#snarfedeh that's ok. velocity is subjective anyway. disagreeing is totally fine.
#snarfedi just want to make sure indieweb community members feel like it's ok to both publish and consume rss/atom for years to come, until they truly don't need to
#snarfedwhile still working toward the h-feed future
#tantekplus then it encourages individuals to make up their own mind, based on both the facts and the spectrum of opinions
#tantekbut I hear what you're saying, that the page appears to reflect too narrow a set of opinions currently. let's fix that.
#snarfed…ah, i see. heh. finally parsed your last question. i meant, "i'm happy to step up as rss/atom supporter within the indieweb community" :P
#kbswould it be hitting too many hot buttons to ask if there was a page comparing rdfa with the microformats approach? :-)
#tantekkbs - I think you want #microformats for that ;)
#snarfedkbs: probably not. rss/atom tend to inspire conversation here, but rdf and semantic web mostly get crickets
#tantekkbs - no one* who selfdogfoods actually bothers with rdfa. selfdogfooding is an awesome inefficiency filter. *with the exception of tommorris ;)
#tantekkbs - anyway - there's more frequent (still not much, but more) discussion of rdfa vs microformats etc. in #microformats, so you may have more luck finding folks who want to discuss that there.
#kbsah, okay - thanks for tips all (and for pointer to #microformats :-)
#tantekbtw - a lot of the generic design of microformats2 was inspired by both microdata, and bits of rdfa1.1
#tantekpretty sure I've credited that in every talk I've given on microformats2
#kbsbtw, love the indiewebcamp idea [stumbled on this via kevin mark's last TWiG appearance] but don't have much idea behind the history - I'm awash in a sea of format specs, so forgive any naive questions :)
#tantekkbs - I have a lot of sympathy for anyone awash in a sea of format specs - I'm sorry to hear that.
#tantekhas been awash in a sea of format specs since 1998 :/
#kbsheh, that's a clever idea. Wouldn't a user still need to be tech-savvy enough to know if a given web-app is actually only connecting to their private server?
bnvk joined the channel
#pauloppenheimkbs: i am not positive on the specifics, but it uses containers, so I am not sure that's possible
#gRegor`Ok, caught up. tantek, snarfed: Yeah, I think it's the general "tone" about RSS/Atom that concerns me. I don't view XML and h-feed as an either-or, and it honestly seems like quite a low priority indieweb area. Like, if you run a WordPress blog, you have a working feed out of the box. Encourage people to add h-feed and other microformats, of course, but the anti-RSS tone just seems unnecessary.
#gRegor`Re "if you don't want to put a feed on your home page - no problem, then you can make it a separate URL", tantek, that violates DRY, which is one of the main criticisms of the XML feeds.
#dariusdunlapOn Wordpress, there is unfortunately still much work to be done. I’ve got the indieweb plugin installed and I’m using an “approved” theme, and yet the h-entry validation on http://indiewebify.me/ still points out several missing elements on https://dunlaps.net/darius
snarfed, pfenwick, tantek and eschnou joined the channel
#tantek!tell gRegor` it's not a DRY violation if you only have one feed.html page which has your latest posts, *instead of* your home page. XML/RDF feeds are typically DRY violations because they *DUPLICATE* what's already on your home page.
#tantekHey Loqi, do you have problems with gRegor` 's backtick? cc: aaronpk
#tantek!tell gRegor it's not a DRY violation if you only have one feed.html page which has your latest posts, *instead of* your home page. XML/RDF feeds are typically DRY violations because they *DUPLICATE* what's already on your home page.
Jihaisse, rtaibah, eschnou and julien51 joined the channel
#julien51tantek snarfed interesting discussion around RSS/Atom. I generally agree that it's a huge pain to consume (both at the developer level and the user level). We (Superfeedr) have been fighting hard to help with that while still preserving 'interoperability'. There are millions of feeds out there and a lot of different ways to consume them. Let's just not throw out the baby with the bath water! For example, one of our recent attempts is
#julien51https://subtome.com and I would really really love to get your feedback on that.
#julien51And if anyone is looking for really nice alternative to Google Reader, I'd suggest https://feedbin.me It's an "indie" solution, created by a wonderful dude who charges for his services. It's simple, fast and a generally "great" product. [they're not even a Superfeedr customer if you're wondering about a conflict of interest on my end!]
#julien51Also, tantek, you know saying "snarfed - is there an RSS/Atom community? I thought any sense of community died as a result of the RSS/Atom wars." is not true. You're clearly substituting your own reality at this point =)
julien51, caseorganic and melvster joined the channel
#gRegor`!tell tantek I'm not sure I understand. Are you suggesting I replace my home page (currently displays one post) with a page that displays X posts? Because I don't desire to do that.
#gRegor`Tantek did try sending it to non-backtick after Loqi didn't respond with the backtick. But he tried first with the backtick
#gRegor`The other Gregor is online, so I can't /nick to clear it.
#ben_thatmustbemeYeah, i saw that, I'm more just curious if Loqi ignores the backtick in certain places or if it just drops lines that have the backtick in it
#gRegor`I think it doesn't respond to it as a valid username.
#ben_thatmustbemeso when he tried to use tell without the backtick (and Loqi responded if i rememeber) it was just queuing up a message for the other nickname
#dariusdunlapCool, that twitter profile_image trick. Though I notice it’s the small version of the image.
#jonnybarnesI'm happy with the default size, I was originally going to link directly to the pbs.twimg.com link brid.gy sends you, but that can go wrong
#jonnybarnesit appears doing username/profile_image will always give me the latest image, so hopefully no breakages
#gRegor`Re: the feed discussion, I put down some thoughts on the talk page last night. Mostly because I can't always keep up with IRC regularly: http://indiewebcamp.com/Talk:feed cc tantek snarfed
#aaronpkgRegor`: yes thanks for that. Frankly I don't see much point in continuing the rss discussion until there is a productive and practical reason to do so
#gRegor`I'm hoping to be productive. :) And I wanted it to be clear that it's not an either-or for me - not dissing h-feed *or* RSS/Atom. The messaging concerns me a bit is all.
#ben_thatmustbemethe more i look at it, the more I see how much wordpress will not suit my needs
#gRegor`It's really extensible, ben_thatmustbeme, but oh my goodness is the code/API a mess. I rather loathe WordPress.
bnvk_ joined the channel
#ben_thatmustbemethe only advantage to fighting through that mess is being able to share it back to other wordpress users
#dariusdunlapjust had some problems doing and indieAuth login to the wiki using Google+ — getting “invalid Credentials”. Tried logging out of all google accounts and logging in to only the main one, tried both http and https. (though I didn’t try to fuss with the rel_me settings anywhere.) I think this used to work. Problem on the Google side? App.net indieauth worked fine.
#ben_thatmustbemehmm, is it worth while for me to create a new project instead since my host doesn't have mongoDB
#dariusdunlapaaronpk - I’m working on related stuff this morning, so if you need any testing, just holler.
julien51 joined the channel
#ben_thatmustbemeaaronpk what language is p3k? I'm not really finding anything I like thus far, and if I don't find something it will probably mean ripping apart opencart again to use their basic MVC setup
#aaronpki probably won't even load the file contents into the DB, just use it as a way to find filenames to read from
#ben_thatmustbemehuh, interesting approach. Yeah, unless you are doing text searching there is no reason to put in the contents. That must make it nice and fast too.
#kylewmvery cool, does it actually use the git metadata at all, or is that just a mechanism for storing flat files of everything?
#aaronpkI want to be able to query by tags, post types, etc, and maintaining flat-file indexes of those is getting hard, so that's the reason for moving the metadata into the DB
#aaronpkjust using git as version control and a way to sync the files between multiple computers I develop on
caseorganic, glennjones, benprew and snarfed joined the channel
#aaronpkalso I am stumped with this google auth error
#aaronpkthis app was working at one point, and now I've gone and made new API keys and even made a new app, and it's still giving me that error
tantek joined the channel
#Loqitantek: gRegor` left you a message 1 hour, 45 minutes ago: I'm not sure I understand. Are you suggesting I replace my home page (currently displays one post) with a page that displays X posts? Because I don't desire to do that.
#Loqitantek: gRegor` left you a message 1 hour, 44 minutes ago: Yeah, Loqi doesn't like my backtick apparently. Only saw this in my IRC mentions. Same for the ++ point/karma thing.
#tantekgRegor`: no I'm not suggesting you change your home page. My point is that if your home page DOES NOT display X posts, then having a separate "posts.html" page that DOES display X posts is NOT a DRY violation.
#tantek(because it's not duplicating functionality)
#gRegor`Why isn't it? The only purpose of that separate posts.html would be for feed readers (which is what my XML feeds do already).
#aaronpkso I have to click through to each article to read it?
#gRegor`If you're browsing the archives page, yes. If you go to the home page, you see the latest post. Each post has links to previous and next.
#aaronpkso it sounds like you have already published all the information and links needed for me as a person or myself using a feed reader to find all your content
#aaronpkbut you just said I can read the full articles if I click on the links on your archives page
#tanteksounds like a lot of work for the (human) reader
#gRegor`Yes, which would . . . not be in the feed reader.
#aaronpkbut the feed reader can pre-fetch the content if it really wanted to
glennjones and kbs joined the channel
#aaronpkI guess my question is why are you making more work for yourself, you've already made your content and the links to it all available for both human and machine consumers
#gRegor`I'm not sure why anyone assumes I'm "making more work" for myself.
#aaronpkyou seem to think you need to publish an rss/atom feed
#gRegor`DRY is one of the best arguments against XML. Yet a shortcoming of h-feed (with my current site setup) is that I would need to set up a separate HTML page, *duplicating* my RSS/Atom feed in HTML form. So the DRY argument loses its power there. Sure, if I dropped the RSS/Atom feed and just used an h-feed, that would make sense.
#gRegor`And the language of the /feed page makes it sound like RSS/Atom should be dropped *today*, meanwhile hardly any IWCers are actually doing that.
#snarfedgRegor`: definitely. we need to update /feed, right now it kind of reflects a minority opinion. that's on my todo list, but feel free to jump in ahead of me!
#tantekgRegor`: yes, we should update the page to reflect both the current situ and trends more accurately. I think snarfed and I will be banging that out tomorrow night.
#kylewmgRegor`: snarfed: the /feed page has been toned down quite a bit, it's still problematic?
#tantekIt feels like the kind of thing that will work better in-person interactively.
#gRegor`Cool, cool. Thanks for the discussion. Hope I wasn't too frustrating. :)
#tantekkylewm - not so much problematic, as could be better :)
#snarfedkylewm: oh maybe not. i've tried as hard as i could to avoid and ignore the debate altogether. (but recently failed :P)
#tantekgRegor`: it's all good. sometimes harder to discuss contrasting approaches/opinions in text.
#snarfedbret, aaronpk: agreed, h-feed to rss/atom and back seems pretty straightforward. not even sure there are any meaningful design decisions to be made, just simple fetch-on-demand and transformation. at least, that's the naive first pass, which is all i'd expect to do.
#kylewmcool -- I did understand how someone could find it a little inflammatory a week ago, but not anymore thanks to barnaby's edits. looking forward to reading new versions :)
#snarfedaaronpk: out of curiosity, were you thinking something more sophisticated?
#kylewmsnarfed: is your latests blog -> tweet evidence that bridgy publish supports picture posts now?
#bretsnarfed: another approach, send a webmention/webhook to the gateway on updates?
#tanteksnarfed, per Postel's law - does it make sense to produce anything other than the strictest/most valid Atom from h-feed? And then consume *any* Atom/RDF/RSS and produce the strictest h-feed?
#snarfedbret: maybe! but that implies the gateway has some state it would need to update. i'd just as soon avoid that.
#tantekideally the proxy shouldn't be polling our h-feeds
#tantekit's about time we simplified PuSH and made it work for native h-feeds
#gRegor`I didn't compare to its recent history. I didn't find /feed inflammatory in its current state, just a bit of concern about the tone/message. I commented last night on the talk page, which I just mostly recapped here in IRC. :)
#snarfedyeah. honestly, personally i would (will) probably only build the fetch-on-demand, nothing more
#snarfedtantek: re caching, sure. if you already serve rss/atom, though, this wouldn't increase your load much, since the number of users of h-feed readers is approximately…2? 3? :P
#Loqibarnabywalters meant to say: my VPS gets about 1TB of monthly bandwidth so feel free to use it, within reason :)
#tantekbarnabywalters - any thoughts of supporting PuSH+h-feed?
#barnabywalterstantek: recently I’ve been trying to get PuSH 0.4 to work with h-feed
#barnabywaltersas my reader and anti-spam projects both need to be able to subscribe to things
#barnabywaltersno success so far but that’s a result of the feed reader trying to also be a programming environment. I think I need to focus it on one purpose and optimise it for that rather than trying to over generalsie
#barnabywalterswhere “being a programming language” is about as general as you can get in this case :)
#barnabywaltersbut current working consensus from julien is: subscribe to a HTML page, get pings when it changes, pings contain the content of the HTML page which you then re-parse
#barnabywaltersIIRC eschnou also work(s/ed) on a service which does similar 1-click installs
#ben_thatmustbemebret: sadly they both sound depressing, I don't want to start from scratch to get far enough to the point of things I want to work on, but I also don't want to try working with the mess that is WP
#barnabywaltersben_thatmustbeme: presumably someone’s already suggested looking at idno as a base to build off?
#snarfedben_thatmustbeme: what problems are you having with wp?
#ben_thatmustbemeyes, I like the look of the code, but alas, sql only host and while I could dev locally, it would not be fun to not be able to launch on the hosting i've prepayed for
#ben_thatmustbemesnarfed, mostly I was looking to hack it pretty heavily and wp doesn't always work the intuitive way for me. I'm leaning toward just sticking with WP for now
#snarfedgot it. good luck! plenty of both wp and non-wp users here to support either path
paulcp joined the channel
#ben_thatmustbemeI suppose the easiest for me is a framework I've already used and could get it together fast, but I wouldn't really be learning as much
#snarfedeh, you'd still learn plenty about indieweb. and honestly, poking at both in parallel isn't so bad! you get something working fast, and you also get to try building from scratch.
#ben_thatmustbemeI think i'm going to just go the MVC route i'm used to. It will take me through a lot more of learning the indie web standards, and thats really what i'm interested in
#ben_thatmustbemei may be back in a few hours belaboring the decision again
kbs`, julien51, rtaibah, KartikPrabhu, KevinMarks and benprew joined the channel
#tantekthe events support will likely force the implementation of webmentions etc. - because RSVPs to real world events are far more interesting to me than comments.
#tantek(personal preference only - obviously not reflective of anyone else's priorities)
#waterpigs.co.ukedited /PubSubHubbub (+547) "/* Testing */ moved to How To section, stubbed subscription how to with link to superfeedr docs" (view diff)
#tantek(note the lack of icon, name, and dangling "Article on" on that comment)
kbs, rtaibah and scor joined the channel
#tantekanyone here do custom CSS for specific blog posts, and if so, what technique(s) do you use?
#tantekI'm currently using 'style' attributes on per-post custom CSS, which is perhaps suboptimal but I'm not sure where else to start for a more "general" solution (without getting bogged down in template-reinvention-hell, that's a project for later)
scor and KartikPrabhu joined the channel
#KartikPrabhutantek: I do custom css for some posts but it is through templates
#snarfedtantek: sigh. that happens when the semantic-linkbacks plugin isn't installed, or i guess if it chokes for some reason
#snarfedi've been ratcheting down the G+ poll interval for exactly that reason. it's at 30m now. maybe i need to drop it more. :/
#snarfedtantek: re custom css, yes, i've done both inline styles and custom class names that i style in a .css file. both are straightforward. not sure it needs any more magic or special techniques beyond those.
#tanteksnarfed, re: "it's because i sometimes hit the 10k queries/day G+ api limit," - perhaps don't send empty interaction webmentions then? So we don't get empty interactions on Bridgy consumers.
LauraJ joined the channel
#snarfedtantek: i definitely understand the motivation. it's not quite that easy, since i don't currently track the number of g+ calls i've made, and i don't think there's a way to ask, but i'll think about it.
#tantekben_thatmustbeme: I'm not quite yet a fan of using scoped style though I have high hopes it will help solve this use case.
#KartikPrabhuscoped styles created more problems i feel. I have setup my templates to look for a CSS file with the same title as the article and insert a <link> if it exists
#pauloppenheimsnarfed: holy shit, that atv story involves a lot of luck, glad you survived
#ben_thatmustbemekartikPrabhu, do you just make a full copy if you want to re-use a custom style then?
#KartikPrabhuben_thatmustbeme: I use a CSS preprocessor mostly so small common styles as in separate resuable files. I've not needed to reuse huge amounts of custom styles.
#ben_thatmustbemeI'm having to force myself to remove proper markup from my new site... just to force myself to learn the hard way to insert the correct tags. I shouldn't have started with a template, I feel like i'm cheating.
#tanteksnarfed, different from issue 15 - if there is no info for the activity, then perhaps don't post it? i.e. if an error is returned.
#snarfedtantek: well, sending a webmention is a separate request from serving the source URL, and the architecture is pretty stateless. i don't currently fetch from G+ when i send the webmention, so i don't actually know whether the later fetch will succeed or not
#tantekstateless is fine - perhaps don't send the webmention if the fields are null?
barnabywalters joined the channel
#aaronpkwow superfeedr is the only hub that supports PusH 0.4?
#snarfedtantek: sorry, i'm probably not being clear. during a poll, i fetch the post and its comments and likes, store new ones, and enqueue tasks to send webmentions for them. later, independently, i try to send those webmentions. (no silo api calls there.) later, independently, the recipient asks me for the mf2 translations, which i use silo api calls for. that last part is the one that failed in this case. when that happens, with this archite
#snarfed…which brings me back to the ideal answer, serve the mf2 from my stored data instead of silo api calls. doable, but a bit harder than it sounds, due to implementation details.
#aaronpkthat is not at all like how I imagined doing it, but that is awesome!
#kbsrandom question - has any 'de-facto format' to embed pgp keys in a h-card (or (hcard I guess) taken root? Right now it seems a bit all over the place.
#tantekgRegor`: feel free to create one - so far we've been using the newsletter vol 1 as an explanation
#cweiskeaaronpk, I once talked to you about accepting webmention via "202 accepted" if they get processed asynchronously, and wished there was a way I would get notified about the processing result
#KartikPrabhutantek: re custom styles: do you think custom styles and scripts for a post should be syndicate-able, i.e. put inside the h-entry so others can consume it if they want?
#tantekKartikPrabhu - not sure. Going to need more indie reader experience to answer that.
#tantekkbs - what's strange is rel="key" makes more sense
#tantekbecause the specific content-type of the key should come from HTTP
#tantekor in an advisory role, from the "type" attribute of <a>
#tantekkbs - it depends on whether you include the ASCII hardened key inline in your h-card or if you keep that in a separate file
#tantekthat's a publishing choice you make independently of any card format
#kbsSounds logical enough - so is the thinking that the hcard class="key" be used for inline keys, and the generic <a rel="key" be used for links to keys?
#tantekkbs - that seems to be the pair of patterns emerging
#tantekthough for h-card, class="p-key" for the inline data
#tantekyou could do both rel="key" class="u-key" on a URL to a key
#cweiske.decreated /stapibas (+694) "Created page with "Standalone pingback server written in PHP, storing data in MySQL. == Features == * Receive webmentions and pingbacks * Send out webmentions/pingbacks to each link in an Atom f..."" (view diff)
#pauloppenheimtantek: sandstorm is not necessarily indieweb related, actually sounds more like a Docker competitor. Also the author is not merely ex-google, he made protocol buffers and cap'n proto
#tantekpauloppenheim: the marketing copy on the home page makes it sound like it is indieweb related
#pauloppenheimtantek: i haven't dug too far into the code, but it seems like a weird way to go about doing what the site describes. Then again I might not be the best person to evaluate that.
#kylewmI find it really disconcerting how the sandstorm.io animation stops abruptly
#pauloppenheimtantek: yeah, i think that's where the intended audience is eventually, but right now the project seems to requre understanding cgroups and chroot
#pauloppenheimbut my housemate is writing an operating system to do something similar, and has been working on it for a few years now, so there are likely traps in there that are as yet unseen
#dariusdunlapHmmm. CSS still gets a lot of validation errors, but that seems to be in bootstrap.min.css — not my fight.
#dariusdunlapI’m a little bored with this… maybe I should fire up Xcode. (ha ha)
#tantekbret - when you click the aim: link on tantek.com - does it automatically open our previous AIM conversation?
#pauloppenheimdariusdunlap: hey man, you can test that "patches accepted" phrase everyone always throws around!
#tantekand activate/select the message entry field (thus showing keyboard etc.)?
paulcp joined the channel
#dariusdunlapYou mean wrt the bootstrap.min.css? heh heh. Seriously, I’m wondering if I have time to stick my head in my ADN Python code to see if I can figure out how to make brid.gy talk to app.net.
friedcell, _6a68 and caseorganic joined the channel