| 03:33 | <Domenic> | Gahhhh. We really need a mode that turns off all authoring conformance requirements, for implementers to look at. |
| 04:11 | <MikeSmith> | Domenic: 100% agreed |
| 04:13 | <MikeSmith> | even me in making changes to the authoring-conformance parts of the spec, I feel a need to reduce the noise that creates for implementors trying to follow changes |
| 04:14 | <MikeSmith> | I try not to fiddle with any of it unless it is actually important (e.g., it is causing actual confusion for authors, or creating problems for me with teh checker implementation) |
| 04:16 | <MikeSmith> | and I still wonder whether we should consider separating out all the authoring-conformance stuff to a separate spec in the long term |
| 04:17 | <MikeSmith> | I think the jury is still out on whether the experiment of conflating authoring conformance requirements and UA requirements in the same spec has been a success |
| 04:18 | <MikeSmith> | I think even it is done as best as it possibly could be it still results in the spec being subpoptimal as a resource for authors |
| 04:18 | <MikeSmith> | because can cannot optimize the spec for both audiences |
| 04:18 | <MikeSmith> | and we should optimize it for implementors |
| 04:19 | <MikeSmith> | which means authors are always going to lose out a bit |
| 04:20 | <MikeSmith> | and UA implementors are going to continue to be distracted/confused by existing author-only requirements in teh spec, and any changes we make to add more o fthem |
| 04:22 | <MikeSmith> | I still plan to make time to (re)add the toggle we used to have that allowed Web developers to see the web-developer-only view |
| 04:23 | <MikeSmith> | but unless I’m confused, we do not have markup in the source to enable a similar thing to allow implementors to see an implementor-only view |
| 04:24 | <MikeSmith> | but hope I am just forgetting and we can actually do it with the current source |
| 08:40 | <tobie> | MikeSmith: have you considered having two builds? |
| 08:40 | <tobie> | Or JS toggles + cookies? |
| 08:41 | <annevk> | MikeSmith: pretty cool to finally obsolete the old DOCTYPEs |
| 10:46 | <mkwst> | annevk: Infra! Yay. :) |
| 10:46 | <mkwst> | I'm happy to see this. Thanks for putting it together! |
| 11:58 | <MikeSmith> | annevk: yeah I wish we could eliminate need for any DOCTYPE at all but I guess for the forseeable future we’re stuck with needing <!DOCTYPE html> at least |
| 11:59 | <MikeSmith> | tobie: about doing two builds, wattsi already does that |
| 11:59 | <MikeSmith> | our build tool |
| 11:59 | <MikeSmith> | I’m not sure there’s even an option for making wattsi *not* do that |
| 12:00 | <MikeSmith> | and we had a JS toggles + cookies thing before for dynamically switching between the “all” view and the “(web) developer” view |
| 12:03 | <MikeSmith> | and the problem we ran into before with publishing the web-developer version separately is that it kind of became an separate redesign project instead of just simply a version with the implementor-specific content stripped out |
| 12:04 | <MikeSmith> | anyway none of what we have now would enable us to generate an *implementor*-only view with all the authoring-conformance-only stuff stripped out |
| 12:04 | <MikeSmith> | because the authoring-conformance-only parts of the source are not actually marked up as such |
| 12:05 | <MikeSmith> | so enabling it would require a human to manually go through the entire spec and identifying the authoring-conformance-only parts and marking them up with class=authors-only (or whatever) |
| 12:06 | <MikeSmith> | as Hixie once did, going through the whole spec to mark up the implementor-only bits with class=impl (now class=no-dev) |
| 12:56 | <mkwst> | annevk: Would you accept patches to infra to bring in some of the string processing bits from HTML's https://html.spec.whatwg.org/#common-parser-idioms? I use a few in CSP and elsewhere. |
| 12:56 | <mkwst> | I imagine I'm not the only one. |
| 12:57 | <annevk> | mkwst: yeah I think so |
| 12:58 | <annevk> | mkwst: everything that's duplicated and has no better place is welcome |
| 12:59 | <mkwst> | Ok. |
| 13:03 | smaug | kicks wicg. such a horrible playground |
| 13:22 | MikeSmith | sends some positive vibes in smaug’s direction |
| 13:23 | <smaug> | yeah, sorry, I was too negative. But dumping random stuff there isn't working very well. |
| 13:25 | <MikeSmith> | smaug: yeah I hear you |
| 13:25 | <gsnedders> | "The WG doesn't like this, send it to WICG" |
| 13:25 | <MikeSmith> | not sure what to do to mitigate that dumping-random-stuff problem |
| 13:26 | <MikeSmith> | kind of an inevitable side effect of the inherent nature of it |
| 13:27 | <MikeSmith> | but "The WG doesn't like this, send it to WICG" is at least somewhat by design |
| 13:27 | <annevk> | tobie: I guess https://github.com/heycam/webidl/issues/240 also argues for creating some kind of longer term story for what we want out of IDL |
| 13:28 | <tobie> | annevk: I don't know if specifically #240 does, but it's clearly something we have to workout. |
| 13:28 | <tobie> | annevk: I'm glad you released the infra standard. That, at least, sets a lower boundary. |
| 13:30 | <tobie> | annevk: was contemplating referencing it from WebIDL pretty much as we speak. |
| 13:34 | gsnedders | still needs to stop typing the CSS colour property |
| 13:41 | <nox> | gsnedders: No you don't. |
| 13:41 | <nox> | gsnedders: Instead you must submit a patch to the spec to allow British spelling, obviously. |
| 13:42 | <tobie> | MikeSmith: I see. That sucks. |
| 13:42 | <annevk> | tobie: yeah, I think that would make sense |
| 13:43 | <annevk> | tobie: we updated biblio.json so referencing it probably already works if that is all still working |
| 13:43 | <annevk> | tobie: Shepherd hasn't been updated to index it yet so referencing any terms might still not work |
| 13:43 | <tobie> | annevk: http://www.specref.org/?q=Infra%20Standard |
| 13:44 | <annevk> | \o/ |
| 13:44 | <tobie> | annevk: it would be nice if shepherd did stuff hourly vs. daily |
| 13:44 | <tobie> | annevk: changes a lot |
| 13:45 | <annevk> | tobie: the main problem with Shepherd is that TabAtkins or plinss have to add and remove standard URLs someplace |
| 13:46 | <tobie> | annevk: they could just pull the list of new W3C and WHATWG standards from Specref |
| 13:47 | <gsnedders> | nox: post a PR next April 1, and leave everyone confused as to how genuine I am? |
| 13:47 | <annevk> | tobie: we don't want everything to be indexed by Shepherd basically |
| 13:48 | <nox> | gsnedders: For April 1st I want to support </sarcasm> in Servo. |
| 13:48 | <annevk> | tobie: e.g., W3C copies of stuff, older specifications such as Typed Arrays that are now obsolete |
| 13:48 | <nox> | gsnedders: TokenSink::take_a_deep_breath :p |
| 13:48 | <tobie> | annevk: you could be really un-pc and filter by editor name :D |
| 13:48 | <gsnedders> | nox: you are aware of the WebKit </sarcasm> bug? |
| 13:48 | <nox> | gsnedders: What's that? :O |
| 13:49 | <gsnedders> | nox: https://bugs.webkit.org/show_bug.cgi?id=45645 |
| 13:50 | <nox> | gsnedders: Ah ah ah ah ah ah. |
| 13:50 | <nox> | notImplemented(), gold. |
| 13:50 | <gsnedders> | "Um… yeah." |
| 13:51 | <nox> | gsnedders: But if we had take_a_deep_breath(), a screen reader could actually take a deep breath. |
| 13:52 | <gsnedders> | I don't think any of them support that, though |
| 13:52 | <nox> | Heh. |
| 13:52 | <nox> | gsnedders: Just transliterate the sound of a deep breath in the Latin alphabet and feed that to the screen reader. :p |
| 14:02 | <tobie> | annevk: editor name, pub date, plus publisher. I can add an API for this if there's interest and if it speeds up things. |
| 14:03 | <annevk> | tobie: I'm not sure I follow |
| 14:03 | <tobie> | annevk: filter which spec need to be integrated by Shepherd automatically. |
| 14:03 | <annevk> | tobie: if you're talking about Shepherd, it's best to file an issue against tabatkins/bikeshed |
| 14:04 | <tobie> | annevk: noted |
| 14:04 | <annevk> | tobie: maybe we can figure something out there with TabAtkins on how to improve things (and the more regular indexing you mentioned that I agree would be nice to have) |
| 14:06 | <younder> | What has happened to my amazon cite? It only shows text, will not accept ssh and is pretty much useless? Seems this is a chromium bug. 'Ubuntu' informs me that this is a probem with too strict conformance criteria in chrome. |
| 16:36 | <astearns> | gsnedders: the April 1st colour thing was done already (and some took it seriously) |
| 16:43 | <gsnedders> | astearns: level of surprise: low |
| 17:57 | <annevk> | astearns: isn't there a serious proposal for property localization every couple of years? |
| 17:58 | <astearns> | annevk: I haven't seen a serious one in the ~6 years I've been at this |
| 17:58 | <annevk> | astearns: well, back in my day... |
| 17:58 | <astearns> | but I may not have been paying enough attention |
| 18:06 | <gsnedders> | annevk: Proof you're old? :) |
| 18:36 | <annevk> | gsnedders: get off my lawn |
| 18:38 | <annevk> | astearns: found https://lists.w3.org/Archives/Public/www-style/2009Feb/thread.html#msg475 |
| 18:38 | <gsnedders> | annevk: hey, I'm positively in my mid-20s nowadays! |
| 18:38 | <annevk> | via https://annevankesteren.nl/2009/02/www-style-thursday by the way |
| 18:41 | <annevk> | Can't really find more quickly though |
| 18:42 | <astearns> | annevk: before my time, but sampling the thread it looks to me like all the 'serious' people were against considering the idea |
| 18:43 | <annevk> | astearns: oh yeah, I wouldn't take that blog post too serious |
| 18:43 | <annevk> | astearns: it never went anywhere, just came up a few times (although I can only find this one instance at the moment) |
| 18:45 | <astearns> | our very very serious @⚐ proposal should have taken off |
| 19:59 | <annevk> | MikeSmith: https://dev.w3.org/2006/webapi/WebIDL does not redirect |
| 20:03 | <smaug> | is there any tool to check usage of some html element? |
| 20:04 | <smaug> | I thought Google has something, but can't recall the url |
| 20:06 | <not-an-aardvark> | smaug: https://www.chromestatus.com/metrics/feature/timeline/popularity |
| 20:06 | <smaug> | that doesn't have all |
| 20:07 | <smaug> | I was thinking the tool to query what kind of code web pages have |
| 20:07 | <smaug> | (because https://www.chromestatus.com/metrics/feature/timeline/popularity is missing stuff) |
| 20:07 | <smaug> | foolip: you would know |
| 20:13 | <younder> | smaug, why do you care? |
| 20:14 | <smaug> | this time I'd like to know about the usage of <template> |
| 20:14 | <smaug> | (we just shipped a regression which is related to <template> and I'm just evaluating how fast it should be fixed. asap or very soon) |
| 20:14 | <younder> | cat you just use a grep .... | wc -l |
| 20:15 | <smaug> | grep <put the whole web here> ? |
| 20:15 | <younder> | I don't know the context |
| 20:16 | <not-an-aardvark> | curl '*.*.*.*' | grep 'template' | wc -l |
| 20:18 | <younder> | yes |
| 20:19 | <not-an-aardvark> | (that was a joke) |
| 20:19 | <younder> | sounds slow thow. Don't you want to fownload it first? |
| 20:19 | <younder> | download |
| 20:20 | <younder> | seriously grep uses a lpa algorithm whch is quite fast even piped |
| 20:20 | <not-an-aardvark> | I think smaug is looking for metrics from existing browsers and/or crawlers, since downloading the entire web and piping it into grep would not be practical. |
| 20:21 | <younder> | just dn't use the -A optin |
| 20:21 | <younder> | -E |
| 20:26 | <tobie> | annevk, MikeSmith: apparently related to dev.w3.org having been moved today. |
| 20:26 | <tobie> | annevk, MikeSmith: trying to get systeam to fix this asap |
| 20:27 | <jsbell> | smaug: There's bigquery and HTTPArchive |
| 20:28 | <jsbell> | Not perfect but you can grep the top 1e6 sites (main resource and script resources) |
| 20:29 | <smaug> | big query, that is what I was thinking, I think |
| 20:30 | <jsbell> | for example, 698 pages contain '<blink' in the 2016-11-01 desktop crawl |
| 20:32 | <younder> | Has it ever occured to you that your we browser is becoming overcomplicetaed. That due to it's compleity it is already on the verge of failing. |
| 20:33 | <jsbell> | TIL: BigQuery's regex match is case sensitive and doesn't have a flag, so you need to LOWER() explicitly. |
| 20:33 | <jsbell> | So make that 721 uses of '<blink' |
| 20:34 | <younder> | Particularly bink needs to shape up. Yoy can't include every thing and the kichecn sink and have no consequences. |
| 20:35 | <younder> | How many more hacks can you pull before it becomes copletely unmentainable |
| 20:35 | <younder> | ? |
| 20:35 | <annevk> | tobie: just redirect it to https://heycam.github.io/webidl/? |
| 20:36 | <tobie> | that's what it used to do |
| 20:36 | <tobie> | annevk: |
| 20:36 | <tobie> | ^ |
| 20:36 | <younder> | You have optimized the compiler, the linker the make syetem. But you still have way to many dependencies |
| 20:36 | <annevk> | tobie: I see |
| 20:37 | <younder> | The firefox guy's are fine.. |
| 20:37 | <annevk> | tobie: for some reason bz was still stuck on the old URL |
| 20:37 | <annevk> | tobie: maybe a bookmark |
| 20:37 | <tobie> | annevk: :) |
| 20:38 | <younder> | After spending a month with the blink source this is my conclusion. |
| 20:38 | <younder> | puuuwee |
| 20:42 | <younder> | I could be more elaborate, but quite frankly I don't think I have to. |
| 20:43 | <younder> | I have a 8 core zenon at 4GHZ and it takes 1.5 hrs to compile the source. |
| 20:44 | <younder> | Firefox 20 Min |
| 20:45 | <zcorpan> | jsbell: 2016_11_01_chrome_pages has 165,162 rows, not 1e6 |
| 20:46 | <jsbell> | zcorpan: Whoops - should have been 1e4 - I misread http://httparchive.org/about.php |
| 20:47 | <jsbell> | zcorpan: and I don't know how where the # of rows comes from. Thanks for pointing out my mistake |
| 20:48 | <foolip> | smaug: I'm not sure what you're thinking of, something like https://developers.google.com/webmasters/state-of-the-web/2005/elements? |
| 20:48 | <zcorpan> | maybe that page is outdated |
| 20:48 | <foolip> | smaug: I think there are internal thingies that I've never used, and of course httparchive |
| 20:48 | <younder> | piss your pant' if you must, but DO SOMTING about it! |
| 20:49 | <foolip> | But that's not in a form where you can just ask "how often is <template> used" of course |
| 20:50 | <foolip> | younder: I spent some time trying to remove unnecessary includes in Blink, seemed hard to make much progress. Probably serious refactoring would be needed to make a dent. |
| 20:57 | <zcorpan> | smaug: 822 pages for this query |
| 20:57 | <zcorpan> | SELECT page, COUNT(*) AS num |
| 20:57 | <zcorpan> | FROM [httparchive:har.2016_11_01_chrome_requests_bodies] |
| 20:57 | <zcorpan> | WHERE REGEXP_MATCH(LOWER(body), r'<template(\s|>)') |
| 20:57 | <zcorpan> | GROUP BY page |
| 20:57 | <zcorpan> | ORDER BY num DESC |
| 20:58 | <smaug> | 822 out for how many? |
| 20:59 | <zcorpan> | 165,162 |
| 21:00 | <smaug> | thanks |
| 21:01 | <zcorpan> | i can paste the urls somewhere if you want to check how broken they are |
| 21:14 | <younder> | foolip, It would never be enough. The package providers would have to go in for theselves and eliminate pckages. The best you can do is find a least common denominator of packages which fits all their needs |
| 21:15 | <younder> | and make them all stick with them |
| 21:16 | <younder> | Reove their headres. Make sure you nly do hat for the active developers. |
| 21:18 | <tobie> | annevk, MikeSmith: fixed |
| 21:20 | <younder> | run out of urage. Most would. You will have to do what I did at Opera. Take a Stand. |
| 21:20 | <younder> | courage |
| 21:33 | <Domenic> | Probably the first thing we should remove is MathML |
| 21:33 | <Domenic> | After that, maybe XML in general |
| 21:34 | <gsnedders> | I'd be amazed if we ever got away with killing XML |
| 21:34 | <gsnedders> | MathML… aren't things like MathJAX starting to build off of the DOM? or is almost everyone using the LaTeX-like API? |
| 21:38 | <TabAtkins> | Apparently MathML might make its way back into Chromium; Igalia is reworking it to be consistent with our Layout-NG project, which negates our main objection to it previously (that it was a big mess of layout hacks that made everything else more complicated). |
| 21:47 | <younder> | TabAtkins, Yes I agree a complete reimplementatin was neccesary |
| 21:49 | <younder> | But I think it can be done within a reasonable amut of time by one preson |
| 21:50 | <younder> | MathML is a lot of tags but it is just shifting UTF8 tags around with coords |
| 21:50 | <younder> | and scale |
| 21:51 | <younder> | The thing is to seperate stype from representation |
| 21:51 | <younder> | style |
| 21:52 | <younder> | So you have a CSS part and a HTMP part |
| 21:52 | <younder> | HYML |
| 21:53 | <younder> | I've looked at categories in LaTex and they are best represtented as a matrix with arrows. Well withhin the realm of SVG |
| 21:57 | <younder> | I think there is a simpe and elegant solution which wil perform efficientltly |
| 21:58 | <gsnedders> | younder: CSS doesn't suffice for MathML, AFAIK |
| 21:58 | <gsnedders> | younder: it certainly didn't used to, and I'm pretty sure it still doesn't |
| 21:58 | <younder> | I need a custom browser for my GUI in Formula (A functional language), but if it can help you as well, I would happily contribte. (in fact by the licence I am obliged to) |
| 22:00 | <TabAtkins> | Yeah, it def doesn't. |
| 22:00 | <younder> | MathML is stil just placing of symols. It follows 3 rules: translation scale and rotation |
| 22:00 | <younder> | SYMBOLS |
| 22:01 | <TabAtkins> | Uh, no, it's quite a bit more than that. |
| 22:01 | <younder> | control that triplet and all else follows |
| 22:02 | <younder> | coolor is an added attibute |
| 22:02 | <TabAtkins> | In particular, there are stretchy glyphs (which are *not* just scaled), and the sizing of things is complicated. |
| 22:02 | <TabAtkins> | (There's some places where an element's width depends on its height, which is the inverse of the CSS model.) |
| 22:03 | <younder> | So it is isolated and made to look simple |
| 22:04 | <younder> | I admit for example I have mostly looked at the implementation of LaTex |
| 22:05 | <younder> | It has it's metafonts of fixed sizes and the need to organize these glyphs |
| 22:06 | <younder> | Vector fonts complicate this somewhat, but it is simplified by fixing a size befor rendering |
| 22:07 | <younder> | A CSS file would do nicely |
| 22:08 | <younder> | TabAtkins, You sir are overcomplicating things |
| 22:08 | <TabAtkins> | lol ok whatever dude |
| 22:15 | <younder> | Let's just let the code decide. You don't have to take it into the main distribution. But I have a feeling you will. |