| 00:07 | <conley> | In html5, If I have a <nav></nav> and I want div-like elments inside...what should those be? <nav> or <section>? |
| 00:08 | <Hixie> | <div>? |
| 00:08 | <Hixie> | depends what you mean by div-like :-) |
| 00:09 | <conley> | hixie: I consider all those div-like...i wasn't sure if div was still valid in html5... |
| 00:09 | <conley> | I did a ton of googling and ironically, right after I posted here I finally found http://html5doctor.com/the-section-element/ |
| 00:13 | <Hixie> | that's a pretty good overview, yeah |
| 00:28 | <conley> | so is it ok to put a <nav> inside a header? |
| 00:29 | <conley> | like, if I have a big banner at the top of my site and some links inside, is that fine to label them that way? |
| 00:30 | <AryehGregor> | Sure. |
| 00:30 | <AryehGregor> | <header> can contain any flow content. |
| 00:31 | <conley> | thanks |
| 02:15 | <drclue> | Long day today , just refueled the jet packs on the truck :) |
| 02:29 | <drclue> | I really would like a method to from within HTML to create some sort of XML data island and use entities like &name; sequences for in page templating. |
| 02:29 | <drclue> | Figure if one can annoy the world with a marquee element , there might be room for something like this too |
| 02:33 | <drclue> | Any particular reason we need embed , applet and object as opposed to merging them into object? |
| 02:34 | <drclue> | For that matter , iframe could be rolled into object too |
| 02:39 | <Hixie> | applet is gone |
| 02:39 | <Hixie> | embed, iframe, and img are basically specialised versions of object |
| 02:40 | <Hixie> | (object has all kinds of problems -- it turns out overloading elements to do lots of similar things isn't a good design after all) |
| 02:40 | <Hixie> | (<input> is similarly a problem) |
| 02:40 | <drclue> | I understand they are specialized , but why keep them around? |
| 02:40 | <Hixie> | they work better |
| 02:40 | <drclue> | emmm how so? |
| 02:41 | <Hixie> | have you tried using <object>? browsers implement it in all kinds of buggy ways |
| 02:41 | <Hixie> | just compare the spec for <object> to the spec for iframe, img, and embed -- the object one is way more complicated for what it does |
| 02:43 | <drclue> | I understand that in use in the current ways things are done that some things work better than others , but in rolling things together |
| 02:43 | <drclue> | it would just be a matter of vectoring existing code so in their rolls the code would still be the same |
| 02:43 | <Hixie> | you'd think |
| 02:43 | <Hixie> | yet it seems not |
| 02:44 | <Hixie> | in practice <object> has had way more bugs than the other three put together |
| 02:44 | <drclue> | I'm just kicking thoughts around. I've been enjoying the show since mosiac |
| 02:45 | <drclue> | How about my other thought of XML data islands and linkages to entities for in page templating. I've been using that approach server side since old netscape |
| 02:46 | <drclue> | "yet it seems not" That one I just find hard to believe , but I'll take your word for it |
| 02:47 | <Hixie> | just test the browsers -- <object> is more buggy |
| 02:47 | <Hixie> | or look in the bug databases for mozilla and webkit |
| 02:47 | <Hixie> | <object> has way more bugs than the others |
| 02:47 | <drclue> | I know what your saying is true, but that really was not my point. |
| 02:49 | <drclue> | I figure you have been messing about with browsers probably as long as I have |
| 02:49 | <Hixie> | since 1998 :-) |
| 02:49 | <drclue> | I started about Mosiac |
| 02:49 | <Hixie> | well, since 1995, but i started doing standards stuff in 98 |
| 02:50 | <drclue> | Started programming before PC's came with screens and keyboards as options |
| 02:51 | <drclue> | So basically neither of us is new to this. |
| 02:51 | <Hixie> | indeed :-) |
| 02:52 | <drclue> | I tend to believe that while object is sorta buggy that there might be a way to spec our way out of the buggy into those places that work better |
| 02:53 | <drclue> | Thats the fun thing about specs is that you can make demands and if those demands leverage the existing code base , they can be reasonable |
| 02:55 | <Hixie> | well we tried that with html4, for embed/img/applet/iframe -> object |
| 02:55 | <Hixie> | and it didn't work :-) |
| 02:56 | <drclue> | Thats because the spec did the migration incorrectly |
| 02:56 | <Hixie> | does html5 do it correctly? |
| 02:57 | <drclue> | You tell me , will object as a result of the spec behave? |
| 02:58 | <Hixie> | i think the bugs is more of a result of the element being overloaded with multiple roles than the spec being poor, in this particular case, so i disagree with the premise of your question :-) |
| 02:59 | <drclue> | Maybe the question is better asked this way. Of all the object flavored elements , which do you think works most correctly cross-browser? |
| 03:01 | <Hixie> | img, them iframe, then embed, then applet, then object, i think |
| 03:01 | <Hixie> | but it's hard to say |
| 03:01 | <drclue> | img does not really do anything comparable, which next in order brings us to iframe |
| 03:02 | <drclue> | iframe is a possible key to this , even if it were called object |
| 03:03 | <drclue> | Leveraging existing good behaviors is something I would encourage |
| 03:04 | <drclue> | Not much to changing an element name in a table |
| 03:06 | <MikeSmithX> | Hixie: I agree completely with what you said about overloading elements to do lots of similar things being bad design -- especially the case of the input element -- but I think command is also an instance of that same kind of pattern |
| 03:06 | <drclue> | As you said object has so many issues , so just call iframe object and back fill it with the object element. |
| 03:08 | <drclue> | [MikeSmithX] I really take issue with having a bunch of different elements that do basically the same thing drifting about in disparate implementations |
| 03:08 | <drclue> | I would rather leverage what browsers do well |
| 03:08 | <Hixie> | MikeSmithX: how so? |
| 03:08 | <Hixie> | MikeSmithX: oh the whole radio/checkbox/option thing? |
| 03:09 | <MikeSmithX> | drclue: we don't always get that choice.. in some cases where are stuck also with retaining what browsers do not so well, and what existing content uses (how every inappropriately it might be using it) |
| 03:09 | <MikeSmithX> | Hixie: yeah |
| 03:10 | <MikeSmithX> | basically, I think every place where the spec says, "If the foo element is in the bar state.." is a red flag |
| 03:10 | <Hixie> | the spec never says that :-P |
| 03:10 | <Hixie> | elements aren't in states :-P |
| 03:10 | <Hixie> | it's attributes that are in states :-P |
| 03:10 | <MikeSmithX> | heho |
| 03:10 | <MikeSmithX> | yeah, you know what I mean |
| 03:11 | <Hixie> | MikeSmithX: i suppose we could have <command>, <command-checkable>, and <command-radio>, or something, but really they're all just menu items (maybe <menuitem> would be a better name) |
| 03:11 | <MikeSmithX> | drclue: there is a rational argument to be made for retaining applet event |
| 03:11 | <MikeSmithX> | *even |
| 03:11 | <MikeSmithX> | the applet element |
| 03:11 | <Hixie> | MikeSmithX: i looked at existing designs when doing <command> and most just have a single type of element for that (e.g. xul just has <menuitem>, iirc) |
| 03:12 | <drclue> | [MSX] I know about legacy, we all know about legacy. In pulling this stuff together I figure one leverages what they do well , so that those having to code it are less resistant |
| 03:12 | <drclue> | I'm a coder of many decades, and and internet type of many many years too, and I write parsers, so I feel that while it is but an opinion , I can contribute one from the vantage point of having worn all the shoes |
| 03:13 | <MikeSmithX> | Hixie: then I would consider a flaw in those specs as well, and an anti-pattern that shouldn't be further perpetuated |
| 03:13 | <drclue> | BRB smoke break |
| 03:17 | <Hixie> | MikeSmithX: i don't know that i agree that in this case it's really overloading... especially if you consider the way that commands ended up being specced out, it's not clear to me that they are that different from each other |
| 03:17 | <Hixie> | MikeSmithX: they are far more similar to each other than, say, the various input types |
| 03:18 | <drclue> | I'm back |
| 03:18 | <fantasai> | jgraham: sorry, I meant bug 137 |
| 03:18 | <fantasai> | jgraham: not 135 |
| 03:19 | <drclue> | My thought really is that the specification related to the multiple object types should try and flow from the object that currently works best cross browser |
| 03:20 | <drclue> | The other objects should then back fill into that one object |
| 03:21 | <MikeSmithX> | Hixie: yeah, agreed in command case they are more similar and certainly less confusing than input |
| 03:21 | <drclue> | Coding wise that would be easier and hopefully provide better acceptance |
| 03:22 | <MikeSmithX> | as far as input, it's not really accurate to say the language has *an* input element.. instead it has 15 or more different input elements |
| 03:22 | <drclue> | It would also give coders an excuse to fix what aint broken (tongue in cheek) |
| 03:22 | <MikeSmithX> | fantasai: hey, so it seems like there has recently been some forward movement on the CSS ruby spec |
| 03:24 | <fantasai> | MikeSmithX: yeah, I hear ishida wants to edit |
| 03:24 | <fantasai> | :) |
| 03:24 | <MikeSmithX> | yeah, I was chatting with him about it earlier |
| 03:25 | <MikeSmithX> | Hixie: btw, can you recall what is the use-case for having multiple ruby bases and texts in a single ruby element? |
| 03:25 | <MikeSmithX> | as the HTML5 spec currently allows |
| 03:25 | <Hixie> | "it works in IE", i think |
| 03:25 | <MikeSmithX> | ah |
| 03:25 | <fantasai> | MikeSmithX: well, the CSSWG approved his editorship if that helps :) |
| 03:25 | <fantasai> | MikeSmithX: I don't think the editor's draft is on dev.w3.org yet |
| 03:25 | <Hixie> | also advanced ruby or whatever it was called allowed something equivalent |
| 03:26 | <MikeSmithX> | fantasai: OK. I hope he might consider a co-editor, but I think we will be able to get him help with it regardless |
| 03:27 | <MikeSmithX> | anyway, I am just glad that it's actually being worked on actively |
| 03:27 | <MikeSmithX> | Hixie: I asked because it's my understanding that the original ruby spec did not allow it |
| 03:27 | <MikeSmithX> | http://www.w3.org/TR/ruby/ I mean |
| 03:28 | <drclue> | So let's say that iframe is the most reliable object for instance. We claim that all objects to be named object take as their basis iframe accept where otherwise noted and expand out from there adding back in the details of object , sorta in spec suggesting one really just goto the element table and write "object" where it says "iframe" and inject the additional details of object to the spec of iframe |
| 03:29 | <MikeSmithX> | but if IE does allow multiple ruby bases and especially if there is existing content that relies on it, well, it would seem hard to justify not allowing it simply for reasons of keeping it aligned with that ca. 2001 spec |
| 03:30 | <MikeSmithX> | that original ruby spec is one that I'm not clear on whether it went to Rec without there actually having been any implementations at the time |
| 03:30 | <Hixie> | this isn't my area of expertise |
| 03:30 | <Hixie> | anyway, gotta go |
| 03:30 | <Hixie> | bbiab |
| 03:31 | <drclue> | ruby is so far afield of HTML , I have a hard time understanding it's presence in the conversation |
| 03:33 | <fantasai> | MikeSmithX: just trace back the versions and see if there was a CR stage :) |
| 03:34 | <MikeSmithX> | drclue: http://dev.w3.org/html5/spec/text-level-semantics.html#the-ruby-element and http://www.w3.org/International/questions/qa-ruby |
| 03:35 | <drclue> | Yes , I know it has made inroads , but that is another matter |
| 03:35 | <MikeSmithX> | fantasai: well, even HTML4 had a CR stage... I just think it lasted for a week or two or something |
| 03:36 | <MikeSmithX> | or maybe it didn't even have a CR |
| 03:36 | <fantasai> | MikeSmithX: I thought the old process didn't have a CR stage |
| 03:36 | <MikeSmithX> | I believe HTML4 went from FPWD to Rec in 5 months |
| 03:36 | <fantasai> | MikeSmithX: LC -> PR |
| 03:36 | <fantasai> | ? |
| 03:36 | <MikeSmithX> | fantasai: ah, OK |
| 03:37 | <drclue> | function,data,presentation one should never mix these drugs |
| 03:44 | <drclue> | Even in reading the cited spec, it still seems warped. Perhaps this would be better served with a CSS pseudo like :before :after and be :above |
| 03:47 | <drclue> | Hard to say , but I don't care for the Ruby thing, but I would not throw a fit about it I guess. Once I saw marquee , I had to just say "Elvis has left the building" |
| 03:47 | <drclue> | and keep in mind that this is a collaborative effort that requires a lot of patience or else nothing will actually get done. |
| 03:55 | <drclue> | Another issue that puzzles me is how multi-col layouts are ending up in CSS and not in layout. I really would like to see the multicol element return |
| 03:56 | <drclue> | I do like the return of the table layout elements , and while I got used to writing CSS for table layouts never could understand how tables were considered style |
| 04:00 | <drclue> | I've seen a lot of weird stuff since 1994 when I switched from general programing to Internet based development and I'm sure I'll see a lot more . |
| 04:01 | <drclue> | whatever spec gets laid down , I'll just code for it and back patch as best I can like always, but some days I just have to scratch my head |
| 04:05 | <drclue> | Can anyone explain to me why an HTML specification would have "5.4 Microdata vocabularies"? |
| 04:06 | <drclue> | Why would layout care about the microdata vocabularies of the moment? |
| 04:07 | <drclue> | It would almost seem like caring about "dBase III" |
| 04:08 | <othermaciej> | 1) HTML is not "layout" |
| 04:08 | <drclue> | HTML should be what then? |
| 04:08 | <othermaciej> | 2) The W3C HTML5 spec does not include them, but they are in the all-inclusive WHATWG copy (which includes many other things like WebSocket) |
| 04:08 | <othermaciej> | 3) dBase III? huh? |
| 04:08 | <drclue> | My point exactly |
| 04:09 | <othermaciej> | HTML is a markup language for hypertext documents and applications on the Web |
| 04:09 | <othermaciej> | it is not a layout language |
| 04:09 | <othermaciej> | CSS is a layout/styling language |
| 04:09 | <drclue> | Of course it is |
| 04:09 | <drclue> | CSS is styling |
| 04:10 | <drclue> | making reference to particular fixed widths and such anywhere is a bad idea in general |
| 04:11 | <othermaciej> | p { display: block; margin: 1.0em 0px; } |
| 04:12 | <othermaciej> | that is the line of CSS that gives the <p> element its default layout behavior in WebKit |
| 04:12 | <othermaciej> | it comes from the file html.css |
| 04:12 | <drclue> | The style of that P element is that it is a block with a margin, and is indeed style |
| 04:12 | <boblet> | foolip: I’ve got some more Live Microdata feedback… |
| 04:12 | <othermaciej> | drclue: that style is what controls its layout |
| 04:13 | <othermaciej> | you put that same style on a <div> or <span>, and it lays out the same |
| 04:13 | <othermaciej> | the way <p> is different is semantics, not layout |
| 04:13 | <othermaciej> | <p> means the contents are a paragraph, but <div> or <span> would not indicate anything specific about the contents |
| 04:14 | <drclue> | I could with layout indicate that the order of appearance and placement in the tree of the P element involved is one thing while the style of tha rendering is another |
| 04:14 | <boblet> | foolip: when I use the card/ical download links the file has a .part extension, so isn’t recognised by relevant apps |
| 04:14 | <othermaciej> | the appearance and placement of the P element on screen is completely controlled by CSS |
| 04:14 | <boblet> | foolip: also the download link doesn’t work in Chrome 5 |
| 04:14 | <drclue> | There is a devision of labor |
| 04:14 | <drclue> | There is some overlap |
| 04:15 | <drclue> | but the primary roles should be separate |
| 04:15 | <othermaciej> | the primary role of HTML markup is semantics |
| 04:15 | <othermaciej> | the primary role of CSS is controlling styling and layout |
| 04:16 | <drclue> | CSS = Cascading "style" sheets. |
| 04:16 | <othermaciej> | in case this means anything to you, my day job is as a WebKit developer, and I am pretty sure I know what our code does |
| 04:18 | <boblet> | foolip: if I change the extension the vcard files downloaded by FF load fine |
| 04:18 | <boblet> | foolip: however the .ics file appears to load but doesn’t show up (iCal being screwy most probably) |
| 04:18 | <drclue> | Well if we are flipping our peckers on the table to lay a ruler next to , I have been programming since before PC's had keyboards and screens as options and working the Internet since Mosiac. I don't think penis wars makes one right or wrong and should not be something by which to decry the merits of ones position |
| 04:21 | <boblet> | drclue: re: you example about style overriding layout, if CSS is disabled then the user agent will indicate what the element is to the user based on the element’s semantics (eg screen reader will pronounce words wrapped in <strong> differently) |
| 04:27 | <drclue> | Having written screen reader code for the blind and participated in the sun conferences , I can say that font-weight and other CSS styling s should indeed effect the |
| 04:27 | <drclue> | style in which content is rendered verbally. |
| 04:30 | <drclue> | The failures of software aimed at the visual market should in no way be used as an excuse , just as penis wars should not be used , but rather the devision of labor in the standards should win out in this regard |
| 04:31 | <TabAtkins> | I'm not sure why the current division of labor is inadequate. It's split differently from how you expect, apparently, but HTML is semantics and structure, CSS is style and layout, and Javascript is behavior. |
| 04:34 | <drclue> | Standards wise , while there may be some overlap javascript = function , XML = data , HTML = layout , CSS=style |
| 04:34 | <drclue> | There are indeed overlaps and accepted abuses and pseudo standards to augment actual standards , but they is what they is |
| 04:34 | <TabAtkins> | That's not the division that currently exists, or is reflected in current standards. |
| 04:34 | <TabAtkins> | It may be how things were intended to be divided in the past, I dunno. |
| 04:35 | <drclue> | I guess we are each free to view the standards from our own vantage point |
| 04:36 | <TabAtkins> | I see you arguing previously that HTML's role is 'layout'. Why do you say this? At what point was HTML ever any sort of layout language, except by default (because CSS wasn't mature enough to do so yet)? |
| 04:41 | <drclue> | HTML is a structural beast as relates to it's nature , CSS is the paint or style you apply to that structure. At times the lines can blur a bit as one reaches into the realm of the other, but they really do have their basic natures. At times , even in spec they have over-reached, like for example tables. When the spec changed , I complied but it was wrong , which is why it has never really worked all that well. TABLES belong to HTML, just |
| 04:45 | <TabAtkins> | Indeed, tables do belong to HTML, because "tabular data" is a unique semantic that doesn't exist elsewhere. It's very useful to be able to indicate that a set of data is correlated along two different axes. |
| 04:46 | <TabAtkins> | On the other hand, table *layout* is part of a CSS module, and capable of being used on any element to provide the layout you want. Grid-like constraint-based layout is useful outside of presenting tabular data. |
| 04:46 | <TabAtkins> | (I use the table layout module, for example, to produce an attractive two-column layout on my company's site.) |
| 04:53 | <drclue> | Well , that two column layout IMHO should have been the MULTICOL tag introduced originally as the vendor specific offering of Netscape. |
| 04:53 | <drclue> | This was a structural layout item and made sense. No matter the size of the display device the general layout was consistent. Using tables or kin |
| 04:53 | <drclue> | sorta requires one to decide the size and dimensions of the display which is really wrong. If one takes their time and distributes the work load properly the desktop rendering should by and large work just as well in almost any other medium |
| 04:55 | <TabAtkins> | Nah, multicol is wrong. For example, I want in some cases to be able to control the number of column based on the screen width - 1 column for small screens, 2 or 3 columns for wider screens. This has *nothing* to do with the document itself; it's entirely a styling issue. |
| 04:56 | <TabAtkins> | In fact, excellent example, the translation application for my company's software can list over 20 languages. The name of a language isn't very wide, so it's ideal to present in multiple columns, but obviously you don't want to make the column too small so it starts wrapping the name. That's just silly. |
| 04:57 | <TabAtkins> | So I do a little "ul.lang-list{ column-width: 20em; }" or whatever, and it automatically adjusts. |
| 04:57 | <drclue> | OK , so you want to contrive several versions of your output or inject function into presentation. I myself would not do same. I might have a collapse option on the multicol but other than that I would rather have the multicol element |
| 04:57 | <foolip> | boblet: hi there |
| 04:58 | <boblet> | hey hey |
| 04:58 | <foolip> | which browser appends a .part extension? |
| 04:58 | <TabAtkins> | If I were to change how I do this column-izing, though, that doesn't change the meaning of the document at all. I shouldn't be changing the document just because I'm presenting it differently. |
| 04:59 | <boblet> | foolip: FF 3.5x |
| 04:59 | <drclue> | OF course you should be able to have one well crafted presentation work across multiple platforms and dimensions which is where I always try to stand |
| 04:59 | <TabAtkins> | So yes, I do want to present several versions of my output, based on the output device. |
| 04:59 | <boblet> | foolip: also tested in Chrome 5.0x, but it didn’t seem to download the file (both Mac) |
| 04:59 | <foolip> | boblet: and iCal, is that on mac? |
| 05:00 | <roc> | FF appends a .part extension while the file is downloading |
| 05:00 | <roc> | when the download is complete, the .part extension is removed |
| 05:00 | <TabAtkins> | In many cases, sure, you want a single presentation because it's just easier. But there are also cases when the best/most natural/prettiest presentation is well-suited for a particular class of output devices, but needs some tweaks for otheer ones. |
| 05:00 | <boblet> | foolip: yes. renaming vcard to .vcs worked fine for Address Book |
| 05:00 | <foolip> | roc: this is a data: URL |
| 05:00 | <drclue> | I just want to have one presentation that through proper design renders well desktop , laptop,cellphone , audio |
| 05:00 | <boblet> | roc: the file is from a data url, and it’s not removing .part for me |
| 05:00 | <roc> | that sounds like just a bug |
| 05:01 | <foolip> | boblet: file bug at mozilla :) |
| 05:01 | <drclue> | I want that one presentation to occur within the standards |
| 05:01 | <boblet> | foolip: will add it to the todo pile :) |
| 05:02 | <TabAtkins> | You can certainly put together that one presentation. But that doesn't mean that one should be limited to that. |
| 05:02 | <foolip> | boblet: if either the vcard or ical output doesn't work in some major software, that may be a bug in the spec (assuming I follow it) |
| 05:02 | <boblet> | drclue: given the state of browsers across those devices, I think you’re going to be making something that’s lowest common denominator. <p> is well-supported :| |
| 05:02 | <TabAtkins> | For example, for a screen reader it's best to have the content come first, and all the miscellaneous navigation and such come afterwards. But if you do that with HTML and relatively basic CSS, you're forced to have the nav stuff either on the right column, or below the content. |
| 05:03 | <MikeSmithX> | othermaciej: fwiw, I just revived the twitter @html5 account (which had gone dormant since December when twitter apparently changed their rest API to reject any messages longer than 140, instead of auto-truncating them as it used to) |
| 05:03 | <othermaciej> | MikeSmithX: neat |
| 05:03 | <TabAtkins> | If you want it on the left column, you simply *can't*. You need to turn to CSS (right now employing some dirty hacks, in the future being able to use stuff like Template Layout to arrange things easily). |
| 05:04 | <drclue> | Your more than welcome to make as many presentations as you like ,and via media selectors in CSS you can make the subtle changes required to render a single presentation in multiple ways or as you prefer make multiple renderings. Specs need not make that choice for you |
| 05:04 | <TabAtkins> | What do you mean by "Specs need not make that choice for you"? |
| 05:06 | <foolip> | boblet: chrome seems to work if you right click and save as |
| 05:06 | <boblet> | foolip: oh, doh. didn’t try that |
| 05:07 | <foolip> | boblet: problem with data URLs is that they don't have a name, to make this sane one would have to bounce the data to a server-side script with the wanted name |
| 05:07 | <boblet> | foolip: probably not a new feature that’ll happen anytime soon huh ;-) |
| 05:07 | <drclue> | Specs should allow one to avoid having to make multiple base audience presentations but rather through as afforded via media selectors make common changes to multiple singular presentations via subtle suggestions in style |
| 05:07 | <TabAtkins> | I'm confused. That's the current state of affairs. What do you think is wrong? |
| 05:08 | <MikeSmithX> | othermaciej: along with watching for changes to the dev.w3.org/html5/spec/Overview.html (the HTML5 spec) and the drafts split out from it (microdata, 2dcontext, etc.) it also watches and notifies about changes to the h:tml draft, the html4-differences, rdfa, the alt-techniques doc from Steve Faulkner, the decision-policy doc, and the issue-status doc |
| 05:08 | <othermaciej> | MikeSmithX: sounds like a useful data source |
| 05:08 | <TabAtkins> | MikeSmithX: Yay, thanks! I was sad when @html5 stopped broadcasting. |
| 05:09 | <MikeSmithX> | othermaciej: if you want me to have it un-watch the decision-policy doc or issue-status doc, I can turn that off |
| 05:09 | <MikeSmithX> | TabAtkins: thank Twitter for changing their API :) |
| 05:09 | <othermaciej> | MikeSmithX: I'd say not the decision policy doc, but watching issue-status is good |
| 05:09 | <MikeSmithX> | OK |
| 05:09 | MikeSmithX | turns off decision-policy watching |
| 05:09 | <foolip> | boblet: well, no :) the main use for the tool is the "live" part after all. but do try importing the output into various software (I only tried gmail and google calendar) |
| 05:11 | <boblet> | ok! |
| 05:11 | <drclue> | What I think is wrong currently is that forays from the realms of one specification are reaching past reasonable overlap into the realms of others ,and that the established and proven separations of data, function and presentation formerly considered and proven to be good programming practice are being forgotten |
| 05:12 | <TabAtkins> | I think you're misunderstanding just what the specs are doing, and what roles they're playing. Likely you have a slightly out-of-date view of what each language is supposed to address, and thus are seeing more overreaching than actually exists. |
| 05:13 | <TabAtkins> | From what I can tell, you think along roughly the right lines, and the specs agree with you. |
| 05:13 | TabAtkins | is going to bed now. ;_; |
| 05:14 | <drclue> | After 40 years of watching the world go by , I notice these things. Not to be a penis war type of comment , but just a citation of the scars. |
| 05:14 | <drclue> | Those who have not learned from history are doomed to repeat it and it that repetition is something I think your missing. |
| 05:15 | <drclue> | No matter , I figure we are doomed to never actually agree and that is OK, diverse opinions ar ewhat make for the diversity the web is built upon |
| 05:23 | <drclue> | I should clarify the 40 years relates to the computing experience not my actual age, but whats a decade or so among friends :) |
| 05:34 | <drclue> | By and large I think that the HTML5 specification is viable and should despite it's issues move forward. If we waited for it to be perfect , no progress would be made at all in a time when we really do need progress. Old crutches like Adobe Flash need to be cast aside. Vacillations over minutia while great conversation and perhaps valid conversation for HTML5.1 or such pale in consideration of larger issues like the threats posed by Ruper |
| 05:34 | <drclue> | confines of HTML5 we can fix with respct to the various code bases the remaining issues |
| 06:07 | <boblet> | first peek at Microformats & Microdata slides from a talk I gave recently: http://oli.jp/slides/microformats-microdata/ |
| 06:08 | <boblet> | if anyone checks it out and has comments please leave here & I’ll check the logs later |
| 06:08 | <boblet> | otherwise via twitter @boblet |
| 06:08 | <drclue> | I'll buy into this conversation. Why would any microformat be a valid HTML spec conversation? |
| 06:09 | <boblet> | drclue: slides = presentation, not conversation :P |
| 06:10 | <boblet> | I’ve gotta go—conversational ticket sales will resume the next time I’m on |
| 06:10 | <boblet> | feel free to ask me again then |
| 06:10 | <drclue> | OK , you say it is a presentation , but I in looking at the specification still ask the question why any microformat would be a valid |
| 06:10 | <drclue> | component of an HTML specification |
| 06:11 | <boblet> | sigh |
| 06:11 | <boblet> | quickly |
| 06:11 | <boblet> | uF are not part of any HTML specification |
| 06:11 | <drclue> | sigh all ya like , I think it is valid conversation |
| 06:11 | <boblet> | sigh = I’m late |
| 06:12 | <boblet> | the closest uF have come to integration in HTML spec is microdata chapter |
| 06:12 | <boblet> | code examples of using vcard and vevent |
| 06:12 | <drclue> | There should be no reference to any spec of the moment , be it ones you cite or dBas III |
| 06:12 | <boblet> | uF specs themselves are made by microformats.org community = completely different |
| 06:13 | <drclue> | This discussion area is about HTML5 so why would any format be it vcard , dBase III or cobalt be relevant |
| 06:13 | <boblet> | as to why spec uF to start with, because they’re useful, in the same way that the HTML spec is useful |
| 06:13 | <boblet> | following it gives common ground on which to build |
| 06:13 | <boblet> | eg Google Rich Snippets |
| 06:13 | <boblet> | eg Technorati uF search |
| 06:14 | <boblet> | if uF don’t float your boat, there’s no requirement to use em |
| 06:14 | <boblet> | and I’m out |
| 06:14 | <boblet> | l8r all |
| 06:14 | <drclue> | Useful and spec related are not the same thing despite any inroads one might have made |
| 06:14 | <boblet> | again—next time |
| 06:15 | <drclue> | OK next time as you like, but HTML should not in any way be tied to a non w3.org spec which itself is not related to the spec in a timeless manner. |
| 06:18 | <drclue> | There could be any number of hangers on that might want to promote a standard by tying it to a rising star. This should in no way be allowed. |
| 06:18 | <drclue> | in general true w3.org standards are timeless and as should be immune from specs of the momment |
| 06:21 | <foolip> | drclue: HTML makes no reference to microformats |
| 06:22 | <foolip> | drclue: microformats piggyback on the class attribute mostly |
| 06:22 | <drclue> | Most w3 standards can cite years if not decades of service and interact with other standards proven not only by the moment but years and decades of service |
| 06:22 | <drclue> | to lower the bar of admission lowers the quality and consideration of the standards themselves and makes them but a of the moment straw poll of the latest music artist or movie star and totally underming the value of having specifications at all |
| 06:23 | <drclue> | I was reading the standard just today and saw such references , so while I may be wrong , they are indeed referenced and in no way should be |
| 06:24 | <foolip> | drclue: please file specific bugs on those sections |
| 06:25 | <foolip> | drclue: if you want good flame material "microdata"is the keyword, not microformats |
| 06:28 | <drclue> | It is a shame that your position should rely on minor typos or terms of reference. These are not bugs but rather invalid intermingling of outside standards of the moment with official standards |
| 06:32 | <foolip> | All problems in the spec are treated as bugs. |
| 06:32 | <foolip> | Still, please list which "outside standards" it is "invalid" to reference |
| 06:34 | <drclue> | Well , if you prefer that problems in th spec be called bugs , I will most certainly allow for that, but I will not waste my time parsing it in that fashion which is my choice just as it is your choice to send to committee that point |
| 06:35 | <drclue> | A reference to a spec that is not in itself a valid w3.org spec is in my opinion invalid |
| 06:35 | <foolip> | how about RFCs? |
| 06:36 | <foolip> | lots of W3C specs I've read reference RFCs |
| 06:37 | <drclue> | If the w3.org should allow a non spec to masquerade as spec that would be invalid too , and as to the RFC's , it should be as valid as RFC 1149 |
| 06:38 | <foolip> | what constitutes a "non spec"? |
| 06:39 | <drclue> | That is up to interpretation but IMHO a true spec is timeless and not effected by the moment |
| 06:40 | <foolip> | an you give an example of a true spec and a false spec? |
| 06:40 | <drclue> | citing RFC's is like citing the Avion carrier protocool |
| 06:40 | <foolip> | anyway, please file bugs for the specific things in the spec you think are wrong, it's not possible to make spec changes based on general notions like this. |
| 06:41 | <drclue> | The IP protocol is a true spec, in which a gaol is achieved and no axe is ground. A false spec is pretty much the rest |
| 06:44 | <drclue> | Spec changes can indeed be made by the beliefs I hold and which have been both derived and reflected by the historical authorings of the w3.org and the attempts to send such perceptions to committee just make me laugh |
| 06:45 | <foolip> | which committee? |
| 06:46 | <drclue> | IF you [foolip] are in a position to undermine the w3.org , I'm stuck , and as I have since 1994 will comply. Otherwise your opinon does not hold a dime more value than mine and we are both free to express our opinions |
| 06:47 | <foolip> | drclue: you greatly overestimate my power to undermine much of anything |
| 06:48 | <drclue> | Hey than you are just like me , someone with an opinion. You have yours , I have mine |
| 06:49 | <foolip> | drclue: you're in the IRC channel of the WHATWG, if you think something in the spec is in violation of some W3C policy then bring it to the W3C (participtation is open to the public) |
| 06:49 | <drclue> | You want to debate your opinion versus mine , great , lets do it. It's not a mater of anything but dueling opinions |
| 06:50 | <drclue> | Lets ask Hixie if it is fair game to discuss our opinions here, what you thing the answer will be? |
| 06:51 | <foolip> | of course it's OK, but we're not getting anywhere |
| 06:52 | <foolip> | which is why I encourage you to be specific, so that other can understand your position. |
| 06:52 | <drclue> | I was not asking if we were getting anywhere , but rather addressing your opinion that I must express such opinions elswhere. Are you done in that regard? |
| 06:53 | <drclue> | What do you think I have not been specific about? |
| 06:53 | <foolip> | you don't *have* to do anything, I'm only telling you what ways of leaving feedback are more likely to have a result |
| 06:54 | <drclue> | I've had conversation with Hixie and while I do not demand he do anything , I am inded allowed to express myself |
| 06:55 | <drclue> | If you want to debate a point , hixie reads the conversation so lets indeed debate |
| 06:56 | <foolip> | You haven't said anything specific enough to be possible to debate, so I will decline the offer. |
| 06:59 | <Dashiva> | I still have no idea what you're talking about either |
| 06:59 | <Dashiva> | It does remind me of the talk about letting page authors set the accept header... |
| 07:00 | <drclue> | Well , I said what I have said , which prompted you to respond , if indeed you have no position on the issue I expressed than indeed that is what you have to say |
| 07:04 | <hsivonen> | things would be so much easier if html had an intelligently designed internal encoding declaration mechanism instead of an evolved one |
| 07:06 | <drclue> | It would be nice if in hind site that the original HTML allowing scientist to exchange ideas had the for site to address the applications it evolved to. |
| 07:06 | <Dashiva> | hsivonen: Kind of like, not letting HTTP headers override document metadata? |
| 07:07 | <drclue> | If wishes were horses beggars would ride |
| 07:07 | <hsivonen> | like not being able to have charset-sensitive stuff before the declaration |
| 07:09 | <drclue> | If your on a totally different vein then that I'm discussing asking if the broader scope should override the specific scope I disagree. a More specific scope should override a broader scope |
| 07:10 | <MikeSmithX> | geez.. http://html5.org/tools/web-apps-tracker?from=4781&to=4782 |
| 07:11 | <MikeSmithX> | is it really worth going to the trouble of having warnings emitted for the "dataformatas" attribute? |
| 07:11 | <Dashiva> | Warnings will be emitted anyway |
| 07:12 | <Dashiva> | This way the warning makes more sense |
| 07:12 | <drclue> | Not even sure what that cited link has to do with my opinions other than wasting my time |
| 07:12 | <MikeSmithX> | that and whatever the hell these other bizarro phantom attributes are |
| 07:12 | <Dashiva> | event and for are IE-isms |
| 07:13 | <Dashiva> | Although I think other browsers support event=load for=window |
| 07:13 | <MikeSmithX> | I see |
| 07:13 | <othermaciej> | Dashiva: I think only IE supports <script for> |
| 07:13 | <othermaciej> | maybe Opera too |
| 07:13 | <othermaciej> | I know Safari doesn't |
| 07:13 | <drclue> | Discussing IE in regard to specs is like discussing freedom in a jail cell |
| 07:13 | <Dashiva> | Not the general case, but onload |
| 07:13 | <hsivonen> | which other browsers? |
| 07:13 | <hsivonen> | oh |
| 07:14 | <hsivonen> | i thought gecko had support but no longer |
| 07:14 | MikeSmithX | only just now notices Hixie "yes, we're scraping the bottom of the barrel now" comment in that commit description |
| 07:15 | <drclue> | Any discussion that resolves to IE is indeed scraping the ultimate bottom of the barrel |
| 07:15 | <hsivonen> | I suspect this is about making a point about Larry's and Julian's requests |
| 07:15 | <MikeSmithX> | hmm |
| 07:16 | <MikeSmithX> | so if Hixie feels like his time is being wasted, the solution is to spread the waste of time more widely so that others share the pain |
| 07:17 | <Dashiva> | Or so others are motivated to help stop the pain |
| 07:17 | <drclue> | IE is irrelevant. The capabilities of HTML5 will make IE to browsing like MS paint is to desk top publishing. At this juncture in life lets all laugh loudly and move on |
| 07:18 | <Dashiva> | Right. |
| 07:20 | <MikeSmithX> | anyway, apologies for my non-sequitor posting of that link and interrupting the previous conversation.. here's a link that I think might be more relevant to that: http://logopoeia.com/wisdom/ |
| 07:21 | <Dashiva> | I don't think you actually interrupted anything... |
| 07:21 | <drclue> | I really hate having to read links, and I hate it more when after doing so my time is wasted |
| 07:22 | <MikeSmithX> | amen to that |
| 07:23 | <Dashiva> | MikeSmithX, stop using your magical powers to force people to read links |
| 07:24 | <drclue> | Hy , if folks have VALID OPINIONS on the topic at hand , I'll trade words till I die, but if there was a nuclear button for folks with no F**King opinion just wasting my time I would be dancing on it |
| 07:25 | <zcorpan> | othermaciej: your impl of details is surprisingly similar to mine |
| 07:25 | <zcorpan> | othermaciej: except you don't check for native support and don't check which key or button was pressed |
| 07:26 | <zcorpan> | well and i listened for clicks on document |
| 07:26 | <MikeSmith> | Dashiva: OK, point taken |
| 07:27 | MikeSmith | turns off his personal Ultra Low Frequency long-range communications transmitter |
| 07:27 | <Dashiva> | It looks like firefox still supports for=window event=onload (but inline, not onload) |
| 07:27 | <drclue> | [MikeSmith] Please go somewhere for an extended period of time before I violate some law |
| 07:29 | <othermaciej> | zcorpan: eww, why would you listen for clicks on document? |
| 07:29 | <othermaciej> | zcorpan: yeah my implementation was lazy |
| 07:29 | <MikeSmith> | drclue, OK, please just don't violate me. I prefer to be more like, say, ravished |
| 07:30 | <Dashiva> | 犯される方がいい |
| 07:31 | <MikeSmith> | Dashiva: you always have a way with words |
| 07:31 | <MikeSmith> | great pick-up line, that |
| 07:31 | <othermaciej> | zcorpan: I also forgot to add a :active state for the triangle |
| 07:31 | <MikeSmith> | I'll have to try that one out |
| 07:32 | <Dashiva> | I didn't know you needed pickup lines |
| 07:32 | <othermaciej> | zcorpan: I also wanted to make it animate but likewise ran out of steam - not sure it can be done with just CSS transitions |
| 07:32 | <Dashiva> | I figured there was already a line of people wanting to be picked up |
| 07:32 | <othermaciej> | zcorpan: actually, maybe it can through grotesque hackery - wonder if my idea would flicker |
| 07:36 | <drclue> | [MikeSmith] Bring it , I will ravish it, if for no other rason than you will be quite for 2,4,6,8 hours |
| 07:37 | <zcorpan> | othermaciej: i had a listener on document for other things as well |
| 07:38 | <zcorpan> | othermaciej: i sent an email to www-style about transitioning between height:0 and height:auto |
| 07:38 | <zcorpan> | well height:1em and height:auto should work too |
| 07:38 | <othermaciej> | zcorpan: I gotta convince Simon to just implement that for me |
| 07:38 | <othermaciej> | zcorpan: but I did think of a way to fake it |
| 07:38 | <othermaciej> | zcorpan: height: 1em is kinda bogus though |
| 07:39 | <zcorpan> | why? |
| 07:39 | <othermaciej> | the summary could be taller than 1em |
| 07:39 | <zcorpan> | true |
| 07:39 | <zcorpan> | but not if you know it's just one word |
| 07:40 | <othermaciej> | if it has a different font than the details its height won't be 1em in <details> space |
| 07:47 | <MikeSmith> | hsivonen: datasrc, datafld, and dataformatas are the first global attributes that have been added to the "Non-conforming features" section. if we use the same mechanism of making them valid in legacy.rnc (and emitting the errors in the assertions code) it presents a new problem: the also need to be added to the embed.attrs.other pattern |
| 07:48 | <MikeSmith> | there are not currently any obsolete-and-nonconforming attributes in embed.attrs.other |
| 07:48 | <hsivonen> | MikeSmith: an alternative is filing a spec bug about this and force a decision against making all legacy stuff conforming |
| 07:49 | <MikeSmith> | hmm |
| 07:49 | <drclue> | F**K this spec bug shit , if there is an issue lets all have it out , the parties that be read this IRC thing probably more than the bug spec thing |
| 07:50 | <othermaciej> | hsivonen: they're nonconforming already, aren't they? |
| 07:50 | <MikeSmith> | hsivonen: or I can just hold off and wait for the drama to play out |
| 07:50 | <othermaciej> | these particular features, that is |
| 07:50 | <drclue> | I think there might be disagreements , but real bugs ,lets hear the song |
| 07:52 | <MikeSmith> | script@for and script@event and table@datapagesize are easy enough to add, so maybe just add specific assertions error messages for those for now and leave the datafld, etc. stuff for later |
| 07:52 | <MikeSmith> | drclue, help commands |
| 07:53 | <othermaciej> | yeah, definitely in the non-conforming section |
| 07:53 | <hsivonen> | MikeSmith: that's one option if you can get someone else to play the drama opposition |
| 07:53 | <drclue> | [MS] ya ain't near enough to obey |
| 07:53 | <othermaciej> | MikeSmith: are you just making sure that all explicitly nonconforming attributes have a custom error message instead of the generic one? |
| 07:54 | <hsivonen> | MikeSmith: if these are at all inconvenient due to the <embed> stuff, I'd just not add any code for these |
| 07:54 | <MikeSmith> | othermaciej: yeah, for consistency with a change made a while back to emit custom error messages for all attributes listed in that section |
| 07:54 | <MikeSmith> | hsivonen: OK |
| 07:55 | <drclue> | HTML5 aint perfect by a long shot , but as a means to move forward , it's ok |
| 07:55 | <zcorpan> | is <embed datafld> conforming? |
| 07:55 | <MikeSmith> | othermaciej: the spec does not explicitly say that the conformance checkers should emit specific warnings for obsolete-and-non-conforming attributes, but it seems useful for end-users to have them |
| 07:56 | <zcorpan> | i guess it's not since the obsolete section says it must not be used |
| 07:56 | <MikeSmith> | especially given the un-usefulness of the generic messages that jing emits for them |
| 07:56 | <othermaciej> | MikeSmith: in general, yeah - for the datawhatever attributes, maybe not so much, since they only ever existed in the world of imagination |
| 07:56 | <drclue> | HTML5 tries to please too many people , but forward versus standstill it rocks |
| 07:57 | <MikeSmith> | othermaciej: good point |
| 07:57 | <hsivonen> | what does IE do with dataformatas? |
| 07:57 | <MikeSmith> | though the "world of imagination" sounds like a great place to be |
| 07:57 | <drclue> | The further IE falls behind , the better |
| 07:57 | <hsivonen> | "As of Windows Internet Explorer 7, the DATAFORMATAS attribute is no longer supported for the input type=button object." |
| 07:57 | <hsivonen> | http://msdn.microsoft.com/en-us/library/ms533706%28VS.85%29.aspx |
| 07:58 | <othermaciej> | I guess I was wrong about the world of imagination |
| 07:58 | <othermaciej> | I did not realize these were from the world of old IE versions |
| 07:58 | <MikeSmith> | a much less invitingly-named world |
| 07:59 | <zcorpan> | MikeSmith: the schema currently allows <embed datafld> so i think it would give the desired result to add these as assertions and not change the schema |
| 07:59 | <hsivonen> | I guess that would be doable |
| 08:00 | <drclue> | Microsoft has so many contractual reasons to lag behind I could not even write them in a single night. The only way out is to be so far ahad |
| 08:00 | <drclue> | as to blow them off the screen |
| 08:01 | hsivonen | wonders if MS is still making long-term promises about IE8 to BigCos |
| 08:03 | <MikeSmith> | zcorpan: but datafld is a global attribute, so not changing the schema would mean that if datafld were used on an element other than embed, two error messages would be emitted (the generic one from jing, plus the assertions one) |
| 08:04 | <hsivonen> | MikeSmith: the assertion could be <embed>-specific |
| 08:04 | <MikeSmith> | I see |
| 08:05 | <gsnedders> | TabAtkins: No, I don't I tend to avoid graph theory at all costs. |
| 08:06 | gsnedders | used canvas for the first time last night |
| 08:08 | <zcorpan> | hsivonen: but then <p datafld> would give the jing message? |
| 08:23 | <MikeSmith> | interesting.. Mac OSX update seems to have removed my JDK install |
| 08:24 | <hsivonen> | zcorpan: yeah |
| 08:36 | gsnedders | still isn't sure he got his internal representation of the hexagonal grid right |
| 08:38 | <zcorpan> | hsivonen: then maybe it's not worthwhile to special-case embed; most people who use datafld probably use it on div or so anyway |
| 08:39 | <zcorpan> | but wouldn't it be relatively easy to add these as global attributes in the schema and then have a global assertion? |
| 08:43 | <MikeSmith> | zcorpan: it's easy to add them to the schema, it's just that currently all the obsolete attributes are in a legacy.rnc file that's not in the "valid" schema repository but is copied over by the build |
| 08:44 | <MikeSmith> | in consideration of the case of uses of the schema outside of v.nu |
| 08:44 | <MikeSmith> | so the legacy.rnc file is just something that v.nu uses internally |
| 08:44 | <MikeSmith> | but the embed.attrs.other pattern is not in the legacy file |
| 08:44 | <MikeSmith> | it's part of the valid schema |
| 08:46 | <MikeSmith> | and the particular nature of the embed.attrs.other pattern is such that there's no way to just append attribute names to in in the legacy file |
| 08:47 | <MikeSmith> | the legacy file would need to basically completely refine it in whole |
| 08:47 | <MikeSmith> | and it's a long list of attribute names |
| 08:48 | <MikeSmith> | and maintaining it in two different places would be error-prone |
| 08:48 | <zcorpan> | i see |
| 08:48 | zcorpan | suggests doing whatever is simplest :) |
| 08:48 | <hsivonen> | hmm. I just noticed that Safari on my system has tabs under the location bar |
| 08:48 | <hsivonen> | didn't some beta at least have tabs in the title bar? |
| 08:48 | jgraham | discovers that pypy is also using html5lib as a benchmark (I guess they inherited it from Unladen Swallow) |
| 08:49 | <zcorpan> | hsivonen: safari 4 beta had them in the title bar but safari 4 final changed it back, i think |
| 08:49 | <hsivonen> | zcorpan: I see. that explains it. |
| 08:49 | <hsivonen> | http://www.xqsharp.com/xqsharp/samples/raytracer/index.htm |
| 08:49 | <gsnedders> | jgraham: We ought to get serializer test there too |
| 08:49 | <gsnedders> | Haven't tried benchmarking that yet, though |
| 08:51 | <MikeSmith> | hsivonen: on OSX, do you use the macports openjdk, or something else? |
| 08:52 | <hsivonen> | MikeSmith: I use Apple's JDK6 |
| 08:52 | <MikeSmith> | I did a "port install openjdk6" and it seems to be installing all kind of X11 dependencies for some reason |
| 08:52 | <MikeSmith> | OK |
| 08:52 | <jgraham> | gsnedders: We hould build some better benchmarks in general, I guess |
| 08:53 | <MikeSmith> | hsivonen: so what do you set your JAVA_HOME to? |
| 08:53 | <othermaciej> | hsivonen: we had titlebar tabs for a while - it did not end up well-loved |
| 08:55 | <hsivonen> | MikeSmith: /Library/Java/Home |
| 08:56 | <MikeSmith> | ok, thanks |
| 09:12 | <zcorpan> | Philip`: if you're going to research [ in doctype any more, maybe you should look for any garbage there |
| 09:13 | <zcorpan> | Philip`: it's possible people use things like <!doctype html public "" !> and expect a certain mode |
| 09:54 | <hsivonen> | I have very little sympathy for authors who aren't using a doctype from http://hsivonen.iki.fi/doctype/#choosing and are trying to be clever with SGMLisms |
| 09:55 | <othermaciej> | I don't really understand the point of that kind of clever sgmlism |
| 09:55 | <othermaciej> | if you use an SGML internal subset to violate conformance rules, then doesn't that render the conformance result from the validator meaningless? |
| 09:55 | <hsivonen> | othermaciej: the point is that you want to use custom attributes and you want a validator say they are valid |
| 09:55 | <hsivonen> | it's the data-* use case |
| 09:55 | <othermaciej> | right but you're basically ordering the validator to lie to you |
| 09:56 | <hsivonen> | othermaciej: you might still want to catch typos even if you don't want the validator to complain about dojoFoo |
| 09:56 | <hsivonen> | but using the internal subset is an even worse idea than using a home-grown public id |
| 09:57 | <hsivonen> | or system id, rather |
| 10:10 | <Philip`> | zcorpan: http://philip.html5.org/data/doctypes.html#%3c%21doctype_html_public_%22-%2f%2fw3c%2f%2fdtd_xhtml_1.0_transitional%2f%2fen%22_system_%22http%3a%2f%2fwww.w3.org%2ftr%2fxhtml1%2fdtd%2fxhtml1-transitional.dtd%22%2f%3e |
| 10:10 | <Philip`> | zcorpan: http://philip.html5.org/data/doctypes.html#%3c%21doctype_html_public_%22-%2f%2fw3c%2f%2fdtd_html_4.01_transitional%2f%2fen%22_%3chtml%3e |
| 10:11 | <Philip`> | zcorpan: http://philip.html5.org/data/doctypes.html#%3c%21doctype_html_public_%22-%2f%2fw3c%2f%2fdtd_html_4.01_transitional%2f%2fen%22_%2f%3e |
| 10:11 | <Philip`> | etc |
| 10:11 | <Philip`> | Those seem like the most common garbages |
| 10:11 | <MikeSmith> | hsivonen: what kind of PC did you order? |
| 10:13 | <hsivonen> | MikeSmith: Shuttle sx58h7, i7 950, Intel SSD drive, 6 GB RAM (I wanted more but the supply chain sucks), passive-cooled nvidia GPU |
| 10:14 | <MikeSmith> | wow |
| 10:14 | <hsivonen> | MikeSmith: hoping to run hg, make and gcc faster |
| 10:14 | <MikeSmith> | well that outta do it I guess |
| 10:14 | <hsivonen> | MikeSmith: wow in what sense? |
| 10:14 | <MikeSmith> | seems like a lot of spec |
| 10:14 | <MikeSmith> | I guess I had been thinking you meant a new laptop |
| 10:15 | <hsivonen> | it's ridiculously hard to get a small enclosure for i7 |
| 10:15 | <hsivonen> | and the supply situation for the barebone, for RAM and for passive-cooled nvidia GPUs are also ridiculously bad |
| 10:15 | <hsivonen> | s/are/is/ |
| 10:16 | <MikeSmith> | I guess if you're doing a lot of serious merging and building, the machine performance can make a significant difference |
| 10:17 | <hsivonen> | well, the alternative fix would be making C++ work with incremental builds like Java |
| 10:17 | <hsivonen> | but that's not gonna happen |
| 10:17 | <hsivonen> | it's crazy to have to have faster hardware because of the C preprocessor and the undecidability of C++ syntax |
| 10:19 | <hsivonen> | everything except maybe the thermal characteristics of i7 950 would fit in a smaller box. It's pretty annoying that smaller boxes aren't available |
| 10:19 | <MikeSmith> | hsivonen: looking at the retail price for it, I see it's pretty reasonable |
| 10:20 | <hsivonen> | I guess the vendors assume that anyone who wants a fast CPU has to want multiple drive bays and card slots and a dedicated GPU instead of an integrated one |
| 10:22 | <MikeSmith> | I suppose needs of average users in their market are bit different than needs of code developers |
| 10:23 | <hsivonen> | it's basically Atom for non-gamers and huge hardware eating a lot of watts for gamers |
| 10:26 | <virtuelv> | hsivonen: until atom is multicore, and the cpu plus the rest of the hardware doesn't choke on fullscreen flash, then atom is not it |
| 10:27 | <hsivonen> | virtuelv: atom is dualcore |
| 10:28 | <hsivonen> | I don't recall if I tested full-screen flash on my grandfather Atom box, but Google Earth and Compiz worked great |
| 10:28 | <hsivonen> | using the integrated Intel GPU |
| 10:28 | <virtuelv> | hsivonen: that depends on which particular atom chip is the flavor of the day |
| 10:28 | <virtuelv> | hsivonen: flash is a pig |
| 10:28 | <virtuelv> | and you can quote me on that |
| 10:28 | <hsivonen> | s/grandfather/grandfather's/ |
| 10:29 | <virtuelv> | I mean, Grooveshark, nice as it is, makes my laptopcore2 cpu speed up |
| 10:29 | <virtuelv> | s/pco/p's co/ |
| 10:30 | <hsivonen> | virtuelv: clearly, Grooveshark should use HTML5 and Vorbis |
| 10:32 | <hsivonen> | once I get the new box, I was thinking of not installing Flash and see how far I can go without it |
| 10:32 | <hsivonen> | though maybe I have to have Flash to get representative performance metrics |
| 10:32 | <hsivonen> | from Firefox that is |
| 10:33 | <virtuelv> | hsivonen: not all of youtube's content seems to be available in html5 video |
| 10:33 | <virtuelv> | (or I haven't been able to make it work, at least) |
| 10:33 | <virtuelv> | and there's the entire needing x264-thing |
| 10:36 | <zcorpan> | Philip`: thanks |
| 10:36 | zcorpan | doesn't have time to analyze today |
| 10:43 | <MikeSmith> | hsivonen, zcorpan - well, after a little more poking around and thanks to finding wakaba_'s http://suika.fam.cx/~wakaba/wiki/sw/n/dataformatas page, I notice that the dataformatas, etc., attributes were never global in HTML4 anyway |
| 10:43 | <MikeSmith> | the reserved PE was only referenced by span, div, input, select, textarea, button, table, object |
| 10:45 | <MikeSmith> | so it seems the spec should list those explicitly instead of stating "...on any element" |
| 10:48 | <zcorpan> | MikeSmith: file a spec bug :) |
| 10:49 | <MikeSmith> | I guess I'll just reopen bug 9000 |
| 10:50 | <zcorpan> | btw for people giving talks on html5, i can recommend picking one subject and diving in instead of trying to cover everything in html5 |
| 10:52 | <zcorpan> | my talk on html5 video felt a lot more useful than my previous talks on html5 because i didn't have to grossly cut corners |
| 10:54 | <zcorpan> | maybe i should convert the talk to an article (in english) |
| 11:12 | <gsnedders> | zcorpan: Or was it just a better talk because you didn't have to speak a strange, weird, language like English? |
| 11:22 | <MikeSmith> | an updated PER draft of XHTML 1.1 was published last week |
| 11:22 | <MikeSmith> | http://www.w3.org/MarkUp/2010/PER-xhtml-modularization-20100213/ |
| 11:23 | <MikeSmith> | there's a diff-marked version at http://www.w3.org/MarkUp/2010/PER-xhtml-modularization-20100213/xhtml-modularization-diff.html |
| 11:23 | <gsnedders> | Hmm, produced by a WG that doesn't exist. |
| 11:23 | <gsnedders> | Odd. |
| 11:24 | <Philip`> | WGs with unfinished business hang around as ghosts |
| 11:24 | <MikeSmith> | the group remains in existence only until this final set of deliverables get published |
| 11:25 | <gsnedders> | On grounds that everything must become a REC or NOTE? |
| 11:25 | <MikeSmith> | gsnedders: yeah, basically |
| 11:26 | <MikeSmith> | and on grounds that there apparently are some organizations that want a final XHTML 1.1 Rec |
| 11:27 | <MikeSmith> | so, anyway, given that it is going to be published, it would nice to make sure there's nothing in it still that's broken |
| 11:27 | <MikeSmith> | so I would appreciate it if anybody with interest/time could check and let me know if you find any problems |
| 11:27 | <MikeSmith> | by beginning of next week |
| 11:28 | <MikeSmith> | they fixed the longstanding usemap datatype problem |
| 11:28 | <Philip`> | The http://ln.hixie.ch/?start=1172653243&count=1 one? |
| 11:29 | <MikeSmith> | Philip`: yeah |
| 11:29 | <MikeSmith> | another they fixed was the problem that the XHTML 1.1 DTD restricted the value of class to NMTOKENS, which meant class="" was not valid in XHTML 1.1 (though it is in HTML4 and XHTML1.0 |
| 11:30 | <hsivonen> | good thing the WHATWG doesn't have the kind of Process that makes a datatype fix to one attribute take a decade |
| 11:31 | <MikeSmith> | well, it was a decade of non-fixing it |
| 11:31 | <MikeSmith> | and probably 10 minutes of actually fixing it |
| 11:32 | <Philip`> | http://www.w3.org/MarkUp/2010/PER-xhtml-modularization-20100213/abstract_modules.html#s_commonatts |
| 11:32 | <Philip`> | "class (NMTOKENS)" |
| 11:32 | <Philip`> | Looks like the forgot to change that one |
| 11:32 | <Philip`> | *they |
| 11:33 | <annevk> | why does XHTML 1.1 matter? |
| 11:33 | <annevk> | it's obsolete by XHTML5, no? |
| 11:33 | <MikeSmith> | Philip`: thanks for catching it, I'll pass it on |
| 11:34 | <MikeSmith> | annevk: it matters because that spec is going to be published, and this is the last opportunity to get it while there's an actual group chartered to make changes to it |
| 11:34 | <MikeSmith> | *to get it fixed |
| 11:35 | <hsivonen> | annevk: apparently the epub thing that Apple and Adobe use depends on XHTML 1.1 specifically |
| 11:35 | <hsivonen> | not 1.0 |
| 11:35 | <hsivonen> | not 5 |
| 11:35 | <annevk> | i don't see how that matters, nobody is going to conform to XHTML 1.1 |
| 11:35 | <Philip`> | (Changing to xs:string seems to do a lot more than merely allowing an empty value, since it removes the restrictions that tokens must be Names) |
| 11:36 | <hsivonen> | annevk: see http://blog.threepress.org/2009/11/28/what-i-would-change-about-epu/ |
| 11:36 | <hsivonen> | annevk: I don't expect anyone to make an XHTML 1.1 -compliant ebook reader |
| 11:37 | <hsivonen> | has it been revealed if the reader on iPad is based on WebKit or has a simpler CSSless epub-specific different engine? |
| 11:37 | <gsnedders> | Or just WebKit without CSS |
| 11:37 | <MikeSmith> | Philip`: that's the original datatype the attribute had in HTML4 and XHTML 1.0 anyway, I think |
| 11:38 | <MikeSmith> | or equivalent of it |
| 11:38 | <annevk> | at some point all the old HTML specs should just be marked as obsolete |
| 11:38 | <annevk> | similarly to how RFCs are marked obsolete at some point |
| 11:39 | <Philip`> | Most HTML5 drafts are obsolete within a few hours of publication, so maybe they should all be marked as obsolete by default |
| 11:39 | <gsnedders> | Is there any way to detect clicks in a certain bit of a canvas |
| 11:40 | <Philip`> | Yes |
| 11:40 | <Philip`> | Use onclick, then write some scripts to test the position |
| 11:40 | <gsnedders> | OK, so that's the only way |
| 11:40 | <Philip`> | potentially using isPointInPath to help |
| 11:42 | <gsnedders> | If I have a hex grid, how can I tell which hex it has hit? Any idea of what the easiest way to do that is/ |
| 11:43 | <Philip`> | Loop through each grid tile, check whether the point is inside its hexagon or not |
| 11:43 | <gsnedders> | That seems awfully slow |
| 11:43 | <Philip`> | (which is like six dot products per tile) |
| 11:44 | <Philip`> | (or an isPointInPath which isn't much faster) |
| 11:44 | <gsnedders> | It seems like there should be some constant time way of doing it |
| 11:44 | <gsnedders> | Philip`: The dot product? How does that work? |
| 11:45 | <jgraham> | Well sure if you literally just have a grid you can calculate which hexagon you are in with some formula |
| 11:45 | <jgraham> | But you said "easiest" |
| 11:45 | <Philip`> | For constant time, I guess you could calculate x mod hex-width, y mod hex-height, and then do a few line comparisons to work out which relative tile it's in |
| 11:46 | <gsnedders> | s/easiest/easiest but not gratutiously expensive/ |
| 11:46 | <Philip`> | and then add that to floor(x / hex-width) etc |
| 11:46 | <Philip`> | or something along those lines |
| 11:46 | <gsnedders> | http://www.gamedev.net/reference/articles/article747.asp#MOUSE |
| 11:47 | <Philip`> | gsnedders: If the point is outside the hexagon, then there's a side of the hexagon which you could extend to infinity and the point would be on the outside of that line |
| 11:47 | <Philip`> | and you can detect whether a point is outside by doing a dot-product with the perpendicular and checking the sign, or something equivalent |
| 11:47 | <Philip`> | so you just do that for each side |
| 11:47 | <Philip`> | and if it's not outside any side, it must be inside the shape |
| 11:48 | <Philip`> | (Should work for any convex shape, I believe) |
| 11:49 | gsnedders | will someday learn maths |
| 11:49 | <jgraham> | Philip`: (the concept of being "outside a line" doesn't make much sense) |
| 11:50 | <jgraham> | (on its own) |
| 11:50 | <Philip`> | jgraham: "On the opposite side of the line to the center of the shape which we are considering" |
| 11:52 | <Philip`> | gsnedders: Anyway, probably quickest to wrap the mouse coordinates to the smallest rectangular tiling region (which won't be the size I said, because I was stupid) and just do half a dozen line comparisons in there |
| 11:52 | <Philip`> | ...like that gamedev.net thing except use maths instead of bitmaps to determine the region |
| 11:53 | <Philip`> | (By "half a dozen" I think I mean "five") |
| 11:56 | <MikeSmith> | Philip`: http://lists.w3.org/Archives/Public/www-html-editor/2010JanMar/0002.html |
| 11:57 | <Philip`> | MikeSmith: Thanks |
| 11:57 | <MikeSmith> | if you (or anybody else) find other problems, probably best to post comments to www-html-editor, and either Cc or forward a copy to me |
| 11:58 | <Philip`> | Presumably it's too late to address any real problems, so only trivial things are useful? |
| 12:06 | <MikeSmith> | no, definitely not too late to address any real problems |
| 12:06 | <MikeSmith> | anything is fair game |
| 12:06 | <Philip`> | Like "processing requirements are not specified for anything"? |
| 12:06 | <MikeSmith> | Philip`: ↑ |
| 12:07 | <MikeSmith> | ah |
| 12:07 | <MikeSmith> | that might be worth sending a comment to www-html-editor at least |
| 12:07 | <MikeSmith> | to have it on the record |
| 12:08 | <MikeSmith> | and I can make a personal list of comments to pass on |
| 12:09 | <MikeSmith> | but yeah, that's not a deficiency that's likely to get resolved |
| 12:09 | <Philip`> | I can't imagine there'd be any practical value in suggesting "the spec should be rewritten to be like HTML5" |
| 12:10 | <MikeSmith> | there might be some value in seeing what kind of response it would elicit |
| 12:10 | <MikeSmith> | I'm not sure that value would qualify as practical |
| 12:11 | <Philip`> | Probably not a useful use of anyone's time, then :-) |
| 12:13 | <Philip`> | (As a specific example, saying usemap is a URIREF doesn't answer questions like what happens with <img usemap="#%"/><map name="%"/><map id="%"/><map name="%"/>) |
| 12:14 | <Philip`> | (but I think people who care about that don't care about XHTML 1.1, and people who care about XHTML 1.1 don't care about that) |
| 12:15 | <Philip`> | (so everyone can be happy) |
| 12:15 | <MikeSmith> | Philip`: yeah, seriously, that's probably an accurate way of describing the situation |
| 12:17 | <Philip`> | (I don't even care enough to comment that a comment in the DTD still refers to "a usemap IDREF") |
| 12:19 | MikeSmith | tries to decide if he needs to care enough to notice Philip`'s non-comment |
| 12:21 | <Philip`> | All specs ought to have a built-in bug reporting feature similar to the HTML5 one, so that complaining via a bug report is no harder than complaining via IRC |
| 12:23 | <MikeSmith> | I wonder how much work it is to port that feature to other documents |
| 12:24 | <MikeSmith> | I could add it to all the various dev.w3.org/html5 drafts at least |
| 12:24 | <Philip`> | Perhaps the hardest part is convincing people that it could possibly work |
| 12:25 | <MikeSmith> | just adding it first and having it working would do a pretty good job of that |
| 12:26 | <MikeSmith> | wow, the marriage of TeX and Javascript |
| 12:26 | <MikeSmith> | one of the lesse-known William Blake works |
| 12:26 | <MikeSmith> | but also now this: |
| 12:27 | <MikeSmith> | http://ajaxian.com/archives/tex-line-breaking-algorithm-in-javascript |
| 12:27 | <hsivonen> | MikeSmith: ouch. it's even more alarming that text layout is being worked around in canvas than that lack of flexbox is. |
| 12:28 | <MikeSmith> | yeah |
| 12:29 | <MikeSmith> | lacking one right solution is a great way to generate lots of wrong solutions |
| 12:30 | <annevk> | i guess it is inevitable that people will build browsers on top of browsers |
| 12:33 | <MikeSmith> | that's probably what will be the final trigger that brings us into the Singularity |
| 12:37 | MikeSmith | discovers that there is a thrash-metal band called "Detente" and wonders if they actually know or care what they word means and considers what worse names they could have chosen |
| 12:38 | <MikeSmith> | maybe "Gentle Breeze" or "Kumbaya" |
| 12:38 | <MikeSmith> | not that "detente" is dumb word absolutely |
| 12:39 | <MikeSmith> | it's actually a really good word, with some relevance to all kinds of group/social interactions |
| 12:40 | <MikeSmith> | but thrash metal is not high on the list of those |
| 12:40 | <annevk> | ARIA could really use some ARIA5 treatment |
| 12:41 | annevk | is reviewing the responses to his comments |
| 12:42 | <Dashiva> | I dream of a future where groups accept external feedback so it isn't necessary to "fix" their work afterwards |
| 12:42 | <othermaciej> | I wonder when they'll get to my comments |
| 12:45 | <MikeSmith> | Philip`: I did send mail about the stray remaining usemap-IDREF comment - http://lists.w3.org/Archives/Public/www-html-editor/2010JanMar/0003.html |
| 12:45 | <MikeSmith> | I'm happy to collect and pass on any further comments |
| 12:45 | <MikeSmith> | well, not "happy", exactly |
| 12:45 | <MikeSmith> | but will do it nonetheless |
| 12:47 | <MikeSmith> | and/or to follow up on any outstanding comments |
| 12:48 | <Dashiva> | Oh wow, cssquirrel linked Hixie's bible |
| 12:48 | <Dashiva> | I thought he was smarter than that |
| 12:49 | <MikeSmith> | "outstanding" in this case meaning "so far, ignored or otherwise not adequately responded to" rather than as in, "man that was some outstanding kush" |
| 12:50 | <MikeSmith> | Dashiva: that is clearly a trick to fool people into thinking that he has no sense of irony |
| 12:51 | <MikeSmith> | and then if somebody points that out to him, he can create an ironic T-shirt mocking somebody's lack of ability to recognize his clever feigning of lack of a sense of irony |
| 12:51 | <Dashiva> | But someone did point it out |
| 12:52 | <MikeSmith> | oh, there you go |
| 12:52 | <MikeSmith> | I anticipate a T-shirt will be forthcoming real soon |
| 12:52 | <annevk> | othermaciej, they should've already |
| 12:52 | <annevk> | othermaciej, if they didn't something went wrong |
| 12:53 | <othermaciej> | annevk: mine might have been after whenever their last call was |
| 12:53 | <MikeSmith> | othermaciej: what annevk said |
| 12:53 | <hsivonen> | hmm. Did CSSquirrel remove the comments written by someone claiming to be Larry Masinter? |
| 12:53 | <othermaciej> | http://lists.w3.org/Archives/Public/public-pfwg-comments/2010JanMar/0038.html |
| 12:53 | <MikeSmith> | oh |
| 12:54 | <othermaciej> | I do not see a response |
| 12:54 | <annevk> | Dashiva, he's mostly doing cheap shots |
| 12:54 | <annevk> | Dashiva, though they're funny sometimes, admittedly |
| 12:54 | <Dashiva> | It also seems nobody has called Larry out for lying about the telcon |
| 12:54 | <MikeSmith> | othermaciej: I thought you guys were talking about XHTML 1.1 comments |
| 12:54 | <MikeSmith> | nm |
| 12:54 | <Dashiva> | He says there never was a formal objection, but on the telcon he was explicitly asked about a FO twice |
| 12:54 | <annevk> | othermaciej, apparently they're going into LC again |
| 12:55 | <annevk> | othermaciej, so I guess we can email reminders then |
| 12:55 | <MikeSmith> | I propose we try to avoid using words like "lying" |
| 12:55 | <Dashiva> | Should we use words like "intentionally avoiding telling the truth" instead? |
| 12:55 | <MikeSmith> | and instead use more words like "gentle breeze", kubaya, and detente |
| 12:56 | <hsivonen> | annevk: I thought the first HTML5 CSSquirrel was funny but since then, the comics haven't really been funny |
| 12:56 | <othermaciej> | I have not seen a message by Larry that claimed to be a Formal Objection |
| 12:56 | <othermaciej> | at least, regarding this topic |
| 12:56 | <Dashiva> | Not in the message itself |
| 12:56 | <hsivonen> | othermaciej: the minutes say "<masinter> do I need to repeat objections?" |
| 12:56 | <Dashiva> | And then paulc: the co-chairs are aware of the formal objection |
| 12:57 | <hsivonen> | othermaciej: which to me means no scribe misheard Larry but he wrote the word "objections" himself |
| 12:57 | <Dashiva> | He doesn't bother to correct that |
| 12:57 | <annevk> | http://www.cssquirrel.com/comic/?comic=06 is okay, though not about HTML5 |
| 12:57 | <othermaciej> | other people did call it a Formal Objection, and he did not correct the record right away |
| 12:57 | <hsivonen> | othermaciej: granted, without the word "formal" |
| 12:57 | <Dashiva> | "paulc: plh and larry, can you post the FO on the public-html list and the affects on the plans?" |
| 12:57 | <annevk> | anyway, I sort of stopped following it because his blog is just a few one-liners pulled together without even trying to consider both sides |
| 12:57 | <Dashiva> | If he didn't think it was a FO, how could he agree to post a FO? |
| 13:03 | <hsivonen> | hmm. reading http://www.cssquirrel.com/comic/?comic=13 , it's not as funny as I thought it was |
| 13:04 | <othermaciej> | Larry did call his comment an objection initially, and he didn't correct others who called it a Formal Objection |
| 13:04 | <othermaciej> | later he said it was not formal |
| 13:04 | <gsnedders> | Who is the other person in the final panel? |
| 13:04 | <othermaciej> | now I think he doesn't even want to call it an objection |
| 13:04 | <Dashiva> | Has he tried to justify posting it on a private list yet? |
| 13:05 | <othermaciej> | I believe he implied that my suggestion to avoid lengthy scope discussions on public-html led to his choices of where to post |
| 13:05 | <othermaciej> | honestly, I no longer care what Larry did or why |
| 13:06 | <hsivonen> | othermaciej: when I read your suggestion, I thought you meant www-archive |
| 13:06 | <hsivonen> | othermaciej: maybe that's worth saying next time |
| 13:06 | <othermaciej> | I am just waiting for the w3c team to actually get us an answer from Tim, since they insist on treating Adobe's non-Formal non-Objection as a matter that must be handled by the Director |
| 13:07 | <othermaciej> | hsivonen: he did initially post to www-archive |
| 13:07 | <othermaciej> | he also asked for a ruling from the Chairs and the Team, as I suggested he should if he didn't want to leave it up to the WG |
| 13:07 | <othermaciej> | I figured after that his logical next step, based on my suggestion, would be to prepare a Formal Objection against the WG Decision to publish, should such a decision be forthcoming |
| 13:08 | <Dashiva> | So if you don't agree with something, you can skip all the existing processes and just say "I want a direct ruling here" |
| 13:09 | <othermaciej> | I am not certain of the sincerity of Larry's claim (a week after the fact) that he was just trying to follow my directions |
| 13:09 | <hsivonen> | I thought there was already a publicized ruling on canvas |
| 13:09 | <othermaciej> | my understanding is that you can bypass the Formal Objection process and instead make an Appeal of a Chair's Decision, but only if you feel you are not getting due process |
| 13:10 | <othermaciej> | I thought there was already a publicized ruling on canvas too, but Larry apparently was not convinced that it applied to a separate draft of only the canvas API |
| 13:10 | <othermaciej> | note also that the Director has to approve FPWD so there is also effectively a published ruling on HTML+RDFa |
| 13:10 | <MikeSmith> | Dashiva: purely hypothetically speaking, yes |
| 13:10 | <MikeSmith> | http://www.w3.org/2005/10/Process-20051014/policies.html#WGAppeals |
| 13:11 | <Dashiva> | MikeSmith: It seems quite non-hypothetical since it happens right now |
| 13:11 | <MikeSmith> | yeah, well, I'm referring to the hypothetical case |
| 13:11 | <MikeSmith> | "The Team Contact must inform the Director when a group participant has raised concerns about due process." |
| 13:11 | <othermaciej> | MikeSmith: is that really supposed to apply even when you don't have "concerns about due process"? |
| 13:11 | <MikeSmith> | would be what would happen in that case |
| 13:11 | <hsivonen> | MikeSmith: so an "appeal" doesn't need to be public? only known to the Director and Team Contact? |
| 13:12 | <Dashiva> | Besides, how can you claim you're not getting due process if you haven't even tried following the process? |
| 13:12 | <MikeSmith> | hsivonen: that particular sentence does not mention "appeal" |
| 13:13 | <othermaciej> | I didn't see a message where Larry raised a concern about due process, he just disagreed with a decision, so I would have expected the correct escalation path to be a Formal Objection |
| 13:13 | <MikeSmith> | othermaciej: in any given case, I would think it will always be arguable about what constitutes legitimate "concerns about due process" |
| 13:13 | <othermaciej> | it is unclear to me at this point when one should raise a Formal Objection and when one should raise an Appeal of a Chair's Decision |
| 13:13 | jgraham | finds recent goings on in the HTMLWG distressingly close to the experiences of friends who teach teenagers for a living |
| 13:14 | <hsivonen> | MikeSmith: the section about appeals doesn't mention "appeal" using that word at all (except when mentioning that AC appeals are yet separate) |
| 13:15 | <othermaciej> | MikeSmith: sure - I would just expect the person to at least nominally say "I have concerns about due process because X" |
| 13:15 | <MikeSmith> | yeah, the title of that section is not ideal |
| 13:15 | <MikeSmith> | hsivonen: ↑ |
| 13:15 | <MikeSmith> | othermaciej: yeah, ideally |
| 13:15 | <othermaciej> | as opposed to "I disagree with decision Y for reason Z" |
| 13:15 | <othermaciej> | disagreeing with a decision is not in itself a concern about due process |
| 13:16 | <hsivonen> | jgraham: in what sense? |
| 13:16 | <othermaciej> | it would be nice if the Team would clarify when it is appropriate to use a Formal Objection vs. an Appeal of a Chair's Decision, whether the latter category in fact exists, and whether there are any other kinds of appeals that can apply to decisions in a Working Group |
| 13:16 | <othermaciej> | because I thought I understood it from reading the Process document but clearly I don't |
| 13:17 | MikeSmith | taught in a high school for 3 years and can see some relevance in jgraham's allusion |
| 13:17 | <MikeSmith> | othermaciej: I suppose the Director can clarify that |
| 13:17 | <othermaciej> | the reason I ask is because the appeal to a Team Contact bypasses the chairs and is guarateed to get to the Director right away instead of by the next transition request, so if an Appeal is allowed any time you disagree, why would anyone choose to raise a Formal Objection instead? |
| 13:18 | <gsnedders> | Bunch of teenagers… |
| 13:18 | <Dashiva> | Maybe the document was written without considering people gaming the system |
| 13:18 | <othermaciej> | I assumed the point of the process where you bypass the chairs was for the case where you think the chairs are not acting in good faith and thus failing to give you due process |
| 13:18 | <MikeSmith> | othermaciej: true that would not be a path that it would be prudent to encourage going forward |
| 13:20 | <Dashiva> | "In order to vote to resolve a substantive issue, an individual MUST be a group participant in Good Standing." |
| 13:20 | <MikeSmith> | I think circumstances now and then end up such that the Director gets involved, even if the letter of the process doc would not seem to absolutely require it at that point |
| 13:20 | <Dashiva> | Doesn't that make htmlwg's exception for IEs invalid? |
| 13:21 | <MikeSmith> | othermaciej: a possibly good outcome of this particular case might be that it gives some greater clarification about what the Director expects should happen the next time a case like this might come up |
| 13:23 | <othermaciej> | MikeSmith: that would be good, yes |
| 13:26 | MikeSmith | tries to concentrate on sentiments like "look for the silver lining" and communal songs like kumbaya, in contrast with, say, "mutual assured destruction" |
| 13:44 | <gsnedders> | Anyone know how to get the amount of memory used by a process in Java? |
| 13:45 | <MikeSmith> | gsnedders: top(1) ? |
| 13:46 | <gsnedders> | MikeSmith: And then parse it? That seems hidious |
| 13:46 | <gsnedders> | And so system dependent |
| 13:47 | <MikeSmith> | well, java VMs are system dependent |
| 13:47 | <gsnedders> | Yeah, but there's not even anything that will work on all POSIX systems here |
| 13:47 | <MikeSmith> | they only appear system independent to the applications they run |
| 13:48 | <Dashiva> | stackoverflow says http://support.hyperic.com/display/SIGAR/Home |
| 14:00 | <Dashiva> | krijnh: How about "next day" and "previous day" links at the bottom of the logs? |
| 14:00 | <annevk> | othermaciej, do you know if Safari made the withCredentials thing affect same-origin requests? |
| 14:01 | <annevk> | per spec that's not supposed to work, but maybe we should allow it... |
| 14:03 | <hsivonen> | does anyone have a test case that shows if Flash gets the attributes of <embed> in the "right" order? |
| 14:03 | <Dashiva> | krijnh: Also, I'm getting only parts of the page most times. The content just stops. |
| 14:06 | <annevk> | hsivonen, didn't hixie have tests for that? |
| 14:06 | <hsivonen> | annevk: thanks. I'll look for it on Hixie's site |
| 14:09 | hsivonen | fails to find a test under http://www.hixie.ch/tests/adhoc/ |
| 14:10 | <annevk> | cannot see it either now, but hixie did test attribute order for plugins somehow |
| 15:02 | <asmodai> | I love what people come up with: |
| 15:03 | <asmodai> | http://www.bramstein.com/projects/typeset/ |
| 15:06 | <hsivonen> | I rather wish browser didn't compete on perf so much on desktop and dared to implement more expensive line breaking algorithms |
| 15:07 | <hsivonen> | or at least did it for justified text |
| 15:07 | <hsivonen> | justified text probably doesn't show up of perf benchmarks too much |
| 15:09 | <asmodai> | hsivonen: I'm sure we can push some people's buttons to add that ;) |
| 15:09 | <TabAtkins> | That's pretty cool. We need more effort on good typography. |
| 15:10 | <asmodai> | TabAtkins: Agreed. |
| 15:10 | <hsivonen> | (that is, I appreciate competition on perf as long as rendering beauty isn't stifled by perf concern) |
| 15:10 | <TabAtkins> | I mean, how often do perf tests measure justified text anyway? I would *love* to see this justification algorithm in. I implement justified text generally across my company's site, but it has less-than-pretty effects in some places. |
| 15:11 | <gsnedders> | hsivonen: Which in the case of justified text, it normally is |
| 15:11 | <TabAtkins> | Also, the quality of that shape-filling justification is just wonderful. |
| 15:11 | hsivonen | would love to see the (La?)TeX hyphenation algorithm in browsers, too |
| 15:12 | <TabAtkins> | I think "We make the web more beautiful" could be a successful advertising point to counter perf losses. |
| 15:13 | <hsivonen> | the #1 beauty problem is Windows font rasterization though |
| 15:13 | <TabAtkins> | This is true. |
| 15:13 | <hsivonen> | and the Ubuntu/Fedora FreeType defaults, too |
| 15:14 | <asmodai> | hsivonen: How's that? I don't know that background. |
| 15:15 | <hsivonen> | asmodai: Windows XP doesn't have even ClearType on by default |
| 15:15 | <hsivonen> | asmodai: and ClearType is very ugly compared to Quartz or FreeType |
| 15:15 | <hsivonen> | worse, the Windows rasterizers use legacy TrueType hinting |
| 15:16 | <hsivonen> | designed for low-res non-AA rasterization |
| 15:16 | <hsivonen> | which has the side effect that fonts that aren't hinted with Windows in mind can look nice on Mac and Linux but totally, totally horrible on Windows |
| 15:17 | <asmodai> | Mmm, wonder if that's still the case on 7 |
| 15:17 | <hsivonen> | it was on 7 beta |
| 15:18 | <hsivonen> | I doubt they've changed the rasterizer since then |
| 15:20 | <asmodai> | Indeed doubtful. |
| 15:22 | <Philip`> | gsnedders: /proc/$PID/status gives easier-to-parse memory stats, on Linux |
| 15:22 | <Philip`> | (Easier than top) |
| 15:23 | <Philip`> | Doesn't seem to make much sense to have a system-independent memory usage thing, because the concept of "memory usage" is itself system-dependent |
| 15:24 | <MikeSmith> | hsivonen: very short page for you to review if you have a couple minutes |
| 15:24 | <MikeSmith> | http://pastebin.ca/1802736 |
| 15:24 | <MikeSmith> | this is pretty much the same as the earlier page-template patch |
| 15:24 | <MikeSmith> | it just does the same for the form part of the page |
| 15:25 | <hsivonen> | MikeSmith: assuming the page emitter works, I trust this does the same thing for another file |
| 15:26 | <MikeSmith> | OK |
| 15:28 | <MikeSmith> | hsivonen: btw, do you think it would make sense to change the build so that it does actually generate the Java code from the (default/existing) page templates? |
| 15:29 | <MikeSmith> | that is, have the SaxCompiler actually generate the validator/src/nu/validator/servlet/PageEmitter.java file as part of the build |
| 15:30 | <MikeSmith> | from the default validator/xml-src/PageEmitter.xml file |
| 15:30 | <hsivonen> | MikeSmith: that would be ok as long as the generated file goes into the current location for convenient access from Eclipse |
| 15:31 | <MikeSmith> | OK, yeah, it would definitely go in the same location |
| 15:31 | <MikeSmith> | I will write up a patch for it later |
| 17:11 | <JonathanNeal> | Well, guess what, guys. |
| 17:11 | <TabAtkins> | What? |
| 17:12 | <JonathanNeal> | The website works so well in HTML5 that we're going to move the entire product to HTML5. |
| 17:12 | <TabAtkins> | Hah, awesome! |
| 17:12 | <JonathanNeal> | Yea ... kinda nervous about that one, though I pushed this. |
| 17:17 | <AryehGregor> | What HTML5 features did you use here? |
| 17:19 | <JonathanNeal> | On our website? |
| 17:20 | <JonathanNeal> | Mostly structural to build the website as an outline, but we're also using <video> with a graceful flash fallback. |
| 17:22 | <AryehGregor> | And this really gives you serious concrete benefits? I'm a big fan of HTML5, but it's only marginally useful so far as far as I can tell. |
| 17:23 | <JonathanNeal> | Well, it's the way the web is going, and nobody else is seriously going for it, so we figure it worked so well for the website |
| 17:23 | <AryehGregor> | Like, I wouldn't use <video> with Flash fallback except as opt-in, since it doesn't support things like fullscreen (except in FF3.6) . . . I'd make it opt-in and/or use Flash with <video> fallback. |
| 17:23 | <JonathanNeal> | bam, let's do it. |
| 17:23 | <AryehGregor> | I've been trying to get Wikipedia to switch, but it's backlogged way too much on code review, so it hasn't deployed my new less-broken HTML5-enabling code yet. :( |
| 17:23 | <Dashiva> | Well, marginally is the name of the game |
| 17:23 | <JonathanNeal> | HTML5 elements in use in the core product would be, hgroup, nav, section, article |
| 17:24 | <AryehGregor> | Dashiva, true. |
| 17:24 | <Dashiva> | You don't have an atomic HTML5 vs not |
| 17:24 | <JonathanNeal> | probably header and footer as well, though I'm still not sure how the W3C or WHATWG wants us to use them. |
| 17:24 | <Dashiva> | Just a large set of tiny changes that mesh together |
| 17:24 | <AryehGregor> | JonathanNeal, so you have the JS hack for IE? |
| 17:24 | <JonathanNeal> | Yes. |
| 17:24 | <JonathanNeal> | And we haven't received a single complaint about it. |
| 17:25 | <JonathanNeal> | And traffic is very high on our site from the enterprise end, who would be the ones on the lowest end machines. |
| 17:25 | <JonathanNeal> | No complaints about not being able to see the videos either, in fact, only a few compliments from developers who noticed the <video> implementation. |
| 17:25 | <JonathanNeal> | For those who don't know what we did, they didn't even notice. |
| 17:27 | <JonathanNeal> | Oi, just feel like I'm playing with the fetus of fire. |
| 17:28 | <AryehGregor> | This reminds me, I have to re-add IE5.0 and 5.5 CSS support to MediaWiki. Someone gratuitously removed it. |
| 17:29 | <Dashiva> | Do you really have to? |
| 17:29 | <JonathanNeal> | AryehGregor, what were the benefits of adding it back in? |
| 17:29 | <JonathanNeal> | And what were the benefits of taking it out? |
| 17:29 | <AryehGregor> | The only benefit of taking it out is two lines saved in the <head>. |
| 17:30 | <AryehGregor> | The benefit of adding it back in is tens of thousands of IE5 users will still be able to read Wikipedia. |
| 17:30 | <AryehGregor> | Probably hundreds of thousands. |
| 17:30 | <AryehGregor> | Hmm, well, at least if you assume they view at about the same rate as other users. |
| 17:31 | <Dashiva> | I'm guessing you don't pay heed to the moral argument of not aiding and abetting entrenchment of outdated browsers :) |
| 17:32 | <AryehGregor> | No, Wikimedia's goal is to make knowledge freely available to everyone. Not only to people who use up-to-date browsers. |
| 17:32 | <AryehGregor> | *Personally*, well, you can look at the source code of http://aryeh.name/ to see my attitude there. :P |
| 17:32 | <JonathanNeal> | Tens of thousands of IE5 users? |
| 17:32 | <JonathanNeal> | Are you sure you're not talking about IE6? |
| 17:33 | <daedb_> | IE6 still has many millions of users. |
| 17:34 | <AryehGregor> | I'm talking about IE5, on Wikipedia. |
| 17:34 | <AryehGregor> | Wikipedia has hundreds of millions of users. |
| 17:34 | <AryehGregor> | IE5 is not down to 0.01% market share yet. |
| 17:34 | <JonathanNeal> | Sure, and I honestly didn't know tens of thousands of them (I understand it's a small percentage) were still using IE5. |
| 17:34 | <AryehGregor> | You think IE5 is all the way down to 0.01%? |
| 17:35 | <JonathanNeal> | I thought it wasn't even listed anymore. |
| 17:35 | <AryehGregor> | The statistic that was given as justification for the removal was 0.3%, which would be hundreds of thousands of users. |
| 17:36 | <JonathanNeal> | Yea, I can understand that. |
| 17:36 | <JonathanNeal> | You are correct, not that you needed me to say that. |
| 17:36 | <AryehGregor> | http://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm |
| 17:37 | <AryehGregor> | MSIE 5.5 4,982 0.11% |
| 17:37 | <AryehGregor> | MSIE 5.0 932 0.02% |
| 17:37 | <AryehGregor> | MSIE 5.01 902 0.02% |
| 17:37 | <AryehGregor> | Actually, you were closer to correct. |
| 17:37 | <AryehGregor> | Still thousands of hits, but possibly only hundreds of users. |
| 17:38 | <AryehGregor> | Oh, wait, that's times a thousand. |
| 17:38 | <AryehGregor> | So about 5,000,000 hits from IE5.5. |
| 17:38 | <JonathanNeal> | Did the removal of the two lines kill IE5.5 though? |
| 17:38 | <JonathanNeal> | I thought IE5.5 was its own beast. |
| 17:39 | <AryehGregor> | One line was for IE5, one was for IE5.5. |
| 17:39 | <hsivonen> | I wonder how many of those are bots pretending to be ie |
| 17:39 | <Dashiva> | If you could cross-reference requests to pages with requests to style sheets... |
| 17:39 | <karlcow> | http://lists.xml.org/archives/xml-dev/201002/msg00064.html |
| 17:40 | <hsivonen> | a couple of years ago I was looking at my logs and concluded most ie5 requests weren't from browsers |
| 17:40 | <AryehGregor> | Dashiva, that's an idea, we could look at the stats for how often those particular stylesheets were requested. |
| 17:40 | <AryehGregor> | Bots wouldn't request stylesheets either. |
| 17:40 | <Dashiva> | GranParadiso was the FF 3.0 alpha, wasn't it? |
| 17:41 | <karlcow> | AryehGregor: it's not done every month? |
| 17:41 | <AryehGregor> | karlcow, I dunno, I don't see newer data. |
| 17:42 | <hsivonen> | Dashiva: Ubuntu backports use the code names for browsers that are built from the same codebase as firefox |
| 17:42 | <TabAtkins> | karlcow: Is that list open to post to? I can provide reasoning for why, frex, data urls were rejected as a solution. |
| 17:43 | <Dashiva> | hsivonen: Isn't that IceWeasel etc? |
| 17:43 | <karlcow> | TabAtkins: I guess the list xml-dev has a moderator |
| 17:43 | <hsivonen> | Dashiva: firefox-3.5 in jaunty-backport gives you a browser called Namaroka |
| 17:44 | <hsivonen> | Dashiva: that's debian |
| 17:44 | <Dashiva> | Namaroka is supposed to be 3.6 |
| 17:44 | <hsivonen> | oops sorry. I meant Shiretoko |
| 17:45 | <Dashiva> | But I get your point |
| 17:45 | hsivonen | lives on the trunk |
| 17:45 | karlcow | wonders how many people are using Safari with VoiceOver, or jaws with IE. not visible in the stats, I guess |
| 17:56 | <JonathanNeal> | I noticed that the WHATWG example for HTML5 shows <nav><h2>Navigation</h2> ... </nav> without an <h1> |
| 17:56 | <JonathanNeal> | Why is that? |
| 17:57 | <Philip`> | The absolute values of the numbers are meaningless |
| 17:57 | <Philip`> | (Only relative values matter) |
| 17:58 | <Philip`> | So I guess the example is attempting to demonstrate that it's valid to do that |
| 17:58 | <TabAtkins> | Also, using lower numbers sometimes can be helpful for no-CSS display in legacy browsers. |
| 17:58 | <JonathanNeal> | So <h1> or <h2> would have been fine, it's completely arbitrary unless relative to another <h*> |
| 18:00 | <JonathanNeal> | I've read somewhere that <hgroup> is a specialised form of <header>, but that would not appear true. |
| 18:00 | <Dashiva> | <hgroup> is for grouping headers and subheaders |
| 18:00 | <Dashiva> | So they aren't interpreted as separate headings |
| 18:00 | <TabAtkins> | Nah, it's just a way to group multiple <hn> into a single heading, so it won't mess up the outline. |
| 18:01 | <JonathanNeal> | Right, that's what I thought. |
| 18:01 | <JonathanNeal> | That's how we're using them. |
| 18:10 | <JonathanNeal> | Okay guys, let me share my "strategy" for HTML5 in our product. |
| 18:10 | <JonathanNeal> | I can pastie it, actually. |
| 18:11 | <JonathanNeal> | http://pastie.org/833062 |
| 18:13 | <AryehGregor> | Why bother with <html>, <head>, <body>? |
| 18:13 | <AryehGregor> | They're implicit. |
| 18:14 | <AryehGregor> | You don't need the /> in the meta, <meta charset="UTF-8"> will work fine. |
| 18:14 | <AryehGregor> | Also, why <header id="header">? What's the id needed for? |
| 18:14 | <AryehGregor> | Likewise <nav id="navigation">, etc. |
| 18:14 | <AryehGregor> | Part of the point of those elements is so you don't need those id's. |
| 18:15 | <AryehGregor> | You also have some div wrappers of unclear purpose -- do you need those for styling? |
| 18:15 | <JonathanNeal> | AryehGregor, yes the div wrappers are there for style guys to muck around with the content. |
| 18:16 | <JonathanNeal> | the IDs are important because they target a particular singular instance versus a generic element which can exist elsewhere in the site. |
| 18:17 | <Philip`> | Calling them something like id="main-header" might help disambiguate them, and stop people asking why that header is header and other headers aren't header |
| 18:17 | <AryehGregor> | Or you could put a class or id on the body element and use that. |
| 18:17 | <AryehGregor> | So you'd only need one class instead of lots of id's. |
| 18:18 | <AryehGregor> | Or, better yet, just stick with no id's or classes at all until you actually need them. That's what I've done so far for my personal site, although admittedly it only has a few pages. |
| 18:18 | <JonathanNeal> | Well, I don't see how that would be easier AryehGregor, I'd still end up having .someBodyClassName > div > header or something to specifically target those elements, that's what IDs and classnames are for. |
| 18:18 | <TabAtkins> | I've used "body > header" to target my top-level stuff. |
| 18:18 | <AryehGregor> | Me too. |
| 18:18 | <AryehGregor> | That's what selectors are for. |
| 18:19 | <AryehGregor> | Especially if you don't need IE support, then you barely need classes/ids at all. |
| 18:19 | <JonathanNeal> | We'll need IE support. |
| 18:19 | <JonathanNeal> | Also, it's easier for a web designer to target an ID than moving through selectors. |
| 18:19 | <TabAtkins> | Well, IE6 is the problem there. Doesen't support child selectors. |
| 18:20 | <TabAtkins> | So an #id helps there. |
| 18:20 | <AryehGregor> | Yeah. Sigh. |
| 18:20 | <JonathanNeal> | AryehGregor, not to mention that it's easier to develop on. |
| 18:21 | <JonathanNeal> | id is implicit that it is unique to the page, the only instance, therefore #header and #footer mean "The ONE header" and "The ONE footer" |
| 18:41 | <JonathanNeal> | Okay updated @ http://pastie.org/833113 |
| 19:10 | <MikeSmith> | the spec really ought to include an explicit description of the rationale/background behind the srcdoc attribute |
| 19:19 | <Philip`> | Even though it doesn't describe rationale/background for any other feature? |
| 19:44 | <karlcow> | http://www.bramstein.com/projects/typeset/ |
| 19:44 | <karlcow> | >This is an implementation of the Knuth and Plass line breaking algorithm using JavaScript and the HTML5 canvas element. The goal of this implementation is to optimally set justified text in the new HTML5 canvas element, and ultimately provide a library for various line breaking algorithms in JavaScript. |
| 19:49 | <AryehGregor> | Oh, neat, Opera is releasing something as open-source. |
| 19:56 | <Philip`> | AryehGregor: What something? |
| 19:56 | Philip` | doesn't follow enough blogs and/or news sites |
| 19:56 | <AryehGregor> | http://my.opera.com/dragonfly/blog/opera-dragonfly-open-for-business |
| 19:57 | AryehGregor | only follows Slashdot, which is enough to get most of the notable tech news |
| 19:57 | <Philip`> | Ooh, Mercurial |
| 19:58 | <karlcow> | DragonFly has always been under BSD licence. |
| 19:58 | karlcow | goes read |
| 19:58 | Philip` | no longer follows any tech news sites at all, because it takes up too much time and anything really important will be repeated on IRC |
| 19:59 | <karlcow> | so they put the code in a repository now. I guess that is the diff with before. |
| 21:04 | <tumbleweed> | howdy. I can't seem to create ElementTrees with html5lib 0.90: http://pastebin.com/m298aa482 |
| 21:19 | <bfrantz> | is it outside the scope of html5 to establish how user agents should treat a content body when issued a 404? |
| 21:20 | <bfrantz> | seems like more of an http issue but thought i'd check |
| 21:22 | <GPHemsley> | Hixie: How come this doesn't say that the default text alignment for <th> is center? http://dev.w3.org/html5/markup/th.html#th |
| 21:22 | <AryehGregor> | Sounds like HTTP. |
| 21:25 | <zcorpan> | GPHemsley: MikeSmithX is editor for that document |
| 21:25 | <Hixie> | GPHemsley: dunno, i didn't write that document |
| 21:49 | <zcorpan> | http://www.zazzle.com/cssquirrel |
| 21:50 | <Dashiva> | The funny part is that it's probably grounds for a libel suit in England |
| 21:53 | <Philip`> | Not if it's true |
| 21:53 | <AryehGregor> | Wait, you need grounds for a libel suit in England these days? |
| 21:54 | <AryehGregor> | That's an awesome T-shirt. |
| 21:54 | <Dashiva> | Philip`: I don't think truth is relevant |
| 21:54 | <Dashiva> | As long as feelings are hurt |
| 21:55 | <Philip`> | http://en.wikipedia.org/wiki/English_defamation_law#Justification - "A claim of defamation is defeated if the defendant proves that the statement was true." |
| 21:57 | <KevinMarks> | right, but the burden of proof is reversed |
| 21:58 | KevinMarks | waits for libel prosecutors to discover Gödel's theorem |
| 21:59 | <Dashiva> | That's libelous to say, KevinMarks |
| 22:19 | <GarethAdams|Home> | I'd love a libel suit |
| 22:19 | <GarethAdams|Home> | with a libel waistcoat and libel trousers |
| 22:24 | <JonathanNeal> | haha |
| 22:26 | <abarth_> | othermaciej: according to http://dev.w3.org/html5/status/issue-status.html it looks like the deadline for ISSUE-4 change proposals as past |
| 22:27 | <deltab> | bfrantz: not if the missing resource is a stylesheet, script, image, video... |
| 22:43 | <TabAtkins> | Philip`: Just watched Hot Fuzz on your recommendation (a while back). Good movie! |
| 22:46 | <bfrantz> | deltab: i have not explicitly tested for stylesheets, script, etc. i stumbled over an apache configuration that implements the ErrorDocument directive to send 404s to index.xml which in turn issues a meta redirect to a script. FF3.5, Safari 4, and Opera 10 follow the redirect where IE8 and Chrome 4 ignore it. I figured I would raise this behavior here so folks are aware of it. It may or may not be a behavior worth standardizing on. |
| 22:47 | <bfrantz> | all browers follow the redirect when visiting index.html directly |
| 22:48 | <deltab> | meta refresh, you mean? |
| 22:48 | <bfrantz> | yea, my bad |
| 22:53 | <GarethAdams|Home> | following meta refreshes is often a preference in a browser |
| 22:54 | <GarethAdams|Home> | as an author, using a meta refresh is a bit amateur |
| 22:58 | <bfrantz> | GarethAdams|Home: I agree, but thats not the reason i'm mentioning this. The meta refresh is just an example of the UA rendering the entity body when encountering a 404. |
| 22:59 | <GarethAdams|Home> | bfrantz: my mistake, I'm in a lot of webdev support channels, I sometimes forget I'm also in this one when flicking through |
| 23:11 | <Hixie> | othermaciej: for the websocket nonce thing... i'm trying to come up with something equally hard to predict or fake, but which can be done trivially (e.g. without implementing MD5 or using a library). What do you think of the client sending 16 bytes of noise, and the server return 8 bytes of noise consisting of the first 8 bytes from the client xored with the second 8 bytes? |
| 23:12 | <GPHemsley> | MikeSmithX: How come this doesn't say that the default text alignment for <th> is center? http://dev.w3.org/html5/markup/th.html#th |
| 23:13 | <Philip`> | TabAtkins: Hmm, did I actually recommend it? |
| 23:13 | <TabAtkins> | Pretty sure. |
| 23:13 | <Philip`> | I thought I usually just mention things without actually giving an opinion because my opinion might be wrong and that would be embarrassing |
| 23:16 | <Hixie> | clearly you are in fact wrong about not giving an opinion |
| 23:17 | <TabAtkins> | Shouldn't have given that opinion, I guess. |
| 23:22 | <zcorpan> | Philip`: the 2008 doctype data seems a bit dated; i found that the 116 doctypes were all gone by now, and looking further at the others doesn't give much fruit either |
| 23:25 | <zcorpan> | Philip`: could you maybe do a new query for: (?i)<!doctype\s*html\s+public\s*("[^"]*"|'[^']*')\s*[^"'>] |
| 23:43 | <othermaciej> | Hixie: that doesn't convey the origin at all |
| 23:44 | <othermaciej> | Hixie: or include anything that specifically indicates WebSocket and not some other protocol |
| 23:46 | <othermaciej> | Hixie: the client could use a one-time pad to encrypt the request handshake, but you really want the server to decrypt and then repack what it got |
| 23:46 | <othermaciej> | Hixie: if you're basing this on abarth's concept |
| 23:58 | <Philip`> | zcorpan: Do you want just a list of URLs, or also the matching excerpts? |
| 23:59 | <Philip`> | (The latter is harder because my multi-line grep program only outputs URLs) |
| 23:59 | <zcorpan> | urls is fine |