both supported data:images ??!
wolftune joined the channel
KevinMarks___ joined the channel
# 00:31 KevinMarks___ I suppose if I give an image too big for local storage could mung it client side in canvas
KevinMarks_ joined the channel
# 00:37 Vendan hrm, for the indie messaging stuff, wouldn't it make sense to check to see if the target has a u-key for public/private key encryption, and encrypt the message with that key?
bengo joined the channel
# 00:40 kylewm but do you really want your private key on your webserver?
# 00:50 KevinMarks_ But I could set something up to fan out to tumblr, WordPress and blogger
# 00:51 acegiak ah ok, neat. I'm interested in how people do their tumblr posseing
mlncn joined the channel
# 00:52 Vendan heh, I may posse notes to twitter, but I'm not really on any other social network
snarfed, bengo, wolftune, KevinMarks_, KevinMarks___ and KartikPrabhu joined the channel
# 02:23 acegiak Kevinmarks, not really because I'm pretty happy with my POSSE setup as it is
# 02:23 acegiak but I'm keeping an eye out for any solutions that can do what I'm already doing with my hacky code but properly
KevinMarks_, KevinMarks2, elima_ and bengo joined the channel
# 03:04 Vendan well, I've managed to start posting notes that are marked up with h-entry, the main page has a h-feed, and I've got a h-card set up with a smattering of info
tvn joined the channel
# 03:17 KevinMarks__ what kyle built with feverdream points towards a "micropub as POSSE" path
# 03:22 KevinMarks__ right, but you can send them with indiewebify.me to check your markup works
Unifex and frzn joined the channel
# 03:24 Vendan markup isn't working for them, actually, there is no markup for them really
# 03:24 Vendan can't currently put a arbitrary html in, and there's no parsing to turn urls into links yet
# 03:25 Vendan golang's template system is rather hardcore about disallowing XSS and similar
# 03:26 Vendan it's kinda crazy, it parses the html in the template to the point where it knows if it's in an html attribute or a text node or what, and automatically escapes the inserted content
# 03:31 KevinMarks__ '(?:\\@[_a-zA-Z0-9]{1,17})|(?:(?:(?:(?:http|https|irc)?:\\/\\/(?:(?:[!$&-.0-9;=?A-Z_a-z]|(?:\\%[a-fA-F0-9]{2}))+(?:\\:(?:[!$&-.0-9;=?A-Z_a-z]|(?:\\%[a-fA-F0-9]{2}
))+)?\\@)?)?(?:(?:(?:[a-zA-Z0-9][-a-zA-Z0-9]*\\.)+(?:(?:aero|arpa|asia|a[cdefgilmnoqrstuwxz])|(?:biz|b[abdefghijmnorstvwyz])|(?:cat|com|coop|c[acdfghiklmnoruvxyz])|d[ejkmoz]|(?:edu|e[cegrstu])|f[ijkmor]|(?:gov|g[abdefghilmnpqrstuwy])|h[kmnrtu]|(?:info|int|i[delmno
# 03:34 KevinMarks__ i'm not sure quite how tantek came up with that, but we use it and stand well back
snarfed joined the channel
# 03:37 Vendan yeah, I looked at the source of it, what you pasted got cut short
# 03:38 Vendan not sure about the concept of doing that hardcore of validation in regex
# 03:39 Vendan I'd feel better getting a rough handle on where the url is, letting golang's stdlib parse it for me, then do some further validation if I want to throw out stuff like .mobi or .jobs
KevinMarks_ joined the channel
# 03:47 acegiak KevinMarks_: I hadn't thought of using micropub for posse!
gRegorLove and bengo joined the channel
# 04:17 kylewm cassis does actually do some validation and post-processing outside of the regex
# 04:17 kylewm for example it removes trailing )'s unless there is a ( in the path part of the url
wolftune, arlen, snarfed, loic_m, glennjones, KevinMarks___, KartikPrabhu, KevinMarks_, KevinMarks2, csarven, killerog, bengo, fkooman and elima_ joined the channel
# 06:48 raucao figured out why bridgy stopped working with my known
# 06:48 raucao apparently it relies on links within the tweet to send webmentions to
elf-pavlik joined the channel
# 06:49 raucao i thought the twitter plugin would somehow announce what it publishes so that bridgy can keep the url with the tweet id or sth
# 06:50 raucao would hate to re-add post links to every tweet, as it means less characters
# 06:50 elf-pavlik yeah, i missed the camp as well :(
# 06:50 raucao i meant we missed you at the camp, not that i missed it :)
# 06:50 elf-pavlik i understood
# 06:51 elf-pavlik i missed both, camp and you :(
# 06:51 elf-pavlik how did it go?
# 06:51 elf-pavlik in Braunschweig, will come to Berlin soon
# 06:51 elf-pavlik hey rhiaro :D
# 06:54 elf-pavlik also teach and learn, since one also learns by teaching
# 06:54 elf-pavlik needs to redraw it...
# 06:54 KartikPrabhu elf-pavlik: i don't think the digital world accomodates all physical world activities yet
# 06:55 rhiaro KartikPrabhu: people still post about them, but what level of modeling for online is hard
# 06:56 rhiaro elf-pavlik: maybe this conversation should be in #social :)
# 06:56 KartikPrabhu I have found it easier to replicate existing digital world social behavior on my site, rather than worry about the most general possible digital expression
# 06:59 elf-pavlik "add detailed logs of my usage of public transport, I take public transport without buying tickets and would like that everyone can fully analyse my consumption habits in this field"
# 07:01 elf-pavlik does it matter? if i don't start publishing it we will have chickent and egg issue
# 07:01 elf-pavlik public transport agencies i hope
# 07:01 elf-pavlik KartikPrabhu, i don't focus on 'blogger' centric requirements
# 07:01 elf-pavlik i focus on *social* and also *economic*
# 07:01 KartikPrabhu hmmm personally I'm not interested in "agencies" or "organisations" consuming what I publish. More about my friends and so on
# 07:02 elf-pavlik well, i hope i will make friends also with people who run public transportation
# 07:02 KartikPrabhu I focus on "what I want to publish" so that my friends "might wan to see"
KevinMarks_ joined the channel
# 07:03 elf-pavlik similar here, many people ask "how do you manage to leave without using money", by publishing detalied logs of my consumptions / contributions everyone can take a look and answer this question to oneself :)
# 07:05 elf-pavlik please note #ownyourdata element to it :)
# 07:05 KartikPrabhu in that respect, it would be good for "everyone" if there existed tools to interpret this data. Pure data is mostly useless unless you know how to extract meaning from it
# 07:05 elf-pavlik i agree, but you often need raw data to build tools for interpreting it
# 07:06 elf-pavlik s/you often need/one often needs/
# 07:06 Loqi elf-pavlik meant to say: i agree, but one often needs raw data to build tools for interpreting it
# 07:06 KartikPrabhu true. but unless those tools are available the data is mostly useless
# 07:06 elf-pavlik again, we need to deal with chicken and egg issue all the time
# 07:06 KartikPrabhu speaks from experience listening to experimental science talks about data :P
KevinMarks___ joined the channel
# 07:07 KartikPrabhu elf-pavlik: lest I come off as dismissive, I do realise that having data is the first step to using it :)
killerog joined the channel
# 07:10 KartikPrabhu rhiaro: btw good post testing aaronpk's fancy new article posting tool
KevinMarks_, friedcell, KevinMarks___, killerog, nloadholtes and evalica joined the channel
LukasRos and petermolnar joined the channel
# 08:18 elf-pavlik ahoy aaronpk o/
interactivist joined the channel
# 08:29 elf-pavlik i don't understand why h-cite would have to do with display?
# 08:31 elf-pavlik for me h-cite suggests: "In this context consider this data as embedded/trascluded from elsewhere"
# 08:31 rhiaro elf-pavlik: the wm spec doesn't have anything to do with what people do with replies when they get them, including how to mark up
# 08:31 elf-pavlik "ceivers SHOULD moderate Webmentions, and if a link is displayed back to the source, SHOULD link to source with rel="nofollow" to prevent spam."
# 08:32 elf-pavlik "4. Alice's publishing software verifies that Bob's post actually contains a link to her post and then includes this information on her site."
# 08:33 elf-pavlik of course i don't say that it should put normative requirement on using h-cite or microformats at all :)
Sebastien-L, KevinMarks_ and loic_m joined the channel
# 08:36 aaronpk elf-pavlik: that's there because it was in pingback
# 08:40 elf-pavlik aaronpk, i guess by 'display' you meant including in page without refering to how it appears to person looking at it?
# 08:41 aaronpk i mean anything that isn't just verifying the webmention and storing it
# 08:42 elf-pavlik i agree that it fits the same bucket as rel="nofollow", i also suggested possibility of adding it in the same section
# 08:43 aaronpk which is what i thought i said on the github thread
# 08:43 elf-pavlik yes, i misunderstood the 'display' part
# 08:46 elf-pavlik replied on gh, thx!
Sebastien-L, fkooman, stream7, adactio, elf-pavlik_, stream7_, j12t, Pierre-O, elf-pavlik and nloadholtes joined the channel
j12t joined the channel
elima_, j12t, frzn, friedcell and mlncn joined the channel
# 12:26 GWG Okay, photo posting support in my webactions plugin, check. I'm now ready to go picture taking.
Erkan_Yilmaz, Jihaisse and mlncn joined the channel
j12t joined the channel
# 13:35 kylewm !tell raucao you don't have to include links in you twitter copies for bridgy to work. See /posse-post-discovery. Let me know if you want help debugging bridgy
# 13:35 Loqi Ok, I'll tell them that when I see them next
# 13:35 raucao kylewm: yer, i found the faq entry about links
# 13:37 raucao wanted to first try and add the feed link to the site that was previously linked in my twitter profile. do you know if it checks every time? if i had written it, i'd probably cache things
fourtonfish joined the channel
# 13:40 GWG zachdonovan: We have at least 50% more attendance than last time
# 13:41 kylewm raucao:yes it caches the url in your profile, if you change the url and reauth Twitter, it will pick up the new one
# 13:44 GWG zachdonovan: Tantek is very charismatic. I imagine that others will follow.
# 13:45 GWG zachdonovan: I am arranging to beat the traffic this time
# 13:46 GWG zachdonovan: It was smooth going home that night
snarfed joined the channel
j12t, Erkan_Yilmaz, chalettu, KevinMarks_, KevinMarks___ and friedcell joined the channel
# 15:21 kylewm I added a "N New Posts" button that appears when new entries are available but hidden, at the top of reader.kylewm.com ... was fed up with it scrolling away from me while i'm reading a long article
KevinMarks_, KevinMarks___ and killerog joined the channel
# 15:38 ben_thatmustbeme kylewm: what do you think about being able to discover multiple feeds when adding a person?
# 15:39 ben_thatmustbeme search the page for h-feed but also for links with rel='feed' title='photos only feed' etc
# 15:40 kylewm ben_thatmustbeme: makes sense to me, does anyone publish feeds like that?
j12t joined the channel
# 15:52 Vendan microformats parser can read the rel links easily anyways, just sift through the feed key of the rels portion. Only issue is you lose all the extra context of the link
# 15:52 kylewm ben_thatmustbeme: is there a good way to know whether rel="feed" will be text/html and not RSS?
# 15:53 kylewm Vendan: check out the new rel-urls feature on mf2py :)
# 15:54 Vendan k, I'll see about adding that to my parser as well
# 15:56 Vendan heh, so you can sift through rss and atom feeds and pick out the microformat ones?
# 15:56 kylewm it's a little bit weird that i'll have to iterate through the rel-urls key to find that link
tantek joined the channel
# 15:56 Vendan kylewm, why iterate, the url is the key to the rel-urls array
# 15:57 kylewm Vendan: right, if my use case is "find all the URLs that are rel=feed and type=text/html", then i need to iterate over the whole hash
wolftune and KartikPrabhu joined the channel
michel_v joined the channel
# 16:00 ben_thatmustbeme i feel like that would be pretty important for the UI of the reader to have a description with it
# 16:01 Vendan right, though you could grab the rels["feed"] and cross reference at that point
# 16:01 Vendan ben_thatmustbeme, that's what rel-urls is for
# 16:02 Vendan kylewm, wonder if there should be a h-remotefeed or something, so people that are exposing multiple feed types could have them grouped together
JasonO and Sebastien-L joined the channel
# 16:05 kylewm KartikPrabhu: cool, I'll try adding support for the rn
tilgovi and KevinMarks_ joined the channel
# 16:08 ben_thatmustbeme though i think i shoudl probably now filter my list of categories that show up to only those that have more than 2 posts or something
gRegorLove joined the channel
# 16:16 tantek I think we should not depend on nor recommend anything to use "nofollow" - it's been horribly abused and now I think it's bad for web publishing in general.
# 16:19 tantek primary abuse of nofollow has been social silos using it on *2nd party* content (posts), not just 3rd party content (comments). And nevermind the wikipedia abuse.
# 16:20 aaronpk i'm all for dropping it from the webmention recommendation in favor of talking about h-cite
friedcell joined the channel
# 16:21 tantek yes let's drop all mention of nofollow from webmention
# 16:22 tantek ah - it was only in ()? expressions, so we can drop it
# 16:23 tantek if an actual problem comes up with nofollow links, let's find another way to solve it
snarfed joined the channel
# 16:23 tantek every link on every post on every silo you post has nofollow on it, even though its on an account that you authenticated etc.
# 16:24 tantek so there's tons of nofollow links that are perfectly fine, as compared to comment spam (original motivation for nofollow)
# 16:25 Vendan main reason I've seen for nofollow has been for stuff like wikis, where anyone could make lists pointing at sites to bump up their link rating
# 16:25 tantek the solution to that is to revert those edits
# 16:25 Vendan kinda pointless nowadays, as most search engines lower your rating for doing that now
# 16:26 kylewm ben_thatmustbeme: KartikPrabhu: OK added h-feed discover, please try it out!
# 16:26 Vendan and seems potentially harmful for social web building
# 16:26 kylewm KevinMarks++ that would've been much more difficult without the new rel-urls stuff!
# 16:26 tantek.com edited /Vouch (-33) "rm 2 instances of nofollow usage, make the 3rd (in old notes) optional since it's purpose is not clear" (
view diff )
# 16:27 Vendan if someone's used webmention to insert bad links, get rid of the bad links, don't just say "oh, ignore the links"
# 16:28 tantek all nofollow did was allow pollution of content to humans while providing hints at the pollution to search engines
# 16:28 tantek we should have come up with a solution to the pollution instead of annotating it :P
KartikPrabhu and j12t joined the channel
# 16:30 tantek and what do you think of an 18:30-19:30 quiet writing hour before the meetup?
# 16:33 Loqi Ok, I'll tell them that when I see them next
snarfed joined the channel
# 16:34 gRegorLove With the removal of nofollow from /Vouch, how do we link to [bad person] without acting as a vouch for them?
# 16:35 tantek where's KevinMarks to suggest bringing back VoteLinks?
# 16:35 ben_thatmustbeme for that matter, how do we stop one person getting spammed from preventing spam to everyone they are associated with (and recurse)
KevinMarks_ joined the channel
# 16:36 tantek ben_thatmustbeme: if you have friends/sites you vouch for but don't want to use as a source of vouches yourself, you blacklist them for that purpose
# 16:36 tantek it's like real life, you say, I trust this person, but I don't trust their choice of friends :P
# 16:38 ben_thatmustbeme tantek, has nothing to do with trust, one hole exposes the entire network for spam immediately (assuming immediate posting of comments)
# 16:38 tantek but that was true before too - nothing to do with nofollow
# 16:39 tantek so instead let's mark them up as …. ***comments***
# 16:39 tantek rather than some vague pseudo-semantic search-engine robots.txt inspired nonsense
# 16:40 ben_thatmustbeme oh, i see, you are saying adding some other value no rel=nofollow but that would work for the same purposes?
# 16:40 tantek to identify is this a comment, then do what you want accordingly
# 16:41 tantek rather than making a judgment for all comments
# 16:41 tantek that was the problem with nofollow - so poorly named that it started to get used for EVERYTHING
# 16:41 ben_thatmustbeme my original idea was to basically do no follow for anyone i am not following / friend of
# 16:42 gRegorLove Hmm, so you mean a new rel-value to indicate they're comments, or a class attribute?
# 16:44 tantek right. there are multiple possible approaches to consider
# 16:45 tantek !tell aaronpk were there any group photos taken at IWC Germany?
# 16:45 Loqi Ok, I'll tell them that when I see them next
# 16:46 ben_thatmustbeme kylewm: another idea (and i'll probably run this by you again once i am properly publishing it) is having some system to pull off of my "following" (yet to be published)
# 16:46 gRegorLove This is relevant to my interests as I'm in process of writing WM+Vouch plugin and was using the nofollow route so far.
# 16:46 gRegorLove So if I should be fundamentally changing that, now would be a good time.
# 16:46 aaronpk Eye fi didn't like the venue wifi for some reason
# 16:49 tantek aaronpk - re: eye fi - note that Flickr uploading was broken this weekend.
killerog joined the channel
# 16:55 aaronpk oh yeah huh.. i seemed to be able to upload screenshots fine all weekend
# 17:01 tantek news is supposed to be about post about indieweb in general, not just a specific indiewebcamp per se
joskar joined the channel
# 17:03 LukasRos Both which I copied are about the event so I added them to the event. I left them in the news section as well.
# 17:04 LukasRos Not sure whether they should be removed from news section, aaronpk added them there so I didn’t want to touch that.
# 17:04 tantek skimming both those posts - they're both about indieweb in general as well as the camp
# 17:05 tantek aaronpk: it's a soft distinction - we don't want to clutter the general posts about the indieweb with too many just event posts
# 17:05 LukasRos Ok. My post is only on the event which I think is fair since it doesn’t provide any generic value.
j12t joined the channel
# 17:25 killerog I hadn't seen that page yet, should I add my post too?
wolftune joined the channel
# 17:44 killerog Than I should rearrange it :-)
gRegorLove joined the channel
# 17:46 killerog brb, time for food
benwerd and LCyrin joined the channel
# 18:24 ben_thatmustbeme brings things to on topic that were not in this room... though he has tried to raise them in the past
# 18:24 tantek have you written this up in a Brainstorming section or a post?
# 18:27 tantek ben_thatmustbeme: but at least a simple list of reasons?
# 18:28 tantek people made similar noises 10+ years ago about (API name) would be better off posting as XML
# 18:28 ben_thatmustbeme KevinMarks__: you really can't do that now without having the auth token in the HTML. most implementations seem to post only to their own site and then use curl from some back-end
# 18:28 GWG Anyone have a good example of a photo post ui?
# 18:30 ben_thatmustbeme it would give us the ability to be able to tag photos at locations which we cannot currently do in micropub
# 18:30 tantek GWG, aaronpk please make that a Pepperidge Farm meme
# 18:32 aaronpk seems like it'd work the same way as people tagging
# 18:32 tantek ben_thatmustbeme, you may be confusing syntax with design methodology
# 18:32 tantek you shouldn't do with a single POST - that's the point
# 18:33 tantek aaronpk, more like "REMEMBER XML-RPC?" "WordPress Remembers"
# 18:34 ben_thatmustbeme for tagging a location within a photo of a person, you have to create the photo post, then create the h-card (which are not really supposed to be objects)
# 18:36 aaronpk to be clear, i'm not 100% opposed to supporting JSON posts for micropub. it's just that the only arguments for it so far have been unconvincing
# 18:36 ben_thatmustbeme aaronpk, i know, i'm trying to make that case, and trying to figure out how to specify tags of people was the first place i was hitting a major problem
# 18:37 KevinMarks__ given the way aaron sneaked images posts in with data urls, html posts have a loegs
# 18:37 Loqi KevinMarks__ meant to say: given the way aaron sneaked images posts in with data urls, html posts have legs
# 18:37 tantek ben_thatmustbeme: you're not advocating for JSON, you're advocating for complex hierarchical POSTs
# 18:38 tantek ben_thatmustbeme: I thought we solved people tagging
# 18:38 ben_thatmustbeme tantek, there are no other arguements you can make, json and form encodings are just encodings, you can always serialize, its just how complex you want to be
# 18:40 tantek also "locations within photos" is ambiguous and misleading - makes it sound like geo locations
# 18:41 ben_thatmustbeme and tantek, we can specify them in MF2... but we are talking about micropub serialization
KartikPrabhu joined the channel
# 18:42 tantek aaronpk: that's what happens when you're a spec person, you see ambiguities in people's IRC :(
# 18:42 aaronpk getting people to accept HTML content type posts is a stretch
# 18:42 tantek especially since there's more demand for markdown
# 18:43 tantek there's nothing indieweb about makerbase, it's just another silo
# 18:43 tantek why do independents keep building more silos?
# 18:43 tantek because they don't know how to build anything else?
# 18:44 GWG tantek: You could invite them to HWC
# 18:46 tantek KevinMarks: from following #indiewebcamp during the weekend and IWC Düsseldorf
# 18:46 tantek ben_thatmustbeme: we already decided that we need to go back to the drawing board for /area-tags
# 18:46 tantek I'm not sure why you're posting that as a strawman
# 18:46 aaronpk ben_thatmustbeme: the way to come up with micropub requests is to look at the microformat markup and work backwards
# 18:47 KevinMarks__ given that aaronpk just made an editor that will let me post complex formatting via micropub
# 18:47 GWG tantek: According to the site, it is a project of ThinkUp LLC.
# 18:48 tantek GWG, we've invited Anil and Gina repeatedly to IWC NYC - Jeremy Zilar even made badges/pins for them - no shows - and mostly no responses
killerog joined the channel
# 18:48 tantek GWG, you're welcome to reach out to them individually on Twitter and invite them to HWC
# 18:48 KevinMarks__ the 'insert' thingy could add other rich types beyond image and video
# 18:49 aaronpk I am thinking a way to specify html vs plaintext content for micropub requests would be useful... similar to how microformats assumes all values are plaintext except for e-* which gets pulled out as html and value
# 18:49 GWG tantek: I'm still annoyed at last week's This Week in Google show.
# 18:50 GWG I should stop watching Internet tech punditry sometimes
# 18:51 GWG KevinMarks__: You are the talking head who adds s counterpoint to the pro silo view
bengo joined the channel
# 18:53 tantek aaronpk - interesting plain text vs. html distinction, also based on microformats p-* vs. e-*
mlncn joined the channel
# 18:53 tantek ben_thatmustbeme: no need for md-* - as md is *supposed* to be just an enhancement of plain text
# 18:54 GWG I'm frustrated with WordPress this week
# 18:54 tantek any markdown that's NOT just an enhancement of plain text violates MD design principles and should be rejected
# 18:54 ben_thatmustbeme tantek: as to why i was going back to the area-tags issue, i don't think the mf2 markup makes sense, we only had a problem with it when we got to mp serialization
# 18:54 gRegorLove Makerbase is a silo, but the indieweb hashtags I've seen here have been people submitting their projects to the Makerbase directory. That one above looks correct since the person is just submitting their own site http://wiobyrne.com/
bengo joined the channel
# 18:55 tantek gRegorLove: it will become spammers submitting their spam
# 18:55 Loqi ben_thatmustbeme meant to say: tantek: as to why i was going back to the area-tags issue, i think the mf2 markup makes sense, we only had a problem with it when we got to mp serialization
# 18:55 tantek both Anil and Gina should know better than that
# 18:55 GWG tantek: No, but there is a WordPress Weekly
# 18:55 gRegorLove Wasn't arguing Makerbase is useful; I see what you mean now, though.
# 18:55 tantek ben_thatmustbeme: I'm not so sure about the mf2 markup for area tags
# 18:56 Vendan might as well just say text is a superset of html
# 18:56 GWG I'm frustrated with not being able to get microformat change in Wordpress
# 18:57 tantek Vendan++ for properly reductio ad absurdum disproving KevinMarks's assertion.
# 18:57 KevinMarks__ you're allowed to put html in markdown to do things markdown can't do
# 18:58 Vendan you are also allowed to put html tags into text...
# 18:58 tantek Kevinmarks - that just means markdown is *ignoring* HTML - it doesn't mean HTML is part of markdown.
# 18:59 tantek and using markdown that way violates the markdown design principles
# 18:59 Vendan also, I wouldn't count on that kinda behaviour, standard setup on golang routes the output through a sanitization so tight you can't do anything that you couldn't do in MD itself
# 18:59 tantek KevinMarks: not at all - auto_link is intended to work on plain text
bengo, killerog and KevinMarks___ joined the channel
bengo joined the channel
interactivist joined the channel
# 19:10 Loqi Ok, I'll tell them that when I see them next
benwerd_, snarfed, KevinMarks_, benwerd and danlyke joined the channel
# 19:42 snarfed got an interesting (to me at least) indieweb use case
Pierre-O1 joined the channel
# 19:43 snarfed a friend asked me to delete a few bridgy-backfed comments on snarfed.org
# 19:43 snarfed i'd like to hide them instead of deleting them altogether
# 19:44 snarfed nothing too special there…wordpress just doesn't support it. i can trash or spam or unapprove them, but i can't mark them private like i can with posts
bengo joined the channel
# 19:44 tantek A makes public comment on B's public post on silo C. B's site copies comment to their own site. A wants copy of comment removed (or hidden)
# 19:45 tantek snarfed, what if you edit the contents of the comment to say "see original on: (link to original comment wherever it was made) "
# 19:46 snarfed (arguably the silo part of the user story is unnecessary)
# 19:46 tantek pretty sure WP allows admins to edit comments
tilgovi joined the channel
# 19:46 snarfed tantek: not a bad idea. and yes wp does. i expect he'll delete the silo copies too though
# 19:46 tantek snarfed: silo part is necessary because the problem is the copy you make *to your own site*
# 19:46 tantek it's necessary to make it clear stuff is happening on/of copies on multiple sites
# 19:47 tantek also - deletion of the original comment is an existing use-case
# 19:47 snarfed but most server implementations store a copy of indie replies etc too
# 19:48 snarfed maybe the only interesting question is etiquette: whether comment recipients should hide or delete their own copy
# 19:49 snarfed #9, single digit! (only exciting to me and kylewm :P)
# 19:53 tantek github spec is out of date because it fell victim to the non-community bottleneck problem
# 19:53 tantek this is why github kinda sucks for open specs :P
# 19:54 snarfed the wm spec duplication/skew is definitely disappointing
# 19:55 snarfed sure, agreed re github for specs. it'd just also be nice to try to solve this specific instance
# 19:59 snarfed seems like delete vs hide policy is an open question
# 19:59 snarfed pro hide: some recipients will want to own their responses permanently
# 19:59 snarfed pro delete: some sources will want to be able to fully delete their own posts and all copies
# 20:00 snarfed not entirely realistic - information wants to be free, etc - but still a clear desire sometimes
# 20:00 snarfed e.g. silos sometimes try to make promises around how fast and how completely they delete your data when you ask
# 20:00 KartikPrabhu snarfed: i think delete vs hide can be left to individual implementations
# 20:01 snarfed KartikPrabhu: in general, i agree. in this case though, there are two stakeholders, the sender and the receiver
# 20:01 snarfed the interesting question arises when they disagree
# 20:02 KartikPrabhu I mean the webmention spec can't recommend a particular way to resolve the conflict :P
bengo joined the channel
# 20:04 snarfed definitely! that's why i keep using the words "etiquette" and "policy"
# 20:05 snarfed it's not a technical problem, so the spec shouldn't try to solve it, but we as a community can definitely come up with guidelines/best practices
loic_m joined the channel
# 20:07 tantek we can document the choices and explain reasons for each, leaving it up to individuals
bengo and fourtonfish joined the channel
# 20:20 kylewm snarfed: I'm sorry I don't know this, does Bridgy already send a 410 for a deleted comment?
interactivist joined the channel
# 20:27 snarfed but i think the mf2 handlers just pass through the silo's 400 or 404 or whatever for an API call for a deleted response
# 20:30 snarfed i'd happily merge a PR that returns 410 instead when we can reliably detect deletes
snarfed, Erkan_Yilmaz and Lancey joined the channel
Erkan_Yilmaz joined the channel
mlncn joined the channel
# 21:22 Lancey not much, i saw the website and have just been working on a little blog
# 21:27 kylewm have you been able to log into the wiki already?
bengo joined the channel
KartikPrabhu joined the channel
# 21:37 Lancey the wiki's running a bit slow for me so that'll take a bit
# 21:40 Vendan Hey, Lancey, what are you using for your blog?
# 21:42 Lancey i hopped on the rust train but that stopped being exciting haha
# 21:43 gRegorLove Yeah, thought that was odd. Also thought the list there was longer
# 21:44 Lancey i saw we had some recommended frameworks for blogs but i don't think i saw anything django based
# 21:45 Lancey maybe if there were more mf2 compatible blog platforms written in django
# 21:45 gRegorLove That's no problem, Lancey. Baby steps. :) Tantek doesn't accept webmentions yet either.
# 21:45 KevinMarks__ the comments about sample code exampels being microblogs was interesting
# 21:46 KevinMarks__ wodner if finding them and sending mf2-adding pull requests would be worth the effort
# 21:47 gRegorLove Guess it's "under construction" though I don't know the latest status of that.
# 21:56 Vendan heh, my blog uses a key/value datastore for primary datastorage
# 22:01 KartikPrabhu Vendan: hmmm can one open the files in a text-editor/viewer and inspect them directly?
# 22:05 Vendan I'm considering going flat file, but bolt does have several features that I do like
# 22:05 Vendan can probably replicate most of them with flat files though, so that's one thing
# 22:11 Vendan ugh, I get the heebie-jeebies when I see someone running gogs
JonathanNeal joined the channel
danfowler joined the channel
# 22:18 Vendan KartikPrabhu, one of the differences, imho, is that there isn't a lot of what is considered "DBA tax"
# 22:19 KartikPrabhu Vendan: oh you mean file storage does not have DBA tax or that this one you are using doesn't?
# 22:22 Vendan interesting, though it took a bit too long to get to an actual description of what it is
tantek joined the channel
# 22:23 Vendan and it immediately failed 2 of my current criteria, though they are low and medium priority
# 22:24 Vendan big one is really "in-process, no extra server software required"
# 22:25 Vendan I would prefer single file with entire db in it, but I'm ok with the idea of "faking" that by supporting import/export as archives
# 22:26 Vendan ^ main reason I don't use file-storage as it stands, but not really a big issue
# 22:27 Vendan my goal is 2~3 files to upload and get it running about anywhere, including on shared hosting
KevinMarks joined the channel
tantek_ joined the channel
wolftune and KartikPrabhu joined the channel
tantek joined the channel
# 22:56 raucao it looks like known is missing h-feed by default
# 22:56 tantek h-feed isn't strictly necessary for common/minimal use-cases
# 22:56 raucao not that it's hard to add, just making sure i can't use the default theme/layout
# 22:57 raucao as long as the document has h-entries somewhere
# 22:57 tantek it's not strictly needed unless you want to give the h-feed its own name from the page its own
# 22:58 Loqi tantek meant to say: it's not strictly needed unless you want to give the h-feed its on name from the page its on
# 23:18 tantek Vendan: I'm back at a keyboard if you had any questions about Falcon's flat file storage format.
# 23:19 Vendan not really interested in parsing html to get the data to render html
# 23:19 tantek it was the easiest storage format for me to personally edit / verify by hand
# 23:20 tantek figuring hand inspectability / editability was paramount for data quality / repairing data corruption etc.
# 23:20 tantek as opposed to random CSV, or JSON or shudder XML
# 23:21 KevinMarks__ I thought it was XML, so you could use path queries - did that change?
# 23:21 tantek others choose markdown instead, yet I don't how you would put / identify multiple markdown posts into a single file
# 23:21 tantek and one post per file = perf problems as kylewm found out
# 23:21 Vendan possibly, but imho, I'd rather optimize for the main use case and go from there
# 23:21 tantek Vendan: I optimized for data quality / reliability over the long run
# 23:21 tantek which to me means human inspectability / fixability
# 23:22 Vendan I can do single file per post with amazing levels of performance
# 23:22 tantek has inspected far too many file formats in byte editors :/
# 23:22 Vendan cause I can just cache stuff into memory and keep it there
# 23:22 kylewm raucao: any clue why it would be going to remotestorage.io?
# 23:22 kylewm raucao: did you previously redirect to remotestorage.io or something?
# 23:22 tantek KevinMarks: my HTML5 is well-formed XML but it's not "proper XML" which has all kinds of assumptions of using custom tags and attributes
# 23:23 Vendan having a single persistent server handling all requests makes it easier
# 23:23 tantek proper XML is when you use XML as it was intended, that is, you make up your own very specific tag names and attributes for a complete one-off bespoke format that no one else can understand or interoperate with.
# 23:27 kylewm raucao: bridgy starts posse-post-discovery, and instead of fetching your updates.kipp.pe page, it fetches remotestorage.io ... i'm wondering if bridgy had an old redirect still cached or something
# 23:28 raucao that's on the same ip as remotestorage.io using sni for tls
bengo joined the channel
# 23:31 raucao kylewm: i have no idea how it would even know of the domain remotestorage.io though
KartikPrabhu joined the channel
# 23:32 raucao might be an outdated curl or sth. what http client is brid.gy using?
# 23:32 kylewm it's on google appengine, they use their own urlfetch thing
# 23:33 raucao sni is very common these days. it's the only way for different domains to use tls on the same ip
mdik_ joined the channel
# 23:43 Vendan shows nothing but a little bit of ui that looks like a map renderer
# 23:44 GWG Anyone know who Chris Aldrich is and why he is following my WordPress microformats stuff with interest?
# 23:44 raucao kylewm: i can't find anything that looks like an app engine UA in my logs
Deledrius joined the channel
# 23:46 raucao just a bunch of irc clients and browsers in the last 2 hours
# 23:46 kylewm it's possible bridgy fakes the user agent to look like a browser, but i cna't find it rn
# 23:48 GWG It says LA on his Twitter profile
KartikPrabhu joined the channel
# 23:50 tantek Vendan - I had to load it in FF, with a lot of scripts. Safari/Webkit didn't work for me (re: that matrix.yaml thing)
# 23:50 Vendan I tried in what is basically stock FF and it ground my computer to a halt till I killed it
# 23:51 raucao kylewm: so maybe it cached that there wasn't any feed linked yesterday?
# 23:51 raucao that would explain it. not sure how remotestorage.io fits in that story
Deledrius joined the channel