/irc-logs / w3c / #webapps / 2013-12-05 / end
Options:
- # Session Start: Thu Dec 05 00:00:00 2013
- # Session Ident: #webapps
- # [00:00] * Quits: ken (~ken@public.cloak) (Ping timeout: 180 seconds)
- # [00:20] * Quits: Hixie (~ianh@public.cloak) (Ping timeout: 180 seconds)
- # [00:24] * Joins: Hixie (~ianh@public.cloak)
- # [00:27] * Quits: tantek (~tantek@public.cloak) (tantek)
- # [00:38] * Quits: ArtB (~abarsto@public.cloak) ("Leaving.")
- # [00:49] * Joins: tantek (~tantek@public.cloak)
- # [00:53] * Joins: ken (~ken@public.cloak)
- # [00:57] * Quits: tantek (~tantek@public.cloak) (Ping timeout: 180 seconds)
- # [01:00] * Joins: tantek (~tantek@public.cloak)
- # [01:00] * Quits: ken (~ken@public.cloak) (Ping timeout: 180 seconds)
- # [01:07] * Quits: tantek (~tantek@public.cloak) (tantek)
- # [01:14] * Joins: tantek (~tantek@public.cloak)
- # [01:34] * Quits: lmcliste_ (~lmclister@public.cloak) ("")
- # [01:42] * Quits: tantek (~tantek@public.cloak) (Ping timeout: 180 seconds)
- # [01:49] * Joins: tantek (~tantek@public.cloak)
- # [01:57] * Joins: ken (~ken@public.cloak)
- # [02:01] * Quits: sicking (~sicking@public.cloak) (sicking)
- # [02:02] * Quits: tantek (~tantek@public.cloak) (tantek)
- # [02:02] * Quits: jsbell (~jsbell@public.cloak) ("There's no place like home...")
- # [02:02] * Joins: jungkees (~jungkees@public.cloak)
- # [02:06] * Quits: ken (~ken@public.cloak) (Client closed connection)
- # [02:07] * Joins: sicking (~sicking@public.cloak)
- # [02:17] * Joins: real_wez (~realwez@public.cloak)
- # [02:17] * Quits: real_wez (~realwez@public.cloak) (Client closed connection)
- # [02:18] * Joins: tantek (~tantek@public.cloak)
- # [02:19] * Quits: tantek (~tantek@public.cloak) (tantek)
- # [02:36] * Joins: davidb (~davidb@public.cloak)
- # [02:36] * Quits: davidb (~davidb@public.cloak) (Client closed connection)
- # [02:38] * Quits: smaug (~chatzilla@public.cloak) (Ping timeout: 180 seconds)
- # [02:42] * Joins: ken (~ken@public.cloak)
- # [02:59] * Quits: skddc (~anonymous@public.cloak) (Ping timeout: 180 seconds)
- # [03:24] * Quits: sicking (~sicking@public.cloak) (sicking)
- # [03:25] * Quits: ken (~ken@public.cloak) (Ping timeout: 180 seconds)
- # [03:41] * Quits: Hixie (~ianh@public.cloak) (Client closed connection)
- # [03:41] * Joins: Hixie (~ianh@public.cloak)
- # [04:04] * Joins: ken (~ken@public.cloak)
- # [04:16] * Quits: ken (~ken@public.cloak) (Client closed connection)
- # [04:20] * Joins: sicking (~sicking@public.cloak)
- # [04:28] * heycam is now known as heycam|away
- # [05:15] * Quits: lgombos (~gombos@public.cloak) (Ping timeout: 180 seconds)
- # [05:19] * Joins: lgombos (~gombos@public.cloak)
- # [05:25] * heycam|away is now known as heycam
- # [05:27] * Joins: ken (~ken@public.cloak)
- # [05:34] * Quits: ken (~ken@public.cloak) (Ping timeout: 180 seconds)
- # [05:43] * Joins: ken (~ken@public.cloak)
- # [06:45] * heycam is now known as heycam|away
- # [06:46] * Quits: jungkees (~jungkees@public.cloak) ("Page closed")
- # [06:47] * Quits: kochi (~kochi@public.cloak) ("Leaving.")
- # [07:11] * Quits: sicking (~sicking@public.cloak) (sicking)
- # [08:08] * Joins: shepazutu (schepers@public.cloak)
- # [08:12] * Quits: shepazu (schepers@public.cloak) (Ping timeout: 180 seconds)
- # [08:12] * shepazutu is now known as shepazu
- # [08:39] * Joins: shepazutu (schepers@public.cloak)
- # [08:44] * Quits: shepazu (schepers@public.cloak) (Ping timeout: 180 seconds)
- # [08:44] * shepazutu is now known as shepazu
- # [10:28] * Quits: ken (~ken@public.cloak) (Client closed connection)
- # [10:39] * Joins: Lachy (~Lachy@public.cloak)
- # [10:58] * Quits: Lachy (~Lachy@public.cloak) ("My MacBook Pro has gone to sleep. ZZZzzz…")
- # [11:08] * Joins: Lachy (~Lachy@public.cloak)
- # [11:10] * Quits: Lachy (~Lachy@public.cloak) ("My MacBook Pro has gone to sleep. ZZZzzz…")
- # [11:25] * Joins: ken (~ken@public.cloak)
- # [11:32] * Joins: smaug (~chatzilla@public.cloak)
- # [11:40] * Joins: Lachy (~Lachy@public.cloak)
- # [11:49] * Quits: ken (~ken@public.cloak) (Client closed connection)
- # [11:59] * Joins: skddc (~anonymous@public.cloak)
- # [12:26] * Joins: abarsto (~abarsto@public.cloak)
- # [12:26] * abarsto is now known as ArtB
- # [12:39] * Quits: ArtB (~abarsto@public.cloak) ("Leaving.")
- # [12:39] * Joins: abarsto (~abarsto@public.cloak)
- # [12:39] * abarsto is now known as ArtB
- # [12:53] * Quits: Lachy (~Lachy@public.cloak) (Client closed connection)
- # [12:53] * Joins: Lachy (~Lachy@public.cloak)
- # [13:08] * Quits: Lachy (~Lachy@public.cloak) ("My MacBook Pro has gone to sleep. ZZZzzz…")
- # [13:11] * Joins: Lachy (~Lachy@public.cloak)
- # [13:12] * Quits: Lachy (~Lachy@public.cloak) ("My MacBook Pro has gone to sleep. ZZZzzz…")
- # [13:14] * Joins: Lachy (~Lachy@public.cloak)
- # [13:16] * Quits: Lachy (~Lachy@public.cloak) ("My MacBook Pro has gone to sleep. ZZZzzz…")
- # [13:16] * Joins: Lachy (~Lachy@public.cloak)
- # [13:57] * Joins: dom (dom@public.cloak)
- # [14:05] * Quits: smaug (~chatzilla@public.cloak) (Ping timeout: 180 seconds)
- # [14:14] * Joins: kawai (~kawai@public.cloak)
- # [14:17] * Quits: lgombos (~gombos@public.cloak) (Ping timeout: 180 seconds)
- # [14:28] * ArtB changes topic to 'WebApps WG; this channel is logged http://krijnhoetmer.nl/irc-logs/'
- # [14:36] * Joins: karl (~karlcow@public.cloak)
- # [14:47] * Quits: MikeSmith (~MikeSmith@public.cloak) (Ping timeout: 180 seconds)
- # [14:49] * Joins: shepazutu (schepers@public.cloak)
- # [14:53] * Quits: shepazu (schepers@public.cloak) (Ping timeout: 180 seconds)
- # [14:53] * shepazutu is now known as shepazu
- # [14:54] * Joins: MikeSmith (~MikeSmith@public.cloak)
- # [15:09] * Joins: smaug (~chatzilla@public.cloak)
- # [15:29] * Joins: lgombos (~gombos@public.cloak)
- # [15:33] * Quits: kawai (~kawai@public.cloak) (Client closed connection)
- # [15:34] * Quits: lgombos (~gombos@public.cloak) (Client closed connection)
- # [15:45] * Joins: davidb (~davidb@public.cloak)
- # [15:47] * Joins: lgombos (~gombos@public.cloak)
- # [15:47] * Joins: ken (~ken@public.cloak)
- # [16:03] * Joins: glenn (~gadams@public.cloak)
- # [16:15] * Quits: dom (dom@public.cloak) (Ping timeout: 180 seconds)
- # [16:16] * Joins: dom (dom@public.cloak)
- # [17:10] * Quits: glenn (~gadams@public.cloak) ("Leaving...")
- # [17:19] * Quits: smaug (~chatzilla@public.cloak) (Ping timeout: 180 seconds)
- # [17:38] * Joins: lmcliste_ (~lmclister@public.cloak)
- # [17:54] * Joins: kawai (~kawai@public.cloak)
- # [18:02] * Quits: Domenic_ (~sid10976@public.cloak) (Client closed connection)
- # [18:02] * Joins: Domenic_ (~sid10976@public.cloak)
- # [18:53] * Quits: ken (~ken@public.cloak) (Client closed connection)
- # [18:55] * Joins: smaug (~chatzilla@public.cloak)
- # [19:05] * Quits: kawai (~kawai@public.cloak) (Client closed connection)
- # [19:30] * Quits: Lachy (~Lachy@public.cloak) ("My MacBook Pro has gone to sleep. ZZZzzz…")
- # [19:30] * Quits: ArtB (~abarsto@public.cloak) ("Leaving.")
- # [19:51] * Joins: sgalineau (~sgalineau@public.cloak)
- # [19:55] * Joins: abarsto (~abarsto@public.cloak)
- # [19:55] * abarsto is now known as ArtB
- # [20:03] * Joins: ken (~ken@public.cloak)
- # [20:10] * Quits: ken (~ken@public.cloak) (Ping timeout: 180 seconds)
- # [20:20] * Quits: davidb (~davidb@public.cloak) (davidb)
- # [20:23] * Quits: chaals (~Adium@public.cloak) ("Leaving.")
- # [20:38] * Quits: dom (dom@public.cloak) ("")
- # [20:51] * Joins: jsbell (~jsbell@public.cloak)
- # [20:51] * Joins: sicking (~sicking@public.cloak)
- # [20:51] <sicking> marcosc: ping
- # [21:14] * Quits: lgombos (~gombos@public.cloak) (Ping timeout: 180 seconds)
- # [21:17] * Quits: sicking (~sicking@public.cloak) (sicking)
- # [21:24] * heycam|away is now known as heycam
- # [21:35] * Joins: lgombos (~gombos@public.cloak)
- # [22:00] * Joins: sicking (~sicking@public.cloak)
- # [22:01] * Joins: davidb (~davidb@public.cloak)
- # [22:58] <marcosc> sicking: pong
- # [22:58] <sicking> marcosc!
- # [22:58] <marcosc> SIIIICKKKING!!! :D
- # [22:58] <sicking> marcosc: what's your target stable date for the manifest? Do you think Q1 is realistic?
- # [22:59] <marcosc> I think so, we are getting close to having the feature set nailed down
- # [22:59] <sicking> marcosc: or rather, do you think we can make Q1 if we drop complex use cases (the cookie stuff comes to mind)
- # [22:59] <sicking> cool
- # [22:59] <sicking> marcosc: your gist was pretty confusing btw. I don't think people understood that you had two separate questions in here
- # [23:00] <marcosc> I only really had one question, but I think people naturally went "this has to be external"
- # [23:01] <sicking> i think people read it as "do we want external or internal" and of course the answer is then external
- # [23:01] <sicking> the option that both could be allowed didn't seem clear
- # [23:01] <sicking> marcosc: one argument for only allowing external manifests is that inline manifests make it tricky for websites to transition into a multi-page app
- # [23:01] <marcosc> yeah, I tried to clarify that... not sure if I succeeded
- # [23:02] <marcosc> what I was primarily interested in was about the json in an attribute thing
- # [23:02] <sicking> marcosc: well.. people answering <link> didn't really answer that question, right?
- # [23:03] <marcosc> well, in some way yes
- # [23:03] <marcosc> but in other ways no
- # [23:03] <marcosc> But those that answered it were relevant. Robin was going to add something today
- # [23:04] <sicking> ok
- # [23:04] <marcosc> I asked Hixie, but he doesn't understand
- # [23:04] <marcosc> the use case
- # [23:05] <Hixie> it's not that i don't understand it, it's that i haven't found where it's described
- # [23:05] <Hixie> i've no idea if i'd understand it or not :-)
- # [23:06] <marcosc> Hixie: the question is really just about if it's ok to put JSON into a the content attribute of meta?
- # [23:07] <Hixie> oh, i thought the question was "is this spec ok"
- # [23:07] <Hixie> hmm
- # [23:07] <Hixie> it's a bit weird to put json there, but not the weirdest thing ever
- # [23:07] <Hixie> probably indicative of an underlying more serious problem though
- # [23:08] <marcosc> Hixie: could be
- # [23:08] <Hixie> how much json are we talking about?
- # [23:08] <Hixie> why does it have to be json?
- # [23:08] <Hixie> what's the data structure?
- # [23:08] <Hixie> how is it used?
- # [23:09] <Hixie> generally i'd warn off using <meta> at all, for browser stuff
- # [23:09] <Hixie> is it browser stuff?
- # [23:09] <marcosc> http://manifest.sysapps.org/#usage-examples
- # [23:09] * Joins: Lachy (~Lachy@public.cloak)
- # [23:09] <Hixie> (i love how that spec calls itself simultaneously a standard and a draft)
- # [23:10] <marcosc> Hixie: yeah... Respec is trying to force my hand :(
- # [23:10] <marcosc> Hixie: will fix that soon
- # [23:10] <Hixie> if that blob in "EXAMPLE 1" is representative of what you want to apear in an attribute, i'd probably raise an eyebrow at it
- # [23:11] <marcosc> Hixie: btw, the problem we are trying to solve is here http://w3c-webmob.github.io/installable-webapps/
- # [23:11] <marcosc> Hixie: example 1 is fairly common
- # [23:12] <Hixie> i disagree with "the web platform lacks a standardized means for a web application to indicate that it can be "installed" and treated as if it were an native application"
- # [23:12] <Hixie> the standardised means of indicating that a page can be installed is for the page to exist
- # [23:13] <marcosc> true, that's badly worded
- # [23:13] <marcosc> will fix
- # [23:14] <Hixie> what's the difference between an application "function as standalone" and not?
- # [23:15] <marcosc> mostly that it runs outside a browser
- # [23:15] <Hixie> sorry, i poorly pasted that
- # [23:15] <Hixie> i meant, what kind of web pages aren't able to function like that?
- # [23:15] <marcosc> ones that rely on a back button
- # [23:16] <Hixie> oh, because iOS has no back button?
- # [23:16] * Quits: davidb (~davidb@public.cloak) (davidb)
- # [23:16] <marcosc> yeah
- # [23:16] <marcosc> and they don't show any browser chrome for standalone apps
- # [23:17] <Hixie> couldn't you just slide in some back button ui once there's a session history?
- # [23:17] <marcosc> sure, but people seem to want to be able to have control of that
- # [23:18] <marcosc> we do that in FxOS already
- # [23:18] <marcosc> (bookmark to home screen gives you access to minimal UI)
- # [23:18] <Hixie> don't they have control, by, like, not adding a session history?
- # [23:18] <Hixie> e.g. using location.replace() instead of location.assign() ?
- # [23:19] <marcosc> I guess, but that means a lot of work for devs in that they have to capture all click events, etc
- # [23:20] <Hixie> if an app has that much navigation, it seems like maybe they should be happy with the browser exposing a back button...
- # [23:20] <marcosc> Hixie: see http://w3c-webmob.github.io/installable-webapps/#ios7-safari-1
- # [23:20] <Hixie> 90% of the things on the requirements list here are things HTML already does, fwiw
- # [23:20] <marcosc> Hixie: yeah, basically... the only advantage is externalizing it
- # [23:21] <marcosc> if that is an advantage, that is
- # [23:21] <Hixie> seems pretty easy to catch all navigations, btw
- # [23:21] * Quits: smaug (~chatzilla@public.cloak) (Client closed connection)
- # [23:21] <Hixie> i mean you just need an event listener
- # [23:21] <Hixie> much like in that example that you cited just now
- # [23:22] <marcosc> ah, yes. Ideally, one can create an app that can be navigated without requiring any script
- # [23:22] <marcosc> Hixie: Chrome Beta for Android was allowing this
- # [23:22] <marcosc> (they yanked the add to home screen feature a few days ago from the beta)
- # [23:22] <Hixie> sure, i'm just talking about if you also want to hide the default back button which would come in when you got a session history
- # [23:23] <Hixie> anyway, what's the manifest got to do with this? doesn't service workers and/or appcache and/or anne's "zip" scheme already solve the bundling issue?
- # [23:23] <marcosc> I guess I'd have to try this out... it feels like managing navigation history to avoid the UI popping up would be a lot of work
- # [23:24] <Hixie> why would it be a lot of work?
- # [23:24] <Hixie> you just catch clicks on links, and use history.replace() instead
- # [23:24] <Hixie> it's almost identical to that script you pointed out
- # [23:24] <marcosc> well, what if you navigated forward a few links, then wanted to go back 3 times?
- # [23:25] <marcosc> you would need some kind of little history db
- # [23:25] <Hixie> oh you want to hide the browser's back button but then provide one yourself?
- # [23:25] <Hixie> that just seems like a lot of pointless work
- # [23:25] <Hixie> what's wrong with the browser's?
- # [23:26] <Hixie> in practice, native apps don't do that, anyway
- # [23:26] <marcosc> yeah, that's what FxOS gets people to do... it proved too much work, so they added a "chrome" manifest member
- # [23:26] <Hixie> they just have one back button per page
- # [23:29] <marcosc> yes, I guess that's right
- # [23:32] <marcosc> I'll shop that idea around a bit, see what potential implementers think
- # [23:37] * Quits: karl (~karlcow@public.cloak) (":tiuQ tiuq sah woclrak")
- # [23:38] <marcosc> Hixie: oh, the last question I had was, if implementers won't budge on using JSON for this, is using <script type="application/manifest+json"> a better alternative to using <meta name="manifest" content='{ ...json... }'>?
- # [23:38] <Hixie> i don't understand the manifest part of this at all
- # [23:38] <Hixie> <Hixie> anyway, what's the manifest got to do with this? doesn't service workers and/or appcache and/or anne's "zip" scheme already solve the bundling issue?
- # [23:39] <marcosc> it seems vendors want a better alternative to the metadata ... that other stuff is separate from just giving an application a name
- # [23:40] <marcosc> also, there is talk of hooking up the service worker through the JSON manifest
- # [23:41] <marcosc> {"serviceWorker": {"url":"worker.js", urlSpace: "foo/*"} }
- # [23:41] <marcosc> or something
- # [23:41] <Hixie> so, we're back to "i don't understand the use case"
- # [23:41] <Hixie> what's the manifest?
- # [23:41] <Hixie> what is it manifesting?
- # [23:41] <Hixie> i had assumed it was the bundling issue
- # [23:42] <marcosc> right now, it's just for name and icons
- # [23:42] <Hixie> that's already solved in HTML. let's not add more redundant ways to do that.
- # [23:42] <marcosc> yeah, I said the same thing
- # [23:42] <Hixie> <meta name="application-name"> and <link rel=icon>
- # [23:42] <Hixie> the latter of which has sizes="" for multiple icons
- # [23:43] <marcosc> is sizes in the spec yet?
- # [23:43] <Hixie> and the former of which has lang="" if you need localisation
- # [23:43] <Hixie> sizes="" has been in HTML for years
- # [23:43] <marcosc> ah, nice
- # [23:43] <Hixie> 2008
- # [23:44] <Hixie> (same as application-name)
- # [23:44] <marcosc> I agree that all the bits are in place. I'll try to find out why those are not sufficient enough. I haven't been able to get a good answer for that
- # [23:45] <marcosc> I think people just think it's easier to deal with the JSON
- # [23:45] <marcosc> as you don't have to parse the HTML and it's not dynamic
- # [23:46] <Hixie> you still have to parse the HTML to get the JSON!
- # [23:46] <Hixie> and then you have to parse the JSON
- # [23:46] <Hixie> that's _more_ parsing
- # [23:46] <Hixie> plus, you have to define how to handle misstructured JSON
- # [23:46] <Hixie> which you get for free on the HTML side
- # [23:46] <Hixie> as far as the dynamic thing goes, if it's obtained via the HTML file, it's exactly as dynamic as the existing solutions
- # [23:55] <Domenic_> marcosc: I feel like I was pro-JSON until you actually wrote down how it would work. But now it seems silly.
- # [23:57] <marcosc> Domenic_: what do you mean?
- # [23:58] <Domenic_> Well all the practical problems of inline vs. external vs. the existing meta tags, and parsing and the preloaded and so on, makes it feel less of an obvious solution now.
- # Session Close: Fri Dec 06 00:00:01 2013
The end :)