IRC logs for #respimg, 2014-04-23 (GMT)

[00:10:12]<eeeps>zcorpan, TabAtkins: 100x50 is bad because it isn’t clear which is the width and which is the height. people use axb notation a lot informally in the art world and it’s rarely consistent.
[00:22:02]* johns has quit ("No Ping reply in 180 seconds.")
[00:26:04]* johns has joined #respimg
[00:29:29]* zcorpan has joined #respimg
[00:46:45]<TabAtkins>Really? Weird. I would have thought that WxH was obvious.
[01:08:27]* zcorpan has joined #respimg
[01:11:48]<eeeps> says always put height before width, and a cursory check shows that’s what the Met and the MoMA do, so that’s probably in a style manual somewhere.
[01:12:09]<eeeps>A few smaller gallery sites I just checked consistently do WxH
[01:12:21]<eeeps>I know I’ve seen BiggerValue x SmallerValue at least once
[01:12:40]<eeeps>and I’d wager that on artist’s personal sites it’s mostly randomness
[01:21:35]<eeeps>the two most common uses in the US are probably 8.5x11 paper (WxH) and photo prints, eg 4x6, 5x7, 8x10 (mostly landscape, so the opposite, HxW).
[01:21:48]<eeeps>also two-by-fours (orientation undefined)
[01:48:24]* Wilto has joined #respimg
[02:03:42]* eeeps has quit (eeeps)
[02:09:12]* zcorpan has joined #respimg
[02:30:18]* Wilto has joined #respimg
[02:41:33]<TabAtkins>True that, eeeps.
[03:10:00]* zcorpan has joined #respimg
[04:02:59]* marcosc has joined #respimg
[04:10:45]* zcorpan has joined #respimg
[05:11:40]* zcorpan has joined #respimg
[05:13:31]* zcorpan_ has joined #respimg
[05:52:02]* marcosc has joined #respimg
[06:14:13]* zcorpan has joined #respimg
[06:42:52]<yoav>TabAtkins: I don't think that adding more 'x' descriptors would be a good idea
[06:43:42]<yoav>Makes room for subtle errors such as "100x50.jpg 100x 50, 200x100.jpg 200x 100"
[06:43:53]<yoav>authoring errors, that is
[06:45:03]<yoav>We could rename 'w' as 'rw' (to stand for resource width) fairly easily, if we are concerned with author confusion around 'w'
[07:08:51]* Kolombiken has quit ("Leaving.")
[07:11:00]* Kolombiken has joined #respimg
[07:14:59]* zcorpan has joined #respimg
[07:41:03]* marcosc has joined #respimg
[08:15:46]* zcorpan has joined #respimg
[09:16:32]* zcorpan has joined #respimg
[09:29:42]* marcosc has joined #respimg
[11:08:56]* zcorpan has joined #respimg
[11:18:34]* marcosc has joined #respimg
[11:39:36]* richt has joined #respimg
[11:40:52]* newtron_ has quit ("")
[11:43:55]* newtron has joined #respimg
[12:18:01]* TimWright has joined #respimg
[13:07:27]* marcosc has joined #respimg
[13:17:42]<yoav>zcorpan: around?
[13:18:00]<zcorpan>yoav: yeah. but in meeting
[13:18:11]<yoav>zcorpan: OK, talk to me when you can
[13:24:14]<zcorpan>i can talk now sporadically, can be more focused in ~2.5h
[13:24:29]<zcorpan>or ~2h
[13:25:35]* newtron_ has joined #respimg
[13:26:18]<yoav>zcorpan: I have some objections to the AxB syntax. I'll comment on the bug
[13:26:57]<zcorpan>yeah i just caught up with the backlog
[13:32:30]<yoav>zcorpan: From impl standpoint, it's all the same (both WebKit and Blink current impl will ignore 120x200, which is a good thing)
[13:33:07]<yoav>zcorpan: I just think we shouldn't be reusing 'x' for 2 distinct descriptors
[13:35:10]<yoav>zcorpan: And anyway, I think i won't try to ship srcset&sizes in the next cycle if this is a concern. would give us another 6 weeks to hash things out
[13:35:39]<newtron_>good morning
[13:35:46]<yoav>newtron_: morning
[13:44:51]<zcorpan>yoav: i guess we're not in a rush, but i didn't intend to delay shipping
[13:47:10]<yoav>zcorpan: you raised a valid concern, and we need to figure the best way to address it. We can (and should) wait on it to figure it out
[13:51:13]* Wilto has joined #respimg
[13:52:07]<TimWright>Hey guys, do you want me to talk about Picturefill 2 on my podcast this week or is it too soon?
[14:05:00]<Wilto>Totally okay by me, TimWright.
[14:05:15]<Wilto>I’m worried about how few bug reports we’ve gotten from the masses, and I am no optimist.
[14:05:27]<Wilto>The more eyes we can get on this sucker the better.
[14:05:48]<newtron_>Wilto: it's still labelled as alpha, right?
[14:06:16]<TimWright>Wilto: I'd assuming it's just the best written code ever. Probably safe.
[14:06:38]<Wilto>newtron_: Yeah. I’m hoping people are assuming it’s a mess at this point, and we’ll get more feedback on the beta.
[14:06:53]<Wilto>TimWright: shh don’t jinx it man
[14:10:36]<newtron_>Wilto: I'm ready for you to spin up that postpone repo whenever you are, but given the confusion around the naming, i don't think it should be called postpone
[14:10:43]<newtron_>maybe something like "deferred image loading"
[14:13:31]* uniqname has joined #respimg
[14:18:31]<yoav>Don't want to make you guys look bad, but eeeps already found a bug in the sizes/srcset impl
[14:19:07]<Wilto>newtron_: “Use Cases and Requirements for Deferred Image Loading?”
[14:19:28]<newtron_>Wilto: sgtm
[14:20:06]<zcorpan>yoav: in picturefill?
[14:21:05]<Wilto>In the real deal, I thought.
[14:21:19]<Wilto>Feel like I saw something about this yesterday, but my memory is fried.
[14:23:10]<yoav>zcorpan: In Blink
[14:23:20]<zcorpan>yoav: what's the bug?
[14:24:00]<yoav>patch at
[14:24:50]<zcorpan>yoav: was it the change we made to the spec to not add src="" as a candidate when there are 'w' descriptors?
[14:25:49]<zcorpan>not quite since the test has the same url in srcset
[14:29:36]<zcorpan>scrap the above line
[14:34:45]<Wilto>Alright, I’m calling it: launches today.
[14:37:01]* marcosc has joined #respimg
[14:54:16]* zcorpan has joined #respimg
[15:10:04]<yoav>zcorpan: Simply a bug :)
[15:22:11]* zcorpan has joined #respimg
[15:39:22]<zcorpan>yoav: here now btw
[15:40:49]* zcorpan 101 switching trains
[15:43:04]* richt has joined #respimg
[15:50:49]* zcorpan has joined #respimg
[15:51:05]* zcorpan back
[15:53:49]<marcosc>TabAtkins: so, I'm still not sure why we didn't get Hixie or hober to fix the srcset algo in HTML? (irrespective of picture)
[15:53:57]<marcosc>or is that not possible?
[15:54:20]<TabAtkins>marcosc: It came up in the middle of the discussion over src-n, so it probably just got lost.
[15:54:30]<zcorpan>fix what?
[15:54:48]<marcosc>dropping the "h" bit, I guess
[15:55:01]<marcosc>just wanting to avoid duplication
[15:55:06]<TabAtkins>Huh? Oh, no, I thought you were talking about something totally different.
[15:55:46]<TabAtkins>The HTML srcset parsing algorithm is broken in some ways, in that it fails to be usefully forward-compatible. I'd have to look up my older email about it to see what the exact problem was.
[15:56:44]<zcorpan>TabAtkins: i'm interested in hearing about that
[15:56:52]<marcosc>seems sad that we can't fix srcset in the HTML spec and we have to replicate it in <picture>. I guess it's easier, so doesn't matter too much.
[15:57:08]<zcorpan>TabAtkins: was it that it wouldn't be compatible to introduce a higher-level grouping separator like || ?
[15:57:15]<TabAtkins>marcosc: <picture> is going to be merged into HTML later anyway.
[15:57:28]<TabAtkins>zcorpan: I can't quite remember. I'll look up the email.
[15:57:56]<marcosc>TabAtkins: yeah, true.
[15:58:08]<marcosc>sorry for the digression
[16:06:11]* eeeps has joined #respimg
[17:19:14]<zcorpan>TabAtkins: i can fix it. (we're getting off-topic for #css)
[17:24:29]<zcorpan>TabAtkins: hmm. "FATAL ERROR: No 'element' refs found for 'picture'.""
[17:24:43]<TabAtkins>update your bikeshed
[17:25:32]<TabAtkins>(I recently added support for <pre class=elementdef>, and so I removed the explicit <dfn element> from the spec.)
[17:26:07]<zcorpan>that fixed it
[17:28:06]* marcosc has joined #respimg
[17:28:26]* Lagann has joined #respimg
[17:28:26]<Lagann>[13picture-element] 15zcorpan opened pull request #163: Use # in the sizes grammar instead of invalid #+. Fixes #162 (06gh-pages...06fix-162) 02
[17:28:26]* Lagann has left #respimg (Lagann)
[17:34:05]<zcorpan>TabAtkins: r? ^
[17:34:48]* Lagann has joined #respimg
[17:34:48]<Lagann>[13picture-element] 15tabatkins closed pull request #163: Use # in the sizes grammar instead of invalid #+. Fixes #162 (06gh-pages...06fix-162) 02
[17:34:48]* Lagann has left #respimg (Lagann)
[17:34:49]* Lagann has joined #respimg
[17:34:49]<Lagann>[13picture-element] 15tabatkins pushed 2 new commits to 06gh-pages: 02
[17:34:49]<Lagann>13picture-element/06gh-pages 14e4c3c83 15Simon Pieters: Use # in the sizes grammar instead of invalid #+. Fixes #162
[17:34:49]<Lagann>13picture-element/06gh-pages 149102786 15tabatkins: Merge pull request #163 from zcorpan/fix-162...
[17:34:49]* Lagann has left #respimg (Lagann)
[17:36:53]* marcosc has joined #respimg
[18:00:13]* marcosc has joined #respimg
[18:24:22]* zcorpan has joined #respimg