Replying to @bertails
Isn't that what Coffee was supposed to be?
February 2015
203 tweets
Replying to @Sniperovitch
En même temps personne n'est obligé d'écrire ça comme ça...
Replying to @Sniperovitch
Ben déjà une des queries sur #message est faite deux fois alors que la val() n'a pas dû changer.
Replying to @Sniperovitch
D'ailleurs je soupçonne que la plupart des queries devraient être faites une fois pour toutes en amont et pas dans le handler.
Replying to @Sniperovitch
C'est plein de bonnes choses dans ces technos, mais c'est comme Perl, faut apprendre à ne pas tenir le couteau par la lame.
Replying to @Sniperovitch
Il ne faut pas avoir peur. C'est un secret bien gardé mais JS en fait c'est le vrai Perl 6 ;-)
Replying to @scshepard and @stevefaulkner, @brucel, @nlhtml5
Some people clearly have too much time on their hands...
Operating as a large community requires built-in filters to maintain a workable signal-to-noise ratio. Does any standards org do it well?
Ability to endure abuse is not one. Neither is resilience in the face of process. Nor the will to kowtow to diktat.
You know, when we were getting excited about Open Source all those years ago, I not once saw this coming. “@OSBTeam ”
I hate telcons with fiery passion but I have to admit a pang of sadness for @w3c’s trusty Zakim being decommissioned. w3.org/2015/07/zakim.…
♥ 7↺ 7
Replying to @briankardell
MIT can’t support the (dying) infrastructure.
Replying to @briankardell
WebEx
Replying to @fgeorges
Well, not for IRC integration for instance.
Replying to @michaelhkay
Yeah, I'm trying to get a couple of it. It's actually Ralph speaking! I think it was part of the deal getting it. @fgeorges
Replying to @masinter
I think that's what it's headed towards, but it also has to actually work 140 days from now. @briankardell
Replying to @null and @marcosc
With XHTML2 you could've just slapped an src on a div and called it a day.
Bit Goo.
Replying to @null and @sgalineau
You can tell in the second paragraph that it is not quite at the quality level of the Gorafi.
Replying to @fgeorges
In terms of writing, silly :-)
Replying to @fgeorges
Malheureusement non!
Replying to @yoavweiss and @patrickhamann, @paul_irish
That might be worth prodding @plhw3org about.
♥ 1
Replying to @ElissaBeth
National wha? National of which nation?
Replying to @2vanssay
Les mêmes batteries servent aussi pour les StormTroopers. C’est le tournevis pour accéder à l’arrière du casque. @bortzmeyer
Promising? «Antibiotics that target mitochondria effectively eradicate cancer stem cells across multiple tumor types» impactjournals.com/oncotarget/ind…
Replying to @fabricedesre
Interesting, thanks!
@iandevlin Looking at github.com/iandevlin/aria… it really hurts me to think I need to duplicate so much information for ARIA to work.
Would this not be better supported by a script that would add the ARIA for UAs that don’t understand the new elements?
Replying to @null and @iandevlin
Sure, but maybe the teaching should focus on what can’t be automated? Then toolify what can be?
Replying to @null and @iandevlin
It’s just my gut reaction: the basic template seems to teach me that ARIA is going to make me a very sad developer!
Replying to @null and @iandevlin
(But please don’t take this as all negative, I’m only suggesting feedback because I think this is a great idea!)
Replying to @null and @iandevlin
I know, I know, but I was thinking about the cases where it is used for older UAs. A script could add role=article on <article>.
Replying to @null and @iandevlin
Yes, but in my experience people misuse most when they have to do dumb, repetitive tasks.
Replying to @null and @iandevlin
So suggesting you should always use <main role=main>, <article role=article> by hand might lead to more misuse.
Replying to @null and @iandevlin
The idea of a script is «We handle the dumb stuff, here are the important cases you need to noodle about».
Replying to @null and @iandevlin
I think that would rock :)
Replying to @LeonieWatson
Yeah, I was more thinking about a build step or a script that just always runs. @iandevlin
Is there some wonderful alternative to PhantomJS for intensive HTML processing I have somehow missed? No, jsdom is too slow here.
Replying to @tobie
Apply lots of processing to a large document. It used to work. Today it started crashing for no reason :)
Replying to @tobie
I think I’m going to have to just resort to jsdom and be patient.
Replying to @tobie
Yeah, things get pretty slow running queries on a really large document.
Replying to @dmvjs
The former can’t work headless nor have access to the DOM, the latter doesn’t help. Sadly.
Replying to @briankardell
Yeah, pretty much indeed. @LeonieWatson @iandevlin
Replying to @null and @fhemberger
Ah, I had completely forgotten about that project. Thanks, I’ll look to see if it helps! @asciidisco
Replying to @tobie
Not the end of the world, but not trivial. It also depends on jsdom switching from using Mutations to track changes to sth internal.
Replying to @tobie
Presumably because mutations are a poor fit, and might interact with some nasties.
♥ 1
Replying to @tobie
Yeah but it looks like they’re already working on that part, for other reasons.
Replying to @ljouanneau
Replying to @mikewest
Which checker, new generation or old one? If the new one, to me. If the old one, use the new one :)
Replying to @stevefaulkner
Whoa, WTF is that? @brucel
Replying to @mikewest
Replying to @mikewest
give w3.org/2014/10/pubrul… a spin. At least if there’s a bug it’ll get fixed.
Replying to @mikewest
Very precisely this github.com/w3c/specberus/…
Replying to @mikewest
What’s your document and what error are you getting?
Replying to @mikewest
Looking at the code, it seems that the checker may not have been updated for the fact that «All Rights Reserved» is now optional.
Replying to @mikewest
So, «Short names differ between This and Latest Versions.» is a real issue.
Replying to @mikewest
And I can see where the copyright issue is coming from.
Pretty SVG things: jfire.io/animations/ (via @recifs)
♥ 2↺ 1
I’m happy to join the the gang at svg-news.com (see also @SVG_news) to help bring forward the cool in SVG. It’s been a while!
♥ 4↺ 5
Replying to @perllover
Sadly, no :(
Replying to @perllover
Poor thing :(
Hey folks, do you ever think in terms of HTML’s «Kinds of content» html.spec.whatwg.org/#kinds-of-cont… or is that just for spec-heads & validator devs?
♥ 2
Replying to @starletdreaming and @_tomchristie
I think «gumpf» was the technical term I was looking for.
Replying to @null and @obiwankimberly
I’ve been wondering to what extent we can drop it and live.
Replying to @null and @obiwankimberly
You can still write a validator with a DTD (and there is one), for that purpose having named groups of things is helpful.
Replying to @null and @obiwankimberly
Validation is orthogonal to DTDs; if it doesn’t matter now it didn’t matter then.
Replying to @null and @obiwankimberly
But it can be a useful tool to assess quality, catch bugs, not rely on fallbacks, etc.
Replying to @null and @obiwankimberly
No, the spec makes no sense to developers — that’s one reason I’m thinking about this. There’s too much gumpf.
Replying to @null and @obiwankimberly
Oh Web devs. Who cares about browser devs? :)
Replying to @null and @obiwankimberly
Up to a point, yes, but who has the time? Jumping into it one is quickly lost.
Numérique ta mère, je dis digital si je veux: blogdumoderateur.com/numerique-ou-d… (merci @ElieSl ).
♥ 1
Oh, look, CC-BY v4 is now deemed GPL compatible. gnu.org/licenses/licen…
♥ 5↺ 6
Replying to @alastc
Remembering what’s block/inline, sure, but does anyone actually do that by memorising the list of palpable, metadata, etc. elements?
Replying to @teleclimber
I'm not sure what you mean, we can have any number of elements with no classification of this sort.
Replying to @teleclimber
Handy for what? :-)
Replying to @TzviyaSiegman
"Hardcore triplers". That's the big hit on YouPorn, right? I, uh, heard about it from a friend.
♥ 1
Replying to @hsivonen
I don't know, I haven't seen an announcement, just stumbled over it. I know it had been discussed for a while.
Replying to @teleclimber
Only if it serves a purpose. I know no one who's memorised the categories and ever uses them when coding.
Replying to @null and @sideshowbarker, @hsivonen
Those discussions broach a very interesting question though, which I hadn’t considered: 1/2
Replying to @null and @sideshowbarker, @hsivonen
2/2 Can a CC-BY/CC0 work be GPL compatible if its source (and, presumably, build tools) aren’t openly available?
WebSpecs now support specs in ReSpec (and vice-versa)! specs.webplatform.org/docs/ cc @stevefaulkner @LeonieWatson @briankardell
♥ 7↺ 6
Replying to @LeonieWatson and @stevefaulkner, @briankardell
Make your source file be index.src.html, add your project to webspecs/the-index, add the hook.
♥ 1
Replying to @LeonieWatson and @stevefaulkner, @briankardell
Happy to walk you through it tomorrow.
♥ 1
Replying to @briankardell
Or you do it tonight :-) I think you can even accept the PR on the-index. @LeonieWatson @stevefaulkner
♥ 1
Is there a tool that works well to convert Sony MXF videos to something useful, preferably on a tous Mac?
Replying to @null and @exsecrari
It’s not listed as supported by Handbrake but that could be a doc bug. I’ll give both a shot, thanks!
Harvard and M.I.T. Are Sued Over Lack of Closed Captions nytimes.com/2015/02/13/edu…
↺ 1
Old, but magical :) «A regular expression to check for prime numbers» noulakaz.net/weblog/2007/03…
♥ 4↺ 2
Replying to @briankardell
In a call now, checking right afterwards. Exciting! @LeonieWatson @stevefaulkner
♥ 1
Replying to @briankardell
LMK if this matches your wants, and if it was easy enough! specs.webplatform.org/#panels-and-pa… @LeonieWatson @stevefaulkner
Adding panel and panel set elements to HTML specs.webplatform.org/common-panel/b…
♥ 5↺ 4
Replying to @briankardell
Sometimes I wonder what I wrote all those docs for :) Look for hook there specs.webplatform.org/docs/how-to/cr… @LeonieWatson @stevefaulkner
Replying to @briankardell
I got an error in the logs, investigating. @LeonieWatson @stevefaulkner
Replying to @briankardell
But fixed — my caching of information really sucks. @LeonieWatson @stevefaulkner
♥ 1
Replying to @anthony_ricaud and @rik24d
File a bug :)
Replying to @jdee
The name will have a dash in it until it's not a prototype.
Replying to @psd and @mahemoff
We could have had them years ago, but the usual myopia prevailed :-)
I work on the Internets; can I write off my cat as a professional expense?
♥ 14↺ 15
♥ 1↺ 1
Replying to @slightlylate
This is not my baby, take it up with them :-)
Replying to @mattur
Report bugs to the editors :-)
Replying to @wycats
It is a rule of the Internets that once people start complaining about a name and trying to replace it, it is guaranteed to stick.
↺ 1
Replying to @wycats
Me neither, but it's there all the same. Portable means something different, too.
Who invented the term "isomorphic javascript"? It's OK, come in now and we won't hurt your cats. @wycats and I just have some questions.
♥ 7↺ 8
Replying to @wycats
Not really. I can run, say, client js in Phantom on my server, it's portable. But that's not isomorphic yet.
♥ 1
Replying to @wycats
I understand what you're getting at, but I still think there's a difference. For instance, it could be isomorphic but Chrome-only.
Replying to @lbljeffmo
Just for the pleasure of being terribly pedantic, wouldn't running the same be "isodromic"? @wycats
♥ 1
Replying to @spikebrehm
If you squint at it right the same goes for portable code. But I agree there's a specificity. @wycats
♥ 1
Replying to @spikebrehm
To be clear I love the idea (in fact hinted at in berjon.com/there-is-still…) but it's true the name gets stares. @wycats @indexzero
Replying to @spikebrehm
But *shrug*, I can't say I truly mind. Having your name bashed is sort of a rite of passage for Web tech :-) @wycats @indexzero
♥ 1
Replying to @brianleroux
We sure could be a nicer bunch (see specs.webplatform.org/docs/policy/an…) but we were kidding about the cats. @spikebrehm @wycats @indexzero
Replying to @brianleroux
♥ 1
Replying to @kaythaney
Scholarly HTML? @ptsefton @petermurrayrust
Replying to @null and @ptsefton
I sort of have a pretty strong interest in that, incidentally :) Can join the conversation somewhere? @petermurrayrust @kaythaney
Replying to @mikewest
I’m pretty sure it would be a net productivity win if we all took 6 months off to go hunt down those problems. @rodneyrehm
♥ 1
Replying to @mikewest
LOL. What’s the problem this time? @rodneyrehm
Replying to @rodneyrehm
A tool that validates W3C specifications for adherence to all sorts of crazy checks. @mikewest
Replying to @mikewest
You only have bugs because you’re not using ReSpec :) (Okay, there are *some* real bugs too…) @rodneyrehm
Replying to @mikewest
Ah, the famed Spec for Specs. People whisper about it… @rodneyrehm
Replying to @mikewest
Funny how when you say «jumped on this» I picture a grenade. You want to cat-herd the cat herders? @rodneyrehm
Replying to @mikewest
I surely don’t disagree; I just think that people have refrained from going meta for fear of a rathole.
Replying to @mikewest
But maybe that’s unfounded fear! Maybe if the ball started running we’d get something nice quickly.
Replying to @mikewest
Mind you, I think that the introduction of the automatic publishing system might force that to happen.
Replying to @mikewest
That is actually planned for. There’s a lot of documentation in the old pubrules, it needs extraction + checking.
Replying to @ourmaninjapan
Are you experimenting with something in particular?
Replying to @LeonieWatson
Happy to expose interfaces for control, too. @ourmaninjapan
Replying to @ourmaninjapan
I think it depends on what you want. A comedic post I would think is best read by the author otherwise machines are better :)
Replying to @null and @ericnentrup
You can use w3.org/community/ to get a group started, you just need five other interested people. @w3c
Replying to @yoavweiss
You want @BrobinBerjon for the extra sulky with a hint of rugged roguishness.
«Loi Macron: pourquoi l'UMP frise l'inconscience» ou le retour de la droite la plus conne du monde latribune.fr/actualites/eco… via @SignOffParis
Quelqu’un bosse sur les horaires de séances de ciné en open data? Parce que @allocine, il y a vraiment de quoi se pendre.
♥ 2↺ 6
Replying to @null and @nhoizey
Je teste, merci :-)
DOM 3 XPath! Y U SO HORRIBLE?
♥ 1
Replying to @null and @tolfsen
I haven’t looked at what’s planned for Complete Selectors; but regular Selectors are missing a lot.
Replying to @SimonSapin
What I was doing right now was «find all elements with an on* attribute» aka //*[@*[starts-with(local-name(), 'on')]] @tolfsen
Replying to @SimonSapin
It’s not that complicated a pattern :) To be serious about replacing XPath, it’s the sort of thing that needs to work. @tolfsen
Replying to @SimonSapin
It’s sort of what @tolfsen was hinting at, it’s also what @tabatkins has mentioned before.
Replying to @SimonSapin
Also, *if* it’s not a goal and XPath remains on the platform, we need to make it actually usable. @tolfsen
Replying to @null and @tolfsen
XPath 1.0 is really not that big, it’s also well thought-out. I don’t think the gap is that huge. @glazou @SimonSapin @cmeyth
Replying to @glazou
But those are in the DOM, and functions cast for consistency. XPath 1.0 implementations aren’t hard. @tolfsen @SimonSapin @cmeyth
Replying to @glazou
Sure, in the CSS context it makes no sense. But for selection I don’t see the issue. @tolfsen @SimonSapin @cmeyth
Replying to @glazou
Sure, it does change the model. And it would be very complex to integrate into the CSS context. @tolfsen @SimonSapin @cmeyth
Replying to @anthony_ricaud and @rik24d
If that exists I’ve missed it! Maybe you’re thinking of [attr^=foo]? @SimonSapin @tolfsen
Replying to @gsnedders
Terribad barely begins to cover it. @tolfsen
Replying to @tabatkins
Yeah, at the very least aria-* and data-foo-* come to mind too. @SimonSapin @tolfsen
♥ 1
Replying to @tabatkins
And who knows, it *might* even be possible to find a better syntax than [@*[starts-with(local-name(), ‘x’)]] @SimonSapin @tolfsen
Replying to @brucel
I’d forgotten! There are some details here vepca.wordpress.com/category/uncat…
Replying to @tabatkins
To be honest, that selector would have confused me; I probably would have expected the bug’s interpretation. @gsnedders @tolfsen
Replying to @w3cmemes
Ah, HTML?
90% des individus qui mangent une tartiflette le font par Internet. #CazeneuveStyle @BCazeneuve
↺ 1
Is there a good, relatively introductory, book on percolation theory?
Why one should read with a pen at hand. nybooks.com/blogs/nyrblog/…
Mastering git submodules — essential reading from @porteneuve medium.com/@porteneuve/ma…
♥ 4↺ 2
Replying to @stevefaulkner
I thought I was :)
Replying to @mathgon
Merci, mais je ne cherche pas à tout aspirer :) @nhoizey
@poulpita Ça cause de toi et de Web Crypto sur SeenThis :) seenthis.net/messages/34373…
OH: "Si même Bob l'Éponge est plus musclé que moi..." @_noemes
@SignOffParis Vous avez quelle source pour l'histoire de @Bruno_LeMaire se faisant refuser une commission sur le Qatar? Je ne trouve pas.
Not new, but an interesting (if debatable) view from David Deutsch on AI: aeon.co/magazine/techn…
"How Network Science Is Changing Our Understanding of Law" technologyreview.com/view/534576/ho…
Replying to @janl
Actually thought of something like that a while back: Web site reviews à la Gordon Ramsay.
"The Origins of Order", S. Kauffman.
Dense, much like a 650pp scholarly article, but astonishing, clear, and well worth the effort.
♥ 1
Why business leaders should think like scientists medium.com/davos-2015/why…
♥ 1
Replying to @null and @_paulverbeek
Short of relying on an HTML diff, no. But if you can describe specific needs it could be done.
Replying to @null and @_paulverbeek
One thing I've thought of is to pare the spec down to sections that have changed, then only show diffs for those.
Replying to @null and @ryosukeniwa
Without the new features there might not be that many users left to fix bugs for.
♥ 2↺ 1
Replying to @null and @ryosukeniwa
I know what you meant :-) The vast majority of end-user visible features are built by Web developers, who need features.
♥ 1
Replying to @null and @_paulverbeek
If you don't mind filing a bug against github.com/w3c/spork I'll try to build an easy way.
@ryosukeniwa If you have to target lagging browsers like old IEs or mobile Safari then that dominates, yes. But that's an upgrade problem.
@ryosukeniwa There's a *lot* missing still in terms of features.
Replying to @null and @ryosukeniwa
I'm not saying it doesn't happen, it certainly does. I just don't think it matters to the exclusion of new features.
Replying to @null and @ryosukeniwa
I'm curious which bottleneck you're hitting though. They can be surprising :-) Have you profiled?
@theWinnower Any chance we can get a mobile style sheet at some point? It would be most lovely!
♥ 1
Replying to @theWinnower
Is your setup/plan amenable to having (some of) your stuff on GitHub?
Replying to @theWinnower
I was thinking code. Like if instead of ranting to you on Twitter I’d make a pull request.
♥ 1
"Design for a Brain", W. Ross Ashby. Disappointing but with nuggets of brilliant precision.
Replying to @torgo
Are they good? I wanted to try but I'm having SOAF issues with this one.
Why is it that ::first-letter only applies to blocks?
↺ 1
Replying to @null and @sgalineau
Heh. Just thinking of styling the first letter of an element.
Replying to @null and @sgalineau
I was looking to underline the first letter of every strong element, in this case.