#gRegorLoveIs the "open external" icon at the bottom supposed to open a Woodwind view of the single post? I though tit would open the original in a new tab.
#aaronpkKevinMarks: lol nice link to "Medium's brilliant engineers" ;)
#tantek__number5__: right, kylewm shared that link earlier.
#tantekthe question is, how the efficacy / validity of the information at such one-off domains can be verified? who checks it? or is it taken on faith that all the magic hex codes in there are correct?
#petermolnarI wonder if we could do that with invisible android emulators running behind a service running actual the instagram app and some kind of glue app to post from anywhere else
#benborgesYunohost stack & i will play with custom Yunohost apps to bundle a Known one
#benborges"self-hosted in a box" labriqueinter.net comes with email, jabber, vpn, etc...it's simply amazing
#endiEverything I did was simple since I was just starting; I'd setup Apache a few times but first time with a full LAMP stack, first time w/ a DB, first time configuring HTTPS
#petermolnarbenborges a ages ago i started with virtualmin which was a pretty impressive bunch of perl scripts setting up everything for you for shared hosting
#petermolnareventually you want more and more control and understanding and will slowly, gradually start building things from (near) scratch
#petermolnar( not gentoo though; that's just too hardcore )
#endiAs I set stuff up I am trying to learn more and get better connected
#petermolnarendi apache?! why not nginx? especially on rpi
#endiI haven't used IRC in ages but I thought I'd start to lurk around see what I can learn
#endipetermolnar: umm just familiarity I suppose? I've had 2-3 sites on Apache, I had done a few virtual servers on a machine etc so I felt it best to start with something familiar
#petermolnarplease, pretty please, take a look at nginx + php-fpm; you'll never look back :)
#benborgeswell Younohost is soo well done and simple (most of automatization is bash scripting) iv seen a non techy geek friend, start from nothing to having part of his needs self hosted in one day
#endinginx lighter on resources / better performance I would imagine?
#petermolnaractually, it's pretty easy to get used to it
#benborgesat first, coming from apache, it might look like, but it's just a matter of getting used to it
j12t joined the channel
#endiWell I've been meaning to move this to an RPi 2, I'm sure it could benefit from the extra resources then maybe I'll repurpose the current Pi for learning nginx :)
#petermolnarI'm not entirely sure; I'll take a look at the search page template, I guess the markdown formatting is not applied for some unknown reasons
#endiGood post. I personally really like Google+ but I hate that they've been so closed.
#petermolnar!tell kevinmarks the internet lack the "explore", "discover", "recommended" & similar tabs on it's own; how could those be replaced, those, present in silos?
#LoqiKevinMarks_: petermolnar left you a message 2 hours, 55 minutes ago: the internet lack the "explore", "discover", "recommended" & similar tabs on it's own; how could those be replaced, those, present in silos? http://indiewebcamp.com/irc/2015-05-22/line/1432313749064
#KevinMarks_these are things that make sense in a reader
#tantek!tell kylewm,cweiske re: rel=self required by Woodwind - I recall a conversation with aaronpk where we concluded that if there is no rel=self, that implied rel=self to the page itself was a good way of handling that, to make the common case even easier for publishers.
#kylewmbenborges: too bad, thank you for providing all the log info and stuff, hopefully Known people will be able to help on that github issue
#Loqikylewm: tantek left you a message 3 minutes ago: re: rel=self required by Woodwind - I recall a conversation with aaronpk where we concluded that if there is no rel=self, that implied rel=self to the page itself was a good way of handling that, to make the common case even easier for publishers. http://indiewebcamp.com/irc/2015-05-22/line/1432328006141
#tantekVendan - "LiteSpeed Open Source Edition / OpenLiteSpeed is a high-performance, open source alternative to Apache that retains many Apache-compatible features." from http://www.litespeedtech.com/
#Vendanand yet, you don't get most of what they tout unless you pay
#Vendan"beats" everything else in horribly inaccurate benchmarks, and the main author is paranoid to the point of thinking other web server authors are out to get him
#aaronpkjust moved a site from apache to nginx and had to rewrite all the rewrite rules and move them out of the htaccess file into nginx config... was not fun
#kylewmI find the nginx equivalent of rewrites to be much more intuitive though, if you are starting from scratch
#VendanI like lighttpd's better then apache's myself, seems like it has more power, and so much of it is just regexs
#Vendanbut I end up doing almost everything just in golang, so it doesn't really make a difference
#tantek!tell petermolnar what are the "explore", "discover", "recommended" & similar tabs that you're referring to? Is this on Medium? Could you make/upload a screenshot to /Medium if so?
#tantek!tell petermolnar The general idea is, if you have silo-feature-envy: 1. take explanatory screenshots of the feature, 2. upload to the wiki, 3. document it in a /silo_name#Features section - that way we have something concrete (in time and URL space) to refer to to start mocking up indieweb versions.
#kylewmso about rel=self, a site that publishes with PuSH MUST include rel=self links, but if they don't, Woodwind as a subscriber could still fall back to just whatever the current URL is right?
#gRegorLoveI'm actually a bit impressed with GoDaddy control panel improvements. It looks more modern and isn't as confusing.
#tantekkylewm: I understand the current MUST include rel=self link requirement, I'm suggesting relaxing that to a SHOULD to make the common "same page" case even less work for publishers.
#kylewmtantek: I re-read the irc logs that aaronpk linked to and was convinced (again) that MUST for rel=self is a good idea
#aaronpkkylewm: i think the hub has to agree on the missing rel=self too
#tantekwhy kylewm? can you provide a counter-argument to my point simpler for publishers? (which is one of the stated design goals for PuSH)
#aaronpkregardless of the outcome, it sounds like it would bea good idea to document this as an FAQ
#kylewmtantek: yes, "<julien51> the reason for specifying the self is because we will only get pings for a specific url from the publishers" ... even if the same resource is served at http, https, /, and /index.html, only one of those URls actually works as the push topic
#kylewmit's a good sanity check that makes sure the hub and subscriber are talking about the exact same URL
#kylewmand not that much more work, since you already have to add a Link header or <link> element, adding another one isn't 100% more work :P
#tantekkylewm: I didn't find that anywhere in the logs that aaronpk linked
#tantekkylewm: no that argument doesn't hold water - OpenID thought the same thing (what's the big deal about requiring two <link> tags? and people got it wrong. it was fragile and dumb)
#tantekor rather, expereince has shown that requiring another Link header causes fragility and brokenness
#tantekhowever, this "even if the same resource is served at http, https, /, and /index.html, only one of those URls actually works as the push topic" - is an already solved problem
#tantekand previously solved it with rel=canonical
#tantekthus if that really *is* a problem for PuSH as well, the PuSH should build upon the pre-existing rel=canonical for the page, rather than *require* rel=self
#tantekyes, re-use existing (especially deployed!) just-as-simple building blocks instead of making up your own
#aaronpki wonder if that was ever brought up in the development of pubsubhubbub
#tantekthe bar for making up your own has to be reasonably high, that is, provides some obv benefit like less work for the publisher (fewer tags/classes/attributes), or more reliable (e.g. visible data instead of invisible metadata), more open (based on an open standard instead of a proprietary snowflake)
#tantekthe point is, make rel=self a SHOULD (not a MUST), then in its absence, consuming code MUST fallback to rel=canonical, and absent that, MUST fallback to the URL itself that they retrieved to discover the hub in the first place
#KartikPrabhuand highlights are trickier because they are phrase-level by necessity which fragmention.js does not yet handle
#aaronpkit's not necessarily "likes" but i think basically the same mechanism
#KartikPrabhuaaronpk: ooo I haven't seen Kindle at all any screenshots?
#tantekif so, one line facepile (of mini-faces, like 1em tall, sparkline style) by default with a "…" ellipsis link that wen clicked expands and shows the rest of the facepile, just like a marginalia comment would
#Loqitantek meant to say: if so, one line facepile (of mini-faces, like 1em tall, sparkline style) by default with a "…" ellipsis link that when clicked expands and shows the rest of the facepile, just like a marginalia comment would
#KevinMarks_fragmention.js would need to inject markup into the page for the highlight
#tantekKevinMarks: the usual :) what is marginalia?
#LoqiMarginalia are interactions or responses to arbitrary parts of a post, in contrast to a traditional comment which refers to the entire post https://indiewebcamp.com/marginalia
#KevinMarks_Ev: "you can incorporate anything from outside as well as anything else can incorporate anything from outside. i.e., links (and embeds). open to more suggestions on how"
#tantekKartikPrabhu: and for reposts: ♺ 👨👸👲👴👳…
#KartikPrabhuno. i mean, if I put a fragmention in an "in-reply-to" or "like-of" my site will go retrieve that part of the original post to display in reply-context