#tantekKartikPrabhu - no I don't but if you want to note any specific issues / topics that we should capture on the wiki, feel free to bring them up. I'd prefer to focus on discussing/debating issues rather than the sources thereof (citations are good of course, but labeling something as someone's personal issue is not)
#aaronpkI must not be rendering the uid that bridgy sets in my /mentions feed
#aaronpkso, the answer to your question is, yes there is a real h-entry in the wild that has no url or uid, it's here http://aaronparecki.com/mentions (I can't send you a link to it obviously cause it has no link)
#tantekKevinMarks gRegor` pdurbin please let's avoid criticisms of Dave Winer as a person, regardless of what he may or may not have done to any of you personally.
#aaronpkso it's not *just* a theoretical problem, because it's a problem with my feed right now
#tantekIt doesn't help discussion of indieweb *topics*, technologies, features, etc.
#gRegor`Sorry, I didn't mean that as a criticism actually. Just a comment.
#gRegor`If I was in his shoes, I might have blocked me too. Not like we have any relationship.
#aaronpkand I assume that if I managed to generate an h-entry with no uid/url then probably someone else will soon too
#tantekgRegor`: I'm not debating whether any such actions were right or wrong.
#tantekI just want to avoid any of us falling into the trap of debating or criticizing an argument based on criticizing or other behavior of the person making the argument.
#tantekwe're (trying to be) better than that - let's discuss ideas on their intrinsic (de)merits, not by who is making what argument or taking what position
#tantekalso - doesn't matter if others *do* criticize arguments by criticizing the people making them. we can set a better example.
#gRegor`Understood. I wasn't even concerned with a topic of argument. It was just an offhand comment about being blocked. No grudge here
#gRegor`But I could have made that more explicit in my remark
#tantekand to be clear, I've made this mistake myself in the past. it's never been to my (or any of my arguments) benefit.
#tantekaaronpk, the Bridgy example is interesting, but I guess I'm still confused by how you got the h-entry in the first place if you can't point to it
#tanteksurely you got a "source" URL in the webmention of it?
KartikPrabhu joined the channel
#aaronpkyes, there is something wrong with the way i'm handling it. my point is that I still was somehow able to generate html for an h-entry that does not have a URL or UID
#pdurbintantek: I'm critical of Dave Winer's words, which smack of homophobia
#aaronpkso my question is what should a consumer of a feed do if there is no url or uid
#aaronpka) find/make some other unique identifier, or b) treat the h-entry as invalid
#tantekaaronpk - how about start with a) the consumer should ignore it, and b) debug why you're generating it
ttepasse joined the channel
#tantekbecause I guess what I'm saying is, I don't see a valid use case for such h-entry's yet
#aaronpkif the consumer ignores it, then I am going to make the consumer report the error back to the producer like how google webmaster tools has a lot of debugging tools
#tantekand it seems like the result of a bug anyway
#aaronpki don't care whether it's a valid use case of h-entry or not
#aaronpkmy question is how does a consumer handle it
#tantekmy point is, if it's not a valid use case, then default ignore
#aaronpkso if it's essentially required for an h-entry to have a uid or url, then I should report missing uid/url as an error
#tantekdefinitely seems like something that can easily ignore things that don't look "real"
#tantekbut yeah, what you said, about ignore / report errors in debug mode seems fine
#aaronpkif there is a "problem" with an h-entry, it needs to make a decision on how to handle it. I want it to be useful for people to know what is wrong with the feed rather than silently ignoring failures
#aaronpkso it sounds like the correct behavior is that a consumer should treat either uid or url as a required field, and report it as an error if not present
#tantekaaronpk - perhaps consider filing "problem with an h-entry" thoughts as feature requests for indiewebifyme? not every consumers of h-entry need to be error reporting / debugging validators
pfenwick joined the channel
#aaronpkignoring h-entries with no url/uid is getting dangerously close to one of those being a required field
#aaronpkbut i'll leave that for another discussion
#Loqisnarfed: cweiske left you a message 2 hours, 9 minutes ago: linkback to my website failed (https://www.brid.gy/twitter/kyle_wm), could you re-run it? I tried to fix it
#Loqisnarfed: cweiske left you a message 1 hour, 55 minutes ago: it worked since bridgy tried after 1 hour again
heath, KevinMarks and caseorganic joined the channel
#caseorganicwaiting in Mozilla lobby with lots of people for Homebrew website club to be let upstairs
#tantekbecause she's constantly managing their FB page etc.
#tanteksnarfed: this is where Amber's talk of "Generations" comes in
#tanteksnarfed: we're slowly making stuff work with less technical work
#tantekdarius: one of the things he did with the HTTPS stuff on the wiki was to edit those pages to make them more clear - to him as he was working on SSL support
#caseorganicBlocking: can't access #indiewebcamp on freenode from mozilla guest network
#caseorga_Alex Linkster is here in portland working on something related to the moon is a harsh mistress and implementing it on his site as a trinary node structure with certain cells connected to other cells
#caseorga_"is there a use case you have in mind?" snarfed
#aaronpk... organizing groups, leaders of groups assigning tasks
#aaronpkbret.io is a jekyll site, done some experiments with setting it up with indieweb stuff
#aaronpk... recently built an atom feed for yesterday's IRC feed
#aaronpk... currently updating at random times during the day (need to fix that)
#aaronpk... next project is a more general h-feed/h-entry to atom gateway, probably written in node
#AmyMac@caseorga I just caught up on the logs, so now I understand why it was mentioned. It's a great club. We have a lot of fun and do a lot of stuff. Site needs a lot of work though ;-)
#LoqiCamlistore Install Fest on Monday, Mar 31, 6:00pm at Mozilla
#dietrichtantek: people -> urls -> launchers was something that's only talked about in the last couple of days and then i saw your tweet last night. #cosmic
hanni joined the channel
#benwerdtantek - re: iOS device detection, it really is all about the user agent. I like the look at this class because it groups by device *type*: http://mobiledetect.net/
#hanninot hanny, although very hilarious that someone has that on a WordPress.com site :)
#aaronpk... currently trying to add micropub and oauth endpoints to try out ownyourgram.com
#hanniaaronpk: i'd love to help with the wp plugin, I'm no dev (see above disclaimer), but ... most things can be done with good intentions and a deisre to learn
#aaronpkknows very little about wordpress unfortunately! But happy to help on the micropub side!
tantek joined the channel
#hanniaaronpk: sounds like we might have the right mix!
#AndySylvesterbret: tried both with my site (andysylvester.com), Barnaby's service works with my site (was able to subscribe to the feed URL), Sandeep's did not return any entries.
#bretyou trigger on some kind of new content event (like a new post or an edit) to send a webhook or webmention to the atom gateway. this parses your h-feed and generates the updated atom feed. when that is ready, ping the pubsubhubbub hub
mattmartin joined the channel
#breti was thinking about having people just deploy their own gateway to some free paass
#mattmartinanyone here willing to give some direction to a newb
#kylewmneed a little advice from real web developers...i want to add a background queue for processing webmentions, sending stuff to syndication, etc. outside of the regular request/reply flow (Python/Flask). google is pointing me to Celery + Redis but seems like that's overkill?
#snarfedkylewm: yeah, celery + redis is both the right idea and also overkill
#aaronpkit's a minimal queuing protocol based on the design of memcache
#aaronpkthat's what I use in p3k (and also that is what has been in production with the geoloqi API for a couple years)
fmarier joined the channel
#kylewmsnarfed: cron is an interesting minimal idea, where would you save information about the jobs being queued?
#snarfedwherever you're storing your posts and other data
#kylewmaaronpk: thank you, I will read up on that too
KevinMarks and benwerd joined the channel
#aaronpkit's simple enough to understand that it's not hard to use in any language, and you don't really need a whole frameworky thing on top of it (like sidekiq and I assume celery)
#jsilvestrepdurbin, for personal cloud platforms that basically need a store, that's even more interesting
#barnabywalterspdurbin: jsilvestre: yep, /store has a bunch of research and brainstorming about store formats and experience for browsing and installing server side software on servers
#jsilvestrei'm looking at it for Cozy (http://cozy.io) even if it's not the focus at the moment, we might take part of the thinking and implementation effort at some point
#barnabywaltersjsilvestre: oh great! bnvk and I were checking cozy out
#jsilvestreyes, bnvk mentioned us the /store in the first place
#gRegor`Me? Chicago didn't have one last night. April 9 will be the first.
#ben_thatmustbemei just read through the logs of it all. There are certainly plenty of projects out there that detect browser from within PHP. browscap.ini being a huge one, but thats a large file for just detecting browser
#ben_thatmustbemewell I live like an hour out side Boston, but I don't really want to start a meetup until I actually find people that want to meet up. especially when I could be at home with my wife
#gRegor`I've used CakePHP framework for a few projects and rather like it. It's fairly bulky, though.
#ben_thatmustbemeyeah I've not played with that many, at this point opencart is what i know best. I decided to use it since thats really not the piece i'm looking to learn right now
_6a68 joined the channel
#ben_thatmustbemePlus I don't want to get half way through a site and realized I don't like the framework.
#kylewmnow trying to leave a regular comment: Forbidden
#tantekand this is why we have "Beyond Blogging and Decentralization" right there on the home page
eschnou joined the channel
#tantekalso - @closetgeekshow's profile links to a *.flavors.me site, not to an "old-style web" personal site. Doesn't seem like they're into IndieWeb in any way, moniker or with their own stuff.
#barnabywaltersha ha yeah ’cause the “old style web” totally had cross-site RSVPs
#tantekaaronpk - I was wondering if there was anything you could do server-side and just not send the link if the client didn't support webrtc
#aaronpkah, you could make an educated guess based on the current support of major browsers, but that's back to maintaining a list of features and versions of browsers
#jonnybarneswoah, tantek's page was scary then, for a second all I could see was a big picture of his face, then it shrunk into the top corner like it should do
#kbsIs it correct to assume that in indieauth, there's provider-specific 'magic' urls to find suitable reverse rel="me" links on the provider's page?
#kbsThe related question is - do such (any?) providers also have ways to embed key urls - it strikes me that it might be a way to gain 'confidence' in a key obtained from any one site. (some sort of public-key analog for indieauth)
#kbsis there already some work around automated validation/proof that a given fingerprint is really from person X? [eg: a special rel=foo link that points to (say) a post on twitter that contains key material, etc.]
basal and tantek joined the channel
#KartikPrabhuI just got the following nugget in an indieweb discussion "even for indiweb you will still end up hosting your content on a webserver which is a thirdparty. so again you technically don't own your data. so for a true indiweb you need to host data on your own servers"
#bearyou can only know about the "inserting bogus rows" issue if you add metrics for database specific things and monitor logs - something that even seasoned ops folk don't do
#bearusing a database works if you need what a database offers - searching and grouping
#aaronpkdatabase barely offer searching, you really need a search app for that
#aaronpki haven't actually looked at it in a reader in a while
#tantekcweiske - I think I saw you asked about articles vs. notes? I include the ActivityStreams Atom object-type element / markup extension for any consumers that want to distinguish between those post types based on an existing spec.
#cweiske"I opened a bug reportâžš and got the not so encouraging response that I should fix that myself since the author was not willing to support "terminally lazy people""
#tantekKartikPrabhu: yes, human-optimized is a better way to design than SEO
#ben_thatmustbemepokes around at everyone else's sites for ideas since on slug format
#onewheelskywardGood for seo, bad for humans, sometimes.
#onewheelskywardLike the slugs that give away the entire article.
#kylewmben_thatmustbeme: one thing i learned from these folks -- consider making the URL uniquely identifiable without the slug, and have the slug just be for humans. so if you went to example.com/20140227/1/blog-post-title or example.com/20140227/1/edited-blog-post-title, it would redirect to the same place
#kylewmbut not everyone does that -- waterpigs.co.uk for articles, the slug is the identifier
#KartikPrabhuonwheelskyward: those are just bad slugs then
#gRegor`There's a start on some permalinks updates. Maybe it would be better to split them out by domain instead of platform? Since most of us are using different platforms anyway.
#gRegor`I thought WordPress would be a good one since so many sites use it, and a few IWCers do.
#KartikPrabhuUI/UX wise i thought it was great that users needed only one page to do everything. It means that the FB UI/UX now is stupider
#kylewmwhoa, "This is truly a nightmare scenario for any CEO: do you take the risk and proceed with the better user experience/product at the expense of short term numbers" ... that should have a VERY easy answer
#gRegor`I am pretty familiar with encodings and converting them. Let me know if I can help.
designdream joined the channel
#gRegor`One of the simplest routes I've seen is perform a mysqldump, search and replace the charset in the create table statements to be utf8, use iconv to convert to utf8, then use the dump to recreate the tables.
bnvk joined the channel
#gRegor`Also alter the database to set its charset to utf8, so all future tables should default to that.
#KartikPrabhuyeah i have mysql in utf8 but I am not sure I am encoding everywhere when my python talks to the db and so on. have most large holes plugged but there might be smaller leaks :)
#tantekI have a feeling as more indieweb folks implement integrated readers, we're going to figure out solutions to the problems that julien51 brings up
#tanteksolutions which provide actual real world incentives to indie web site owners / creators / maintainers
#tantekright now, I've deprioritized a lot of stuff (e.g. subtome, h-entry PuSH etc.) because there is no apparent benefit to me
#tantekor rather, I've prioritized working on things which show direct benefits
#tantekmaybe those of you with integrated readers working can provide an analysis of julien51's post and see how maps to your needs?
#tantekis anyone here actually consuming PuSH (of any variety) on their own indie web site?
#tantekbtw speaking of hosting, just saw these two blog posts by Brent Simmons with real world scenarios / analysis of different hosting options/levels:
#KartikPrabhutantek: re: indiereaders. I do think having our own readers on our own site is the way to go. probably connected via mentions and micopub for noting/replying. we'll surely get to those if someone feels motivated enough... babysteps :)
#tantekapples and oranges cweiske - you're comparing different content
#tantekpoint still stands, same content with h-feed markup < same content with Atom markup
#tantekthe HTTP requests thing is orthogonal to format
#KartikPrabhutantek: there does seem to be a number of people who want to publish only links or links+summary to articles on their homepage (including me). The RSS/Atom way would be to just have those and a separate RSS file with full articles for syndication. But with h-feed one seems to be forced to use more requests or have all articles as full text.
#caseorganicpdurbin: i'm going to do a shorter intro to indieweb with willowbl00 (she's our berkman center contact and also works at mit media lab right now) anything missing from that podcast you'd like included in it?
#kbsnewbie question - are the discussion pages on the wiki reasonable places to ask for feedback on ideas (or work that's already been done) or is there a better approach you'd recommend?
KevinMarks joined the channel
#tantekkbs - discussion pages don't work very well on wikis - they typically get ignored and neglected
#kbsSo, before I blather on - the general area to do with an indieweb proof-of-ownership of pgp keys, without needing things like keybase. Someone stop me before I go on a long discourse, with pointers to work already done?
#kbsThe idea here is this link indicates a "proof", approximately like how one might add a rel="me" link as used by indieauth.
#kbsstep 3. A validator just pulls out the raw content from the target page, and searches for a fingerprint claimed by the canonical page.
#kbsThe validator might [similar to indieauth] only choose to examine specific providers, that lets it identify the silo author from a page, with a degree of confidence
#kbsfinally, It then presents the list of authors (from each silo'ed site it understands) as a measure of confidence that that key. [eg: this fingerprint was found in @t's stream, tantek's g+ page, etc.]
#tantekfingerprints and confidence aren't really a use-case. secure authenticated communication would be a use-case.
#kbstantek: great - going through your edits and the pages - appreciate the crisp summaries :-)
#KartikPrabhutantek: should I start a Brainstorming section on hfeed to document the u-uid idea?
#tantekkbs, sometimes it helps for someone more clueless (me in this case) to try to summarize at a high level to see if the concept is being communicated ;)
#tantekKartikPrabhu: yes please do, starting with the use-case you described
#pdurbincaseorganic: I might put "tentative" next to my name on the guest list. I have two little kids and often these things end up being game day decisions.