2016-12-21 UTC
# 00:00 tantek so if you take / post photos of similar things/events (group workouts, HWC meetups, etc.) then person-tagging gets more automated with each post
# 00:00 tantek is anyone else manually person-tagging who would find that useful?
KevinMarks joined the channel
# 00:00 tantek (figuring that both 1 and 2 can be extracted / written as separable functionality)
# 00:04 aaronpk that method of recommendations sounds much easier to implement than face recognition
# 00:04 aaronpk actually kind of sounds like how gmail recommends additional recipients
# 00:05 tantek no need for any fancy/trendy machine-learning / AI hoo-haw
nitot joined the channel
# 00:09 tantek 3. implement receiving "tag-of" responses and incorporate them as person-tags, using the person-tagging features of silos that I POSSE the photos to as a UI instead, and then have Bridgy backfeed those person-tagging as synthetic tag-of responses.
# 00:15 Loqi [Ryan Barrett] PESOS for Bridgy Publish
# 00:17 tantek not going to push on that feature until I've actually implemented handling tag-of responses myself
aegibsonme joined the channel
# 00:32 tantek (tested it myself before updating the wiki :) )
wolftune, nitot, gkbrk, snarfed, Pierre-O, tbbrown, mlncn and leg joined the channel
gRegorLove, gRegorLove_, nitot, wolftune, Pierre-O, begriffs, snarfed and tantek joined the channel
nitot and KevinMarks joined the channel
# 05:33 KartikPrabhu what should a UI for generating a fragmention from selected text do if the selection is across multiple paragraphs?
# 05:38 Loqi A fragmention is an extension to URL syntax that links and cites a phrase within a document by using a URL fragment consisting of the phrase itself, including whitespace https://indieweb.org/fragmention
# 05:39 KevinMarks It is tricky because of the tension between the selected text being long and coherent versus being just enough to identify that paragraph
# 05:41 KartikPrabhu also the other Challenges should be updated with their now-accepted solutions
# 05:42 tantek KartikPrabhu: you're talking about a challenge within the context of a specific UI
# 05:42 tantek might make more sense inside a specific section for that UI inside Brainstorming
# 05:43 KartikPrabhu ok does a new section called Fragmention Generating UI work? I can ass adactio's book example too
# 05:46 Loqi Ok, I'll tell them that when I see them next
# 05:46 Loqi Ok, I'll tell them that when I see them next
wolftune joined the channel
awolf joined the channel
# 05:51 tantek KartikPrabhu: you can either <nowiki> it and document when it was noticed that the link died, or (more work) dig up an archive.org link to the last functional version
nitot and awolf joined the channel
# 06:10 KevinMarks Github pages for notes no longer ideal - 10 posts an hour limit
KevinMarks_, awolf, nitot and tantek joined the channel
# 07:07 Zegnat good morning IndieWeb, sknebel, KartikPrabhu!
# 07:17 Zegnat I am just reading up on the discussion, KartikPrabhu.
cweiske, tantek and awolf joined the channel
# 07:26 Zegnat Almost caught up on logs. I need a daily-digest-chat-log, I feel ;)
awolf joined the channel
# 07:32 Zegnat KartikPrabhu: do you still need confirmation of the select bug on different OS?
# 07:35 Zegnat I am currently at work, so I can test any W10 browsers. Would it be worth it to add a table to your test page with test results?
Pierre-O joined the channel
# 07:47 Zegnat Lovely, Edge does not let you copy the contents of alert()
# 07:49 Zegnat Neither does Chrome. What's with all this blocking of alert box copying?!
# 07:50 Zegnat Strike that. Chrome did seem to work. Maybe just Edge then
jihaisse joined the channel
# 07:54 Zegnat I didn't bring my MacBook, so I can't verify those results at the moment
snarfed and tantek joined the channel
# 07:57 Zegnat Updated the gist with a link to your page and noting that its W10
# 08:02 KartikPrabhu the spec doesnot say what the actual result is supposed to be so who knows which browsers are at fault here
Pierre-O, dkm_, pindonga and Zegnat joined the channel
# 08:20 Zegnat Feels like an omission in the spec, if they don't specify that the full UT8 should be returned
# 08:21 Zegnat Going to lose this nickname in a few seconds.... ugh
# 08:25 Zegnat KartikPrabhu, so should a spec bug be filed before browser bugs? Or browser bugs before spec bug?
cuibonobo, friedcell and rrix joined the channel
# 08:42 Zegnat KartikPrabhu: not a proper dev? Anyone who detects browser implementation differences and then writes succinct tests for them is a proper dev IMHO
# 08:48 sknebel Regarding the fragmention examples, just for my understanding: the entire document is selected in the first example because it highlights the element that contains the entire selected text?
# 08:48 KartikPrabhu the fragmention.js script picks the nearest ancestor with the whole selection
# 08:49 cweiske if yes: the font size is pretty large on your blog
# 08:50 sknebel KartikPrabhu: sounds like for that case you'd want to highlight all the children of that nearest ancestor that contain parts of the selection instead?
# 08:50 Zegnat It's not really aboout highlighting though, more about where to put .focus()
# 08:51 Zegnat And that's limited to a single element, by definition
# 08:52 KartikPrabhu Zegnat: I don't think the fragmention.js script switches focus at the moment; but that also is a valid expectation, please document it on /fragmention
# 08:53 KartikPrabhu cweiske: what I meant was you could always change the scaling in your own browser
# 08:53 cweiske oh. my browser lets me specify the default font size
# 08:54 Zegnat Wait, it does not put .focus()? I thought that was what it was for :o
# 08:54 cweiske but they seldom do, and most people don't know about that setting
# 08:55 KartikPrabhu Zegnat: fragmention.js does detect the element and put a special attribute "fragmention" on it, I suppose putting focus would be a good enhancement
# 08:55 KartikPrabhu cweiske: if CSS is used it should be respected by the brwoser settings no?
# 08:56 Zegnat Is there a reference fragmention implementation? E.g. a repository for fragmention.js?
# 08:57 cweiske "font-size: 100%" is the size that's set in the browser settings
# 08:57 cweiske if you use px or pt, then you're not respecting the browser setting
# 09:00 Zegnat KartikPrabhu, thanks. So it does element.scrollIntoView(), but as soon as I were to start TABing around I go to the top of the page again because focus was not set to the element
# 09:02 KartikPrabhu would really like feedback on this since I am trying to improve the UX of fragmentions in general
# 09:02 Zegnat sknebel, in some browsers I believe you can. Or you can make it focusable. I am looking to see if I have the code left. Pretty certain I wrote something for it
# 09:03 sknebel Zegnat: apparently you can set tabindex to something != -1 and it makes things focusable. TIL
# 09:05 Zegnat KartikPrabhu: can't currently update the wiki or post to GitHub. Don't have my password vault on my work PC
# 09:06 KartikPrabhu Zegnat: anytime you are free would work. I think fragmentions is a great idea and so would like to see it tested and improved up on
# 09:08 Zegnat I also wonder where screenreaders would expect to be dropped into the document. Probably on the nearest ancestor P. That's also something to think about, if you always grab just the nearest ancestor you might end up on a STRONG in the middle of a sentence. While I would expect to be put on the P, i.e. at the start of the sentence the selection fall
# 09:09 Zegnat So, yeah, focus, and location screen readers expect to end up at. Those are the first two things that come to mind right now.
# 09:09 KartikPrabhu ooh good point! I have been thinking about a11y more these days but not in the context of fragmentions
# 09:09 Zegnat Will update wiki at some point this week. Probably tomorrow. I am starting my trip home to Sweden this afternoon.
# 09:09 Zegnat Or, if anyone else is working on it, just link to this conversation and copy what I wrote ;)
# 09:10 KartikPrabhu once it is documented on the wiki we can try to get the developer's eyes on it
cuibonobo joined the channel
# 09:28 Zegnat KartikPrabhu, I sometimes get a bit obsessive over specs. I don't often see people refer to element categories, which is too bad, because they can be very helpful in conversations like this one
# 09:30 KartikPrabhu Zegnat: yes someone has to do it, and I am horrible are reading these specs. DO any browsers actullay implement these spec categories?
# 10:15 Sebsel aaronpk I'll try to find out why XRay 500s on seblog.nl
pindonga joined the channel
# 10:39 Sebsel But they are not finished, I just post plaintext now, so I have the data
# 10:39 Sebsel Have to think of UI / microformats / posting-interface etc
# 10:40 Zegnat Just plain text documentation is Fine. That's what tantek has.
# 10:41 Sebsel But tantek at least has a standardised format!
# 10:42 Sebsel But I have plans for it, and I will document once I made some steps toward it :)
# 10:43 Sebsel I can also blog about the plans, I guess, but then I need to have my lang="*" fields ready as discussed in #indieweb-dev
# 10:44 Sebsel Having your own website is hard if you want to do all the things 'right' and by yourself
# 10:45 Zegnat That's probably why people see a barrier to entry here. The real thing to understand is that there is no "right". Publishing HTML is all you need to do to do it right.
# 10:46 petermolnar that soon results in too many choices, too many technologies, what is trendy, what is lame, and some will just walk away instead
# 10:50 Loqi [aaronpk] well you don't need microformats if you just want to publish something! you need microformats once you want someone/something else to consume it!...
# 10:50 Loqi cweiske has 66 karma in this channel (74 overall)
loicm, camerongray, friedcell1 and j4y_funabashi joined the channel
camerongray_, Pierre-O, userXYZ, friedcell, friedcell1, camerongray, friedcell2, tbbrown and hs0ucy joined the channel
# 14:25 aaronpk petermolnar: sebsel: I'd say just start simple. it's been surprisingly fun to build my new website by just writing plain HTML files, then slowly adding php includes, so that I focus on actually writing things instead of worrying about the tools https://w7apk.com
Sebsel joined the channel
# 15:01 sknebel from comment on HN it seems like they introduced this sometime earlier this year
# 15:02 sknebel and are probably not strictly enforcing it, since a popular project page with docs or demos should easily exceed 100k requests
tbbrown joined the channel
ehlovader joined the channel
# 15:13 sknebel different topic: can anyone recommend a good cheap DNS hoster? (really just DNS, domain should be kept with the current registrar)
snarfed and nitot joined the channel
# 15:17 aaronpk my main DNS is provided by Linode. I used to run my own DNS server but finally shut it donw.
# 15:18 cweiske i'm able to run my own dyndns because I self-host it
# 15:38 sknebel thanks everybody, lots of options to look through it seems ;)
# 16:04 aaronpk omg i love that quill shows a preview of the thing i'm replying to now
nitot and tbbrown joined the channel
# 16:31 martymcguire[m] aaronpk++ i really love that feature, too!
# 16:31 Loqi aaronpk has 444 karma in this channel (1158 overall)
nitot joined the channel
# 16:48 Sebsel what does it send? a nestled h-item? it should be prepared for that.
# 16:49 aaronpk the top-level object is an h-review, and there's a nested h-product in a property called "item"
# 16:50 Sebsel Ah and I reject everything that's not a h-entry.
# 16:51 Sebsel is it okay to have h-reviews in a h-feed? because that will end up happening if I whitelist h-review in my code.
# 16:52 aaronpk h-review is one of those funny ones where it actually shares most of the same properties as h-entry so maybe shouldn't have been its own top-level item, but came from a long history of microformats1 where it was a top-level item so here we are
# 16:53 aaronpk it does have a little more semantic difference between h-entry compared to something like an h-entry with a photo, plus the name is catchy (as demonstrated by the surprising number of tweets about it still), so it's not too bad that it's a new object type
tantek, awolf and rrix joined the channel
# 17:10 Loqi [indieweb] "Currently debating between #100DaysOfMusic or #100DaysOfIndieWeb for my #100Days project. #100DaysOfMusic - The goal..." on 2016-12-21
jmelesky and cweiske joined the channel
# 17:16 cweiske maybe set max-height on the preview with overflow:scroll-y
# 17:16 aaronpk i'll do that, plus also check for a "name" property and just use that instead
# 17:19 cweiske with the autologin options I can finally add "add a comment" button to my blog posts
tantek and leg joined the channel
nitot, snarfed, wolftune and mlncn joined the channel
nitot, wolftune, snarfed, KevinMarks and ehlovader joined the channel
# 19:06 bear responding to him ^^ in that way so others may see his original request
# 19:14 Loqi bear has 129 karma in this channel (170 overall)
# 19:29 bear oh - this is about making sure that PWA's are "offline first" ... wow, so not interested in that
# 19:31 sknebel also (ideally) completely orthogonal to the goals here
# 19:32 bear yea, I was drawn in to his talk of a decentralized web event
# 19:32 snarfed are you kidding? you don't want a posting UI that works offline or in spotty connectivity? save drafts, sync posts and edit offline, etc?!
# 19:32 Loqi [Ryan Barrett] Posting to the indieweb from your phone
# 19:33 bear I have never posted anything from my phone and likely never will
# 19:33 aaronpk offline first is a great goal. it just seems tangential to decentralized web
# 19:33 snarfed bear: ok. lots of people do though. and probably even more at least want to read on their phones
# 19:34 bear yea, I need to stop posting - later I'll explain why hearing anything from some folks in the nodejs community sends me straight to rant/rage
nitot, ehlovader, snarfed1 and KartikPrabhu joined the channel
# 21:01 petermolnar offline first: I'm with bear on this. offline web apps is a waste in it's current state. ( os layer, gui layer, browser layer, web layer, and I'm wondering why my CPU is baking plastic. )
# 21:01 petermolnar there are these things, like Thunderbird, text editors and stuff, out of the browser, which are, be definition, offline
# 21:02 petermolnar for example, photoshop can upload to many offline services just fine
# 21:03 KartikPrabhu petermolnar: offline first also applies to websites and not just apps
# 21:04 petermolnar and FF mobile has been doing it, so it reloads the cached when it loses connection
# 21:04 KartikPrabhu right, but with things like service workers you can even read a web book offline
# 21:07 petermolnar my old kindle 4 does not have the computing power to handle books like that
# 21:07 KartikPrabhu but I don't have a ebook reader. but I can still read adactio's book offline on my phone
nitot joined the channel
# 21:11 Guest90645 epubs are zip archives, so they do have a storage advantage over plain html. Only real advantage I see.
snarfed, wolftune and benborges joined the channel
# 21:29 bear my take on the type of offline reading your describing above is that it's a solution that is required only because phones don't have a well defined api to store content
# 21:30 bear so a browser based app cannot behave the same way as a desktop based app and that wrankles the nerves of browser app folk
# 21:31 bear that is a great question - my answer is tied up in my bias against browser "apps" I guess
# 21:31 bear I should probably just go sit in the luddit corner
# 21:31 KartikPrabhu bear: I am not concerned about "apps" anyway. Adactio's book is a good couter example
# 21:32 bear my gripe is this in a nutshell - why can I not view a web page, have the browser store the html, css and even some js
# 21:33 bear and make it so if I visit that web page when offline I just get a small icon that says i'm viewing an offline page
# 21:33 bear instead most sites won't even have visible text
# 21:33 bear and yes, service workers "solve" this - but really i'm asking why doesn't this already exist as a solved item in browsers
# 21:34 bear that's why i'm such a huge fan of static with a bit of js added for +1 features, but always have the core fallback of static
# 21:35 KartikPrabhu of course the browsers could do this by default, but service worker type things are a good start
# 21:36 bear I view service workers as another layer in the progressive stack really
# 21:36 bear static html - css - js for events - service layer to cache forms and the like - fully online
# 21:37 petermolnar "> always have the core fallback of static" we're relics from ~2006 when progressive enhancement meant static html first, then styling and just after that JS
# 21:37 KartikPrabhu bear: yes! they are another layer in the progressive enhancement stack
# 21:38 bear you can't do anything with a site unless you have 3 layers of js - heck, most of the service worker stack is js implemented
# 21:39 bear I should be able to click on a form button and have the browser say "hey, i'm going to cache this for you until you are online"
# 21:39 bear but now a days you don't even have forms
# 21:40 bear it's all dynamic html stuff that *requires* some javascript dom manipulation that is then sent to a backend using something other than a POST
# 21:40 petermolnar that bothers me as well, when I see a button but it's not a clickable action handler
# 21:40 bear heck, I would even remove that complaint if the data to submit was at least doable over a REST api
# 21:41 KartikPrabhu i agree that most of this JS stuff has been over-used but that doesnot mean that the JS tech is bad
# 21:41 bear and i'm not ranting against that because I'm stuck in 2006 either - entire swathes of internet communities cannot access those items because they use assisted devices to view pages
# 21:42 bear KartikPrabhu - completely agree, most of my ranting is targeted to the echo chamber that is "modern" web devs
# 21:43 bear who build things because they can and then toss them onto the internet and then yell back at you when you ask for it to be accessible
# 21:43 bear wow - I'm really in a grumpy mood today - I think i'm on my 10th rant
# 21:44 bear yea, I did that yesterday to cool off
# 21:44 bear added the beginnings of websub hub code
# 21:45 bear and added to my event thingy pushing of publication items to subscribers
# 21:46 bear so i'm now kinda waiting for aaronpk to get more websub.rocks items in place so I can continue my coding :)
snarfed joined the channel
# 21:52 cweiske bear, I totally follow you in respect to modern js-driven "development"
snarfed and pfefferle joined the channel
# 22:05 pfefferle !tell barnabywalters can you please add me to the indieweb namespace on packagist.org?
# 22:05 Loqi Ok, I'll tell them that when I see them next
wolftune, pindonga, tantek and snarfed joined the channel
wolftune, mlncn, snarfed and ehlovader joined the channel