| 01:31 | <j0hng4lt> | I'm wondering what will happen to XHTML5 in strict mode. My previous attempts to use XHTML1 resulted in two problems. |
| 01:32 | <j0hng4lt> | Problem #1: Due to a rather extreme interpretation of Postel's law in early browser designs, most developers don't even know what constitutes valid markup and how to escape content. Juniors on the team were clued out when it came to ampersands |
| 01:33 | <j0hng4lt> | Problem #2: External libraries and other "foreign" code (e.g. https seals, ads, etc) always assume you're in text/html and use document.write() |
| 01:33 | <j0hng4lt> | this blows up in Firefox |
| 01:33 | <TabAtkins> | Yes? |
| 01:33 | <j0hng4lt> | (when served as application/xhtml+xml) |
| 01:34 | <webben> | j0hng4lt: I don't see how the new spec can make a dent in either problem. |
| 01:34 | <j0hng4lt> | me neither... |
| 01:34 | <TabAtkins> | XML is always going to be strict, and suffer from the problems you've mentioned. That's why XML isn't very useful for the web. |
| 01:34 | <j0hng4lt> | What I'm afraid of is that everyone will start "switching back to tag soup". |
| 01:34 | <webben> | j0hng4lt: But does it matter? Do you need XML for anything? |
| 01:35 | <j0hng4lt> | The problem is not with XML... No one suffers from, say, a C# compiler not accepting invalid C#. Even VB programmers know how to close their while loops |
| 01:35 | <j0hng4lt> | It's purely historical |
| 01:36 | <webben> | j0hng4lt: Ah but there's a big difference between closely controlled syntaxes (e.g. python) and loose syntaxes (e.g. perl, php) |
| 01:36 | <webben> | j0hng4lt: defined tag soup parsing is pretty much a loose syntax for the web. |
| 01:36 | <j0hng4lt> | I'm just afraid that everyone will start switching back to tag soup and start generating even more libraries that rely on Postel's Law Version 2 (be liberal in what you produce and in what you accept) |
| 01:37 | <webben> | j0hng4lt: And the "solution" for development teams seeking to regulate that syntax is the same as for such languages - use a linter. |
| 01:37 | <webben> | j0hng4lt: So long as the output is defined - does it matter? (Anymore than the looseness of Perl and PHP matter?) |
| 01:38 | <TabAtkins> | j0hng4lt: C# compilers can afford to reject invalid C#, because the programmer is guaranteed to see the failure before the user does. On the web, the opposite is usually the case - users usually suffer from an XML failure before an author sees it. |
| 01:38 | <TabAtkins> | In simple terms, the user is the one holding the compiler on the web, and the author just ships source code. This is obviously completely different from regular applications. |
| 01:39 | <j0hng4lt> | Well... it's a bit of a "chicken-or-the-egg" question... Had the syntax been strict from the start, the authors would see the errors before the users did. The tools would have been different. |
| 01:39 | <j0hng4lt> | But we are where we are... |
| 01:40 | <TabAtkins> | No, that doesn't fix anything at all. People who know what they're doing and are *trying* to produce valid XML fail *all the time*. |
| 01:40 | <j0hng4lt> | and it is impractical / infeasible to convert everything to strict syntax |
| 01:40 | <webben> | j0hng4lt: "the authors would see the errors before the users did". Not necessarily. HTML isn't a programming language - it's often (typically) /generated/ by code. |
| 01:40 | <j0hng4lt> | it's not that hard to produce valid xml if it's done by means other than string concatenation |
| 01:40 | <webben> | So there's no error for the programmer to see. |
| 01:41 | <j0hng4lt> | (templates, linq-to-xml, XmlWriter, etc.) |
| 01:41 | <TabAtkins> | j0hng4lt: Yes, it is. And still, nearly all programmers use string concatenation. |
| 01:41 | <j0hng4lt> | and in environments that escape by default |
| 01:42 | <TabAtkins> | Virtually every framework ever invented uses string concat, and those that do try to do some escaping of the data often still miss things (like forgetting to remove U+0000 from input before shoving it into an element, which'll crash strict XML). |
| 01:43 | <webben> | e.g. the php dom implementation happily outputs characters that can't be legally placed in xml documents. |
| 01:45 | <j0hng4lt> | again, this is a chicken-or-the-egg question... Most people who want to output valid XHTML fail - either because they can easily get away with it or because the tools mess it up for them... |
| 01:45 | <j0hng4lt> | They don't normally fail to output valid XML |
| 01:45 | <j0hng4lt> | (e.g. valid RSS, valid sitemaps, etc.) |
| 01:45 | <j0hng4lt> | non-HTML applications of xml |
| 01:45 | <TabAtkins> | By "most" you mean "almost every single person on the web, given sufficient time or attackers". |
| 01:45 | <TabAtkins> | RSS is quite often invalid. |
| 01:47 | <webben> | locus classicus: http://www.xml.com/lpt/a/1450 |
| 01:50 | <TabAtkins> | Note as well: "tag soup" is bad because everyone interpreted it differently, so you could never be sure what pops out. HTML5, on the other hand, defines with absolute precision how to produce an HTML document from any random bag of bytes you throw at it. As long as everyone follows the HTML5 parsing algorithm, you can know with absolute precision how your document will be interpreted. |
| 01:50 | <j0hng4lt> | In my opinion, loose handling of semantics (e.g. a new element) is fine and dandy... What tends to get frustrating is loose handling of basic syntax... If something does break because of bad nesting (e.g. <form><table><tr><td>...</tr></td></form></table> - and most of the page looks like this) - you can't easily find the error because the validator will show an error on every line... |
| 01:51 | <j0hng4lt> | I don't object to tag soup being handled given the situation |
| 01:51 | <j0hng4lt> | I would have objected to it in the early 90s |
| 01:51 | <j0hng4lt> | but now you have to handle it whether you like it or not |
| 01:51 | <TabAtkins> | It would have failed back then as well, for the same reasons it fails today for people who are actually trying to produce valid XML. |
| 01:52 | <j0hng4lt> | In environments that actually use xml parsers, failing to produce valid xml isn't that common |
| 01:53 | <TabAtkins> | My anecdotes contradict your assertion. |
| 01:53 | <webben> | in fairness, it would be easy to make a "strict" markup syntax that's simpler to produce than xml |
| 01:53 | <j0hng4lt> | And Postel's law was meant for forward and backward compatibility - not for the handling of badly written code |
| 01:53 | <TabAtkins> | In carefully controlled environments that aren't doing anything complicated, sure, valid XML might get produced by accident. |
| 01:56 | <j0hng4lt> | Most applications of XML don't use string concat... They might generate semantic errors (e.g. throwing in an element that's not in the schema), sometimes by design, but you won't see nesting errors, naked ampersands, etc (i.e. well-formedness errors) |
| 01:57 | <TabAtkins> | Actually, a significant number of XML-producing applications do produce wellformedness errors. Usually they get nesting right, but they'll often happily output characters that aren't allowed unescaped in XML. |
| 01:57 | <j0hng4lt> | Standard xml processing libraries in .net, java, and even in php are pretty good at well-formedness |
| 01:57 | <j0hng4lt> | the same goes for other strict formats |
| 01:57 | <j0hng4lt> | (json, .ini files, Apache config files) |
| 01:58 | <TabAtkins> | JSON is strict but also *very simple*, and has a single valid encoding. It's trivial to produce valid JSON, unlike XML. |
| 01:59 | <Hixie_> | i dunno about _that_ |
| 01:59 | <Hixie_> | it's simpler, sure, but not trivial |
| 01:59 | <othermaciej> | JSON has some quirky details |
| 01:59 | <othermaciej> | but indeed, it is much simpler than XML |
| 02:00 | <Hixie_> | JSON also still lacks error handling rules |
| 02:00 | <j0hng4lt> | oh well... I've been seeing a lot of flame wars on the Interwebs along the lines of "xml sucks because it's complicated" or "xml sucks because it's sucks because it's xml"... In my opinion, XML is good for some tasks, bad for others. And generating any language using string concat is error-prone without a liberal parser |
| 02:00 | <TabAtkins> | Hixie_: Right, JSON doesn't define what invalid JSON means. It's "strict", in that invalid stuff just isn't JSON at all. |
| 02:00 | <j0hng4lt> | The main reason for the backlash against XML is its marketing... It was marketed as a panacea for everything |
| 02:01 | <TabAtkins> | But how can you get any simpler than the diagram at JSON.org? |
| 02:01 | <j0hng4lt> | .ini files are simpler than json - but also more limited |
| 02:01 | <TabAtkins> | True. But JSON isn't more limited than XML, so that comparison doesn't seem relevant. |
| 02:02 | <j0hng4lt> | fixed-width records are still simpler than ini files |
| 02:03 | <j0hng4lt> | The popularity of JSON is explained by the fact that it's easy to use in JavaScript. |
| 02:03 | <j0hng4lt> | and very few people construct it manually |
| 02:04 | <j0hng4lt> | most of the time, it is produced by a tool that cannot possibly generate invalid JSON |
| 02:04 | <j0hng4lt> | e.g. |
| 02:04 | <j0hng4lt> | asp.net mvc: return Json(myObject); |
| 02:04 | <j0hng4lt> | php: echo json_encode($array); |
| 02:05 | <j0hng4lt> | Constructing JSON with string concatenation would result in the same road blocks |
| 02:05 | <j0hng4lt> | (e.g. having to escape quotes within strings) |
| 02:06 | <j0hng4lt> | Also, JSON is more limited than xml. It lacks namespaces. |
| 02:07 | <j0hng4lt> | But in the environment where it's normally used (AJAX calls), this limitation is irrelevant |
| 02:07 | <TabAtkins> | No it doesn't. It lacks XML Namespaces. That's not a weakness either. |
| 02:10 | <j0hng4lt> | Without the tooling (XPath, XSLT, namespaces, schemas, etc.), it's a bare-bones format that does the same thing as XML (i.e. it's a text representation of a tree structure). If you reconstruct the tooling, it will have the same strengths and weaknesses as XML except for this: |
| 02:10 | <j0hng4lt> | It will somewhat more succint (less space/bandwidth per payload if it is not compressed) |
| 02:11 | <j0hng4lt> | And it will be slightly less readable / editable |
| 02:11 | <TabAtkins> | I think most people find JSON much easier to read and edit than XML (me included). |
| 02:12 | <j0hng4lt> | in notepad, yes... In an IDE with syntax completion... not really... |
| 02:12 | <j0hng4lt> | But other than that, it reminds me of some of those good old flame wars |
| 02:12 | <j0hng4lt> | vi versus emacs |
| 02:12 | <j0hng4lt> | bash verus csh |
| 02:12 | <j0hng4lt> | *versus |
| 02:12 | <j0hng4lt> | tabs versus spaces |
| 02:13 | <TabAtkins> | The difference being that the web has shown pretty conclusively that XML is not the right solution for webpages, unlike your choice of editor, shell, or indentation character. |
| 02:15 | <j0hng4lt> | I think the reasons for that are purely historical... |
| 02:16 | <TabAtkins> | Most people appear to disagree, and have expressed their reasoning in multiple blog posts and articles across the web. |
| 02:16 | <j0hng4lt> | it tends to go back and forth... |
| 02:17 | <j0hng4lt> | I think the biggest problem with standards in general |
| 02:17 | <j0hng4lt> | (not just strict versus loose syntax) |
| 02:17 | <j0hng4lt> | is that on the web, standardization was a bit of an afterthought |
| 02:17 | <j0hng4lt> | (like with sql) |
| 02:18 | <j0hng4lt> | most of the folks who actually develop web pages know anything beyond what gets displayed in the browser and what doesn't |
| 02:19 | <j0hng4lt> | But I don't see why incorrect syntax couldn't be handled further upstream |
| 02:19 | <j0hng4lt> | (e.g. by development tools) |
| 02:19 | <j0hng4lt> | but theory has to deal with reality... |
| 02:20 | <j0hng4lt> | we have to deal with tag-soup parsing because we have a ton of tag soup... and wishing it disappeared wouldn't make it go away. |
| 02:20 | <j0hng4lt> | What could've handled it back in the day is a strict-mode switch that gets ignored by old browsers and is honored by newer ones |
| 02:21 | <j0hng4lt> | e.g. X-Parse-Mode=xml |
| 02:22 | <j0hng4lt> | the switch that was chosen would break everything in IE |
| 02:22 | <j0hng4lt> | so everyone would serve it as text/html |
| 02:23 | <j0hng4lt> | since most people used ie anyway |
| 02:24 | <j0hng4lt> | many third-party tools would continue with document.write(), direct output of html to stdout (<a href="<?=$foo->getLink() ?>">...</a> resulting in <a href="<b>Error:</b> File not found>...</a>) |
| 02:24 | <j0hng4lt> | this made strict mode next to impossible |
| 02:26 | <roc> | awww I was just about to comment |
| 02:27 | <roc> | I will anyway |
| 02:27 | <roc> | Postel's world was a world where most people producing wire data knew what the standard was and were trying to hit it |
| 02:28 | <roc> | the world of Web standards is one where most producers of wire data have know little about the standards, don't care, and largely wouldn't be able to get it right if they did |
| 02:29 | <roc> | optimal behaviour in these two worlds is completely different |
| 02:29 | <roc> | </end> |
| 02:32 | <MikeSmith> | roc: don't try to cloud the discussion with irrelevant details about the realities of the environment |
| 02:34 | <j0hng4lt> | In my opinion, the phenomenon that is the World Wide Web defies the logic of development / software engineering... It's a huge mess, and I find it rather surprising that it works at all... |
| 02:34 | <j0hng4lt> | But the job of the Web developer has consisted of pushing square pegs through round holes from day 1 |
| 02:35 | <roc> | ha |
| 02:35 | <j0hng4lt> | (remember spacer gifs, two-layered tables to create a one-pixel border, javascript branching, etc... |
| 02:35 | <j0hng4lt> | ) |
| 02:35 | <j0hng4lt> | And it's not anyone's fault... |
| 02:35 | <fecalpatties> | shit, i remember flaming horizontal rules and animated email box icons |
| 02:36 | <j0hng4lt> | The installation base and the heterogeniety of the systems involved is unprecedented. |
| 02:37 | <j0hng4lt> | The most common anti-pattern found on the Web is Abstraction Inversion |
| 02:37 | <j0hng4lt> | http://en.wikipedia.org/wiki/Abstraction_inversion |
| 02:37 | <j0hng4lt> | Here are a few examples: |
| 02:37 | <MikeSmith> | j0hng4lt: I think you mean Inversion Abstraction |
| 02:37 | <MikeSmith> | (common mistake) |
| 02:38 | <j0hng4lt> | Well... I learned the mistake from Wikipedia ;) |
| 02:38 | <j0hng4lt> | here are a few examples: |
| 02:38 | <j0hng4lt> | - Using DIVs to render pixels in plotting libraries |
| 02:39 | <j0hng4lt> | - Using a hidden iframe to communicate with the server from a client-side script |
| 02:39 | <MikeSmith> | j0hng4lt: Wikipedia is yet another example of Inversion Abstraction |
| 02:39 | <j0hng4lt> | (before the invention of ajax) |
| 02:40 | <j0hng4lt> | - using background positioning to load multiple images at once |
| 02:41 | <j0hng4lt> | Another big challenge on the Web is the fuzzy border between users and programmers |
| 02:44 | <j0hng4lt> | (hence the reverse engineering of malformed code) |
| 02:47 | <MikeSmith> | indeed, hence |
| 02:48 | <MikeSmith> | j0hng4lt: you should consider writing about some of these theories |
| 02:50 | <MikeSmith> | we sometimes tend to spend too much time focusing on trying to find pragmatic ways to solve existing problems |
| 02:52 | <j0hng4lt> | I probably will... I'm thinking of starting a small tech blog... |
| 02:53 | <j0hng4lt> | Some people have pointed out that many programmers have failed to produce valid (XHTML|XML|HTML5|SGML|insert your language here). |
| 02:53 | <j0hng4lt> | In my opinion, most people, including myself, will fail to consistently produce anything valid by hand |
| 02:53 | <j0hng4lt> | unless it's a simple example you can literally "run in your head" |
| 02:54 | <MikeSmith> | yes, fail |
| 02:54 | <MikeSmith> | we are all failures and admitting that is Step 1 in getting better |
| 02:54 | <MikeSmith> | j0hng4lt: I suggest titling that blog either "Hence" or "Inversion Abstraction" |
| 02:54 | <j0hng4lt> | lol |
| 02:57 | <j0hng4lt> | I personally believe no matter what syntax you choose, human nature being what it is, there needs to be a tool somewhere between the keyboard and the user's screen that catches the most obvious errors. |
| 02:57 | <MikeSmith> | j0hng4lt: that solution sounds to practical |
| 02:57 | <MikeSmith> | it needs... more theory, more Wikipedia citations, or something |
| 02:58 | <MikeSmith> | 1. We admitted we were powerless over the Web — that the Web has become unmanageable. |
| 02:58 | <MikeSmith> | 2. Came to believe that a Power greater than ourselves could restore sanity to the Web. |
| 02:58 | <j0hng4lt> | Why can C# programmers produce valid C#? Actually, they can't. The compiler cactches the errors for them... And Visual Studio catches most errors before they even hit Compile. |
| 02:59 | <MikeSmith> | that's an interesting theory too |
| 03:01 | <j0hng4lt> | One of the reasons why most of the mark-up on the Web is syntactically invalid, whereas most C# code is valid (despite HTML and XHTML constructs being far simpler than C#) is that the mark-up goes to the browser before anyone or anything sees it, and nothing sees it until it's FUBAR (i.e. until it's so messed up that the HTML parser would need an Artificial Neural Network to figure it out). |
| 03:02 | <j0hng4lt> | "Postel-enabled" browsers aren't going away anytime soon, love them or hate them... |
| 03:02 | <MikeSmith> | yeah, it's clearly something in need of government regulation |
| 03:03 | <j0hng4lt> | Hell no... |
| 03:03 | <MikeSmith> | http://blogs.msdn.com/b/ie/archive/2010/06/28/measuring-web-page-performance.aspx |
| 03:03 | <j0hng4lt> | Would you give the Internet to the people who are running the post office? |
| 03:04 | <j0hng4lt> | I'm thinking of a validator that gets attached to all commonly used development environments (ASP.Net with and without MVC, PHP, Python, Ruby, etc) |
| 03:04 | <j0hng4lt> | as a post-processor |
| 03:04 | <j0hng4lt> | which can be disabled in production mode |
| 03:04 | <j0hng4lt> | (in case you're using third-party tools that could mess up your mark-up) |
| 03:04 | <j0hng4lt> | but catches everything during development |
| 03:05 | <j0hng4lt> | i.e. before it goes to the browser |
| 03:06 | <othermaciej> | Lachy: sadly, I don't think the actual terms of the Bilski decision are good news |
| 03:07 | <MikeSmith> | j0hng4lt: http://about.validator.nu/#src |
| 03:08 | <MikeSmith> | othermaciej: aza raskin had a good tweet along those lines |
| 03:08 | <MikeSmith> | http://twitter.com/azaaza/status/17270383318 |
| 03:08 | <MikeSmith> | "Historic day in patent law. Supreme court says business as usual for business and software patents." |
| 03:09 | <MikeSmith> | j0hng4lt: do you do any Java programming? |
| 03:09 | <j0hng4lt> | Not at the moment, but I've done some |
| 03:10 | <MikeSmith> | othermaciej: would be good if you or others on your team might could take a look at the Web Timings draft |
| 03:10 | <MikeSmith> | http://dev.w3.org/2006/webapi/WebTiming/ |
| 03:10 | <MikeSmith> | if you've not already |
| 03:10 | <j0hng4lt> | I'm familiar with the Java APIs for XML |
| 03:11 | <j0hng4lt> | as well as JSF, Struts (1.x only) and "classic" JSP |
| 03:12 | <j0hng4lt> | Right now, I code in php (my day job) and C#/ASP.net with jQuery + jQuery-UI (on my own time) |
| 03:16 | <j0hng4lt> | MikeSmith: Why were you asking me about Java? Was it in relation to the validator? |
| 03:17 | <MikeSmith> | j0hng4lt: validator.nu is a Java app that runs as a RESTful Web service and has an API that allows you to hook it into any dev environment you want |
| 03:18 | <MikeSmith> | you can check out the source and run it locally |
| 03:18 | <MikeSmith> | or you can set up a instance of it for shared use over a LAN |
| 03:18 | <j0hng4lt> | thanks for the link |
| 03:18 | <MikeSmith> | or whatever |
| 03:19 | <MikeSmith> | http://wiki.whatwg.org/wiki/Validator.nu_Web_Service_Interface |
| 03:20 | <MikeSmith> | if you try it and find that it's missing something you'd like for it to have, the backend Java is well-designed code that is pretty easy to hack on |
| 03:20 | <MikeSmith> | using SAX to hook various parts together |
| 03:21 | <j0hng4lt> | REST is another interesting phenomenon... I do understand the initial arguments (i.e. using a GET request when you're not modifying stuff), but in a way, it's another case of going "back to the 90s" |
| 03:22 | <othermaciej> | MikeSmith: we plan to look at it, and also the associated charter for the potential WG around this |
| 03:24 | <j0hng4lt> | i.e. In practice, you end up having to simplify / flatten the data structures you send to the service (i.e. you go from XML-serialized objects to key/value pairs), and the response could be just about anything as well - xml, json or even key-value pairs... |
| 03:24 | <j0hng4lt> | I remember coding Web services before they were called Web services |
| 03:25 | <j0hng4lt> | That was exactly what you would end up doing... You would send a GET or a POSt with URL-encoded key/value pairs and you would get the same thing back |
| 03:25 | <MikeSmith> | nothing wrong with the 90s |
| 03:25 | <MikeSmith> | we could use some more of the 90s around here |
| 03:25 | <MikeSmith> | the Web has too much of the 00s |
| 03:26 | <MikeSmith> | "The World Wide Web: The 00 Years" |
| 03:26 | <MikeSmith> | in terms of standards at least, not the best years |
| 03:27 | <j0hng4lt> | My biggest frustration with the Web has been that, unlike with desktop development, you never get adequate tooling... The moment tools get built up, either they get obsolete or the Web goes up in arms with torches and pitchforks, calling for simplification |
| 03:27 | <j0hng4lt> | then you get a conspiracy between Mort and Einstein to drive out Elvis |
| 03:28 | <j0hng4lt> | So in the end, you're always coding in notepad (or do the equivalent thereof)... |
| 03:29 | <j0hng4lt> | The worst thing of the 00 years had to do with standards designed by committee |
| 03:29 | <j0hng4lt> | that never got implemented |
| 03:29 | <j0hng4lt> | (e.g. XHTML2, CSS2, the beginnings of CSS3) |
| 03:30 | <j0hng4lt> | the positive part of the OOs was the drive for structure (stricter syntax, schemas, etc.) |
| 03:30 | <j0hng4lt> | + no more if (ie) { doThis() } else if (netscape) { doThat() } else { goToHell() } |
| 03:35 | <j0hng4lt> | But in a way, we're seeing the Revenge of Mort in full swing... Loose syntax, no data structures, more copy-and-paste development, less widget-building |
| 03:36 | <j0hng4lt> | Ideally, I would like to be in a situation where Mort and Elvis can use the same Web without competing for space... |
| 04:13 | karlcow | noticed that less and less people do ascii art in their explanation of features (found cute ascii art for FIG element in 1995) |
| 04:13 | <karlcow> | example: http://lists.w3.org/Archives/Public/www-html/1995Jul/0105.html |
| 04:13 | <MikeSmith> | well, that's a sad change |
| 04:13 | <MikeSmith> | charter for proposed Web Performance WG has been updated - |
| 04:13 | <MikeSmith> | http://www.w3.org/2010/06/webperf.html |
| 04:14 | <MikeSmith> | proposed chair is Arvind Jain, Google |
| 04:42 | <MikeSmith> | my strategy of trying to generate as much animi |
| 04:42 | <MikeSmith> | oops |
| 04:43 | <MikeSmith> | wrong channel |
| 04:43 | <MikeSmith> | or maybe not |
| 04:44 | <MikeSmith> | fwiw, I was saying that my strategy of trying to generate as much animosity toward the W3C as possible seems to be failing |
| 04:47 | <MikeSmith> | I had been led to believe that the recent Stephen Shankland CNET article would engender quite a bit of righteous indignation and some robust W3C-bashing among the widest number of communities possible |
| 04:47 | <MikeSmith> | but I've been disappointed |
| 04:47 | <karlcow> | :) |
| 04:48 | <MikeSmith> | maybe if that article could instead have been written by someone less insightful, less prone to actually considering things carefully |
| 04:48 | <karlcow> | I had not even noticed this article and person before you mentioning it :) |
| 04:50 | <MikeSmith> | karlcow: well, you're clearly not doing enough to try to help whip things into a proper frenzy |
| 04:50 | <karlcow> | hmm whips… |
| 04:52 | <MikeSmith> | anyway, lacking the sufficient quantity of W3C-bashing that I had been expecting to get from that this week, I guess I have to fall back to again mostly relying on some folks here to regularly supply some knee-jerk W3C bashing now and then |
| 04:53 | <MikeSmith> | so I hope people won't disappoint me here like others elsewhere have |
| 04:54 | <MikeSmith> | I'm depending on it |
| 04:55 | <karlcow> | it is a looooooooooong article |
| 04:56 | <MikeSmith> | karlcow: it's designed to be savored |
| 04:56 | <MikeSmith> | enjoy every moment |
| 04:56 | <MikeSmith> | there's only one "first time" for everything |
| 05:08 | <karlcow> | http://infomesh.net/html/history/early/ |
| 05:08 | <karlcow> | The Early History of HTML |
| 05:09 | <karlcow> | Full HTML Test Suite - http://www.w3.org/History/1991-WWW-NeXT/Implementation/Test/backup_of_test.html |
| 05:09 | <karlcow> | :) |
| 05:38 | <hsivonen> | wow. http://www.w3.org/DesignIssues/RDF-Future.html doesn't have a year in the URI |
| 05:41 | <karlcow> | hsivonen: none of DesignIssues have year in URIs |
| 05:58 | <hsivonen> | karlcow: ah. so the directory is so *old* it doesn't have a year. I was hoping *new* articles no longer have a year. |
| 06:05 | <karlcow> | hsivonen: this directory is just the directory of the personal notebook of the guy who created this web site you pointed to ;) |
| 06:05 | <MikeSmith> | " I would in fact prefer, instead of <H1>, <H2> etc for headings [those come from the AAP DTD] to have a nestable <SECTION>..</SECTION> element, and a generic <H>..</H> which at any level within the sections would produce the required level of heading." |
| 06:06 | MikeSmith | had never read that particular bit before |
| 06:07 | <MikeSmith> | ah, I get to the end of page and see that Sean Palmer put this together |
| 06:07 | <MikeSmith> | bravo Sean |
| 06:07 | MikeSmith | wonders what stuff Sean is working on these days |
| 06:08 | <MikeSmith> | http://inamidst.com/sbp/ |
| 06:09 | <MikeSmith> | "As of April 2010, I've been working on automatic table of contents generation using jQuery,".. |
| 06:09 | <karlcow> | sean is one of my heroes :) |
| 06:12 | <MikeSmith> | karlcow: http://inamidst.com/whits/2008/technobunkum is still relevant |
| 06:16 | <MikeSmith> | as are http://inamidst.com/whits/2008/ditching and http://inamidst.com/whits/2008/ambient |
| 06:20 | <karlcow> | hehe but but but MikeSmith if there is no more technobunkum, there is no more Gargantua, Olympus, Valhalla, … |
| 06:20 | <karlcow> | suddenly we would be forced to solve real issues, I mean plain boring stuff |
| 06:21 | <karlcow> | no no no no… this is a nightmare ;) |
| 06:23 | <karlcow> | ambient information is cool. Not always easy to put in place, specifically in a non intrusive way (aka GPS = cool, Celltower loc = bad) |
| 06:23 | <MikeSmith> | I guess a lot of real problems are often not particularly fun puzzles |
| 06:23 | <karlcow> | or you really have to be a geek to really do cool stuff with ambient |
| 06:24 | <MikeSmith> | the problems that are more mundane but are nevertheless difficult to solve |
| 06:24 | <karlcow> | ooops talking about ambient… 1:30am |
| 06:24 | <MikeSmith> | the problems that don't lead to particularly elegant solutions, but instead to kind of messy ones |
| 06:24 | <karlcow> | need to do boring puzzles tomorrow morning for a steak in my plate |
| 06:24 | <MikeSmith> | karlcow: that's early, man |
| 06:24 | <MikeSmith> | hai hai |
| 06:25 | <karlcow> | oyasumi nasai |
| 06:27 | <MikeSmith> | oyasumi |
| 07:04 | <MikeSmith> | dbaron: I'm not sure I understand your latest public-html message correctly |
| 07:04 | <MikeSmith> | so let me give an example and ask if it's the same as what you have in mind |
| 07:05 | MikeSmith | trots out a recent hobbyhorse |
| 07:05 | <dbaron> | ok |
| 07:06 | <MikeSmith> | the example is related to parsing of MathML content |
| 07:07 | <MikeSmith> | if the parsing algorithm encounters a MathML <annotation-xml> element, it essentially expects the content of that to be either and <svg> element or more MathML content |
| 07:08 | <MikeSmith> | but if the byte stream it parsing instead has HTML content in there, it goes back and pops elements off the stack until it encounters either a <math> or <svg> element |
| 07:09 | <MikeSmith> | the end |
| 07:11 | <hsivonen> | dbaron: I believe it is implied that there must be no other data structures hiding in the DOM tree than a Level 2 DOM and the references from form controls to forms |
| 07:11 | <hsivonen> | dbaron: but it's possible that it's not actually stated anywhere |
| 07:12 | <hsivonen> | it probably should be stated to make it clear that the kind of hidden data Opera and IE have is non-conforming |
| 07:13 | <hsivonen> | oh, and there are also the HTMLness flag and the quirkiness enum |
| 07:14 | <dbaron> | "No other data structures" might be too strong; it might make more sense to limit certain classes of behavior to not depending on other data structures |
| 07:15 | <dbaron> | in any case, I looked for a bit, and couldn't find any such requirements |
| 07:15 | <dbaron> | MikeSmith, I don't see how your example is related to what I was saying, although I could be missing something |
| 07:22 | <MikeSmith> | dbaron: no, I don't think you're missing anything -- after re-reading I see that you're referring to something other that what I meant by that example |
| 07:22 | <MikeSmith> | the "differences that must not matter" part in your subject line should have clued me in |
| 07:30 | <othermaciej> | good evening folks |
| 07:31 | <othermaciej> | hsivonen, dbaron: I think it would be a valid statement of fact that no conformance requirements in the spec would depend on the input document in any way other than as it affects the DOM |
| 07:31 | <othermaciej> | but I do not believe there is a requirement that browsers may not invent their own such behaviors |
| 07:31 | <othermaciej> | (so long as those behaviors are consistent with the existing UA conformance requirements) |
| 07:31 | <othermaciej> | I think it would be hard to scope it appropriately |
| 07:32 | <othermaciej> | for example, it's clear to me that a syntax-highlighted source view should be able to depend on the exact text input |
| 07:32 | <othermaciej> | or XHR.responseText |
| 07:39 | <dbaron> | but could one say, for example, that anything that depends on the element type (other than a display of the tag name) must not depend on things other than X? |
| 07:51 | <hsivonen> | othermaciej: source view isn't really part of the interoperable platform |
| 07:53 | <othermaciej> | it's hard to construct conformance criterion that's a universal negative |
| 07:54 | <othermaciej> | maybe it would be preferable to just cite specific things that MUST NOT be affected by anything in the source document besides the output of the parsing algorithm |
| 08:08 | <MikeSmithX> | dbaron: btw, I wanted to ask if you might be planning an updates to the CSS3 syntax spec |
| 08:09 | <MikeSmith> | in particular, to the grammar and tokenizer parts |
| 08:09 | <MikeSmith> | I ask because Dan Connolly recently wrote a new CSS Validator (in Scala) |
| 08:10 | <MikeSmith> | with is designed to take a CSS grammar as an input |
| 08:10 | <MikeSmith> | http://dvcs.w3.org/hg/css-val/raw-file/f5931981d7dd/README.html |
| 08:11 | <MikeSmith> | " the CSS 2.1 specification includes a uniform grammar, applicable to all CSS versions. The original (2002) W3C CSS Validator predates this uniform grammar and maintenance is becoming an undue burden as a result. This is an attempt to build a validator using the uniform syntax." |
| 08:11 | <dbaron> | I'm not planning an update; it's currently basically unowned. |
| 08:12 | <dbaron> | somebody writing a validator might make a good owner, though :-) |
| 08:12 | <dbaron> | anyway, g'night |
| 08:13 | <annevk> | oh god -- even the Microsoft canvas tests assume there's no such thing as a high-resolution canvas backing bitmap |
| 08:14 | <annevk> | nobody is gonna get it right |
| 08:16 | <hsivonen> | did Apple ship high-res backing bitmap by default in Mobile Safari on iPhone 4? |
| 08:16 | <hsivonen> | 'cause if they didn't, who is going to? |
| 08:17 | <annevk> | well yeah, the iPhone 4 would certainly be interesting to test on this |
| 08:22 | <annevk> | I don't get the Google China thing |
| 08:22 | <annevk> | they cannot redirect to google.com.hk but they can point to it and users can access it either way but the former is forbidden but the latter is ok? |
| 08:22 | <annevk> | it just doesn't make much sense |
| 08:26 | <MikeSmith> | karlcow: http://inamidst.com/whits/2008/technobunkum is still relevant |
| 08:26 | <MikeSmith> | oops |
| 08:26 | <MikeSmith> | kennyluck: http://inamidst.com/whits/2008/technobunkum |
| 08:26 | <MikeSmith> | kennyluck: http://inamidst.com/whits/2008/ditching and http://inamidst.com/whits/2008/ambient |
| 08:29 | kennyluck | frowned a little bit when he saw words such as "psychological" and "technobunkum". |
| 08:38 | <foolip> | annevk: they just redirected, at least when I left Beijing a few weeks ago |
| 08:39 | <foolip> | annevk: also, expecting the mainland "regulation" of the internet is... funny :) |
| 08:40 | <annevk> | foolip, they're switching to linking, see latest blog post |
| 08:47 | <foolip> | hehe, they even used an image of a search form that links to .com.hk |
| 08:47 | <foolip> | this ICP thing they're talking about is actually quite funny |
| 08:48 | <foolip> | in http://www.miibeian.gov.cn/share/liuchengtu2.jsp there's a flowchart for how to register |
| 08:49 | <foolip> | the first step is "Use Windows IE6 browser to log on to www.miibeian.gov.cn" |
| 08:49 | <foolip> | that's Chinese e-government for you :) |
| 09:36 | <asmodai> | Is ff 3.7 nightly what will become 4.x? And does anyone know if it can coexist with a stable ff (without screwing up any existing profiles). |
| 09:37 | <hsivonen> | asmodai: the version number of nightlies got changed less than 6 hours ago. |
| 09:37 | <hsivonen> | asmodai: so the next round of nightlies isn't "3.7" anymore |
| 09:38 | <hsivonen> | asmodai: I wouldn't run the nightlies with the same profile as 3.6.x |
| 09:38 | <hsivonen> | asmodai: I did, and Firefox Sync and password manager got confused |
| 09:40 | <asmodai> | hsivonen: I will run it on a Windows box. So I should use the profile manager, create a 4.x profile, but can I automate something to have 4.x use that profile (or at least ask which one to use prior to starting)? |
| 09:41 | <hsivonen> | asmodai: the profile manager has a checkbox for making it ask before starting |
| 09:42 | <annevk> | so when is 4 shipping? |
| 09:47 | <asmodai> | annevk: at the end of the year I think I just read |
| 09:47 | asmodai | looks for link |
| 09:47 | <hsivonen> | annevk: https://wiki.mozilla.org/Releases has a tentative lower bound data point |
| 09:47 | <asmodai> | October/November |
| 09:47 | <asmodai> | http://geektech.geektech.netdna-cdn.com/wp-content/uploads/2010/06/Firefox4ReleaseDate.png |
| 09:48 | <hsivonen> | 3.6.6 shipped already |
| 09:48 | <hsivonen> | and there was no 3.6.5 |
| 09:49 | <asmodai> | hsivonen: I guess the 3.7 nightly will update itself to whatever the new one has as version number? |
| 09:49 | <hsivonen> | asmodai: it should |
| 09:50 | asmodai | tries ;) |
| 09:54 | <asmodai> | I forgot, can't run regular and new side by side |
| 09:54 | <hsivonen> | asmodai: it looks like the nightlies haven't cycled yet after the version number change |
| 09:57 | <asmodai> | hsivonen: That's ok -- got a 3.7apre6 now |
| 09:57 | <asmodai> | a6pre even |
| 10:10 | <gsnedders> | annevk: thx |
| 10:12 | <annevk> | gsnedders, timely |
| 10:12 | <gsnedders> | annevk: 13 hours? Nothing. |
| 10:20 | <annevk> | gsnedders, at some point you have to start living in the present :p |
| 10:20 | <gsnedders> | annevk: It's hard to leave the past behind |
| 10:23 | <jgraham> | gsnedders: There is hardly any past behind you |
| 10:24 | <annevk> | I was gonna say "drink more" but it would have just gone down from there and it was pretty bad already :) |
| 10:25 | <gsnedders> | annevk: Hey, I had a beer or two when you told me I should have a beer for you in your abscence :P |
| 11:28 | <asmodai> | hsivonen: Well |
| 11:28 | <asmodai> | hsivonen: Do people depend heavily on those litmus tests? |
| 11:40 | <hsivonen> | asmodai: which litmus tests? |
| 11:40 | <asmodai> | litmus.mozilla.org |
| 11:40 | <hsivonen> | asmodai: I have no idea |
| 11:41 | <asmodai> | Grrrr *slaps the test runner* |
| 11:41 | <asmodai> | Cute, press a button to submit a single test result, instead of sends all >_< |
| 11:43 | <hsivonen> | hmm. bulletproofing new elements by using <div>s also pretty much defeats any convenience benefits from the new elements |
| 11:43 | <hsivonen> | see http://tantek.com/presentations/2010/06/html5-vtm/ |
| 11:47 | <jgraham> | and yeah, that nested div thing seems pointless |
| 11:49 | <jgraham> | the microdata vs microformats opinion seems biased |
| 11:50 | <jgraham> | web storage already has wide support, no? |
| 11:50 | <jgraham> | and postMessage + friends |
| 12:38 | <annevk> | hmm, why do Minefield releases "installed" in a separate directory use plug-ins from the version of Firefox that ships with the distribution? |
| 12:38 | <annevk> | specifically, the Ubuntu theming nonsense |
| 12:39 | <crash\> | there's a seperate plugin dir |
| 12:39 | <annevk> | okay, just curious |
| 12:39 | <annevk> | i wonder why that confusing bookmark menu is there |
| 12:39 | <annevk> | has that always been there? |
| 12:40 | <crash\> | seems all Firefox version looking for that directory |
| 12:40 | <annevk> | you'd think people stopped using bookmarks the moment search engines started to work |
| 12:41 | <annevk> | Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:2.0b2pre) Gecko/20100629 Minefield/4.0b2pre |
| 12:41 | <annevk> | Opera/9.80 (X11; Linux x86_64; U; en) Presto/2.6.30 Version/10.60 |
| 12:42 | <annevk> | and I thought ours was bloated :/ |
| 12:43 | <hsivonen> | annevk: you can disable the Ubuntu addon via the add-on manager if you don't like it |
| 12:43 | <crash\> | It bloated since old script use bad detection for browsers :( |
| 12:43 | <annevk> | it just disabled itself automatically it seems |
| 12:43 | <annevk> | yeah, that's why we have that Version/10.60 crap |
| 12:44 | <annevk> | we shouldn't have done that I think |
| 12:44 | <annevk> | Presto/... is enough imo |
| 12:44 | <crash\> | you should see the Chromes UA string |
| 12:44 | <annevk> | rather not :) |
| 12:45 | <crash\> | Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_6_4; en-US) AppleWebKit/534.2 (KHTML, like Gecko) Chrome/6.0.451.0 Safari/534.2 |
| 12:45 | <annevk> | lol |
| 12:45 | <hsivonen> | annevk: if you put Opera there, trying to get people to sniff for Presto is about as futile (or more futile) than getting people to sniff for Gecko instead of Firefox |
| 12:46 | <hsivonen> | the same goes for WebKit and Safari |
| 12:46 | <crash\> | UA string sniffing is bad at all |
| 12:46 | <crash\> | I never needed it |
| 12:46 | <hsivonen> | what's the deal with the underscores in the OS X version in Chrome's UA string? |
| 12:47 | <crash\> | donno |
| 12:47 | <annevk> | the idea was that you could sniff for the engine common across Opera deliveries; obviously not a very good idea |
| 12:47 | <hsivonen> | crash\: well, if a browser vendor doesn't want to be sniffed, the vendor should kill the User-Agent header or freeze it to a constant |
| 12:47 | <annevk> | though in theory it sounds nice |
| 12:47 | <gsnedders> | hsivonen: IIRC it's how it is represented in some system define |
| 12:47 | <gsnedders> | But that would be compile time |
| 12:47 | <gsnedders> | Hmm |
| 12:47 | <gsnedders> | Dunno |
| 12:47 | <hsivonen> | annevk: "normal people" don't think in terms of engines. they think in terms of browser names |
| 12:48 | <crash\> | hsivonen: they use it themself, for example at addons.mozilla.org |
| 12:49 | <hsivonen> | crash\: unfortunately, I'm very aware of Mozilla's own sites using the UA string |
| 12:49 | <crash\> | to detect if an addon is avaiable for the curent Firefox version |
| 12:49 | <annevk> | I guess at the time this discussion went on I should have weighed in and pointed at what a failure Firefox/Gecko is for the Mozilla community but I did not really care back then :/ |
| 12:49 | <hsivonen> | Dear lazy IRC, how do I tell pulseaudio on Lucid to use a USB headset for Skype but built-in audio output for everything else? |
| 12:50 | <annevk> | Well, it's not like I really care now, but trimming HTTP weight and fingerprinting variables is something that's somewhat fun |
| 12:53 | <roc> | Gecko clearly isn't a complete failure, since Webkit adopted it |
| 12:53 | <jgraham> | roc: I plan to quote that out of context :) |
| 12:54 | <hsivonen> | roc: they adopted it before Firefox was as well known as it is today |
| 12:59 | <roc> | jgraham: yeah, I realized after I said it that that could get me in trouble :-) |
| 13:00 | <roc> | hsivonen: maybe so |
| 13:00 | <roc> | the whole engine vs browser distinction is a quandry |
| 13:01 | <roc> | it's good in principle but it confuses people |
| 13:17 | <Dashiva^2> | hsivonen: lazy IRC is too lazy to reply |
| 13:19 | <Dashiva^2> | Re: nested div, I thought the IE shiv fixed all the problems with new block elements? |
| 13:29 | <boblet> | anyone know about HTML5 support in .net? |
| 13:31 | <hsivonen> | boblet: someone from the .net side of Microsoft was implementing the parsing algorithm |
| 13:32 | <boblet> | hsivonen: trying to answer a query about <input type="number"> getting an invalid type error, and Google/Bing can’t help. did you hear an eta? |
| 13:32 | <boblet> | (what idiot approved using such a generic name. jeez) |
| 13:33 | <hsivonen> | boblet: I don't know of an eta. |
| 13:33 | <hsivonen> | volkmar: ^ |
| 13:34 | <boblet> | hsivonen: thanks anyhow. good to be able to pass on that it’s in progress |
| 13:38 | <hsivonen> | boblet: I don't know of progress, either. In fact, I don't know what the issue is. |
| 13:38 | <boblet> | hsivonen: “Our dev team use .net 3.5. We are attempting to use HTML5 form input fields. When using <input type=”number” /> we get a “Parser Error Message: 'number' is not a valid type for an input tag".” |
| 13:40 | <hsivonen> | boblet: oh. that's probably something on the ASP.Net layer |
| 13:42 | <boblet> | hsivonen: oho! that’s improved my search results no end. thanks! |
| 13:47 | <boblet> | woah, forum.asp.net is pretty scary |
| 13:50 | <hsivonen> | http://inamidst.com/whits/2008/ditching is interesting |
| 13:51 | <MikeSmith> | hsivonen: I figured you had read that previously.. I thought I was the only one who hadn't |
| 13:51 | <MikeSmith> | kennyluck and I were talking about that earlier today |
| 13:52 | <MikeSmith> | boblet: learning asp.net stuff is fun.. there's a whole series of books and courses about it.. you can spend as much time as you want and there's always more to learn |
| 13:53 | <boblet> | MikeSmith: you had a typo there s/as much time as you want/as much money as you want/ |
| 13:53 | <MikeSmith> | hey man |
| 13:53 | <MikeSmith> | don't disrespect technologies |
| 13:53 | <MikeSmith> | "to each his own" as they say |
| 13:54 | <boblet> | no disrespect intended |
| 13:54 | <hsivonen> | MikeSmith: I hadn't read it before |
| 13:54 | <boblet> | (at least not of that part) |
| 13:54 | <MikeSmith> | boblet: I remember the introductory book for asp.net is called "How to distinguish your ass from a hole in the ground" |
| 13:54 | <MikeSmith> | I found that book really helpful |
| 13:54 | <hsivonen> | why is Modernizr dual-licensed under BSD and MIT? |
| 13:54 | <kennyluck> | I think the document is just too old. |
| 13:55 | <MikeSmith> | boblet: and then of course there is a natural progression of the various asp.net up from that fundamental level. it's all very scientific almost |
| 13:55 | <hsivonen> | kennyluck: Sean Palmer's blog post? |
| 13:55 | <boblet> | hsivonen: woah this is some strong stuff |
| 13:55 | <MikeSmith> | paul_irish: <hsivonen> why is Modernizr dual-licensed under BSD and MIT? |
| 13:55 | <kennyluck> | Yes, hsivonen. |
| 13:56 | <hsivonen> | kennyluck: even though it's 2 years old, the content seems on-point and fresh to me |
| 13:57 | <annevk> | hsivonen, hmm, that was a couple of years ago |
| 13:57 | <annevk> | hsivonen, weird that I haven't seen it then |
| 13:57 | <kennyluck> | He indeed shows the point about RDF -> you can easily write a RDF client. no documentView or whatever that's needed to implement a HTML client. |
| 13:58 | <kennyluck> | His client failed to attract folks, and I think this might be what his frustration came from. |
| 13:59 | <kennyluck> | s/what/where/ |
| 14:01 | <annevk> | heh, that's so vastly different from what I read into it |
| 14:01 | <boblet> | use cases and practical examples are a big help for anyone trying to read a spec too |
| 14:01 | <hsivonen> | kennyluck: I didn't read it that way |
| 14:02 | <kennyluck> | Well, that's my personal explanation. |
| 14:02 | <hsivonen> | kennyluck: I think the point about use cases and solutions looking for problems makes sense regardless of the popularity of his software |
| 14:02 | <kennyluck> | Yes, I agree. |
| 14:03 | <kennyluck> | But I believe we all know about this. It shouldn't be associated with Semantic Web. |
| 14:04 | <annevk> | Why not? It's been repeatedly cited as a problem of RDF... |
| 14:05 | <kennyluck> | We can blame those professors. :) And those who mention the word "ontology" again and again. |
| 14:06 | <boblet> | lol: http://iscss3partofhtml5.com/ |
| 14:07 | <boblet> | (sorry if that’s old news) |
| 14:07 | <hsivonen> | kennyluck: you might be in denial, since the URL in your /wii points to an N3 resource :-) |
| 14:07 | <zcorpan_> | boblet: next up is ishtml5partofhtml5.com |
| 14:08 | <annevk> | also http://isgeolocationpartofhtml5.com/ |
| 14:08 | <boblet> | zcorpan_: stop it, you’re killing me here! |
| 14:08 | boblet | needs to stop laughing so he can breathe |
| 14:11 | <Dashiva^2> | annevk: Missing a "yet" in the domain name :P |
| 14:20 | <annevk> | someone should get partofhtml5.com |
| 14:23 | <kennyluck> | By the way, Sean Palmer post gives us a hint on a use case of Semantic Web/RDF - decentralized issue tracking systems, so that we can make sure what we are working on (an issue) has a purpose (an issue that blocks this issue). |
| 14:24 | <kennyluck> | The reason why we need "decentralization" is the following use case. MikeSmith might want to create an issue "(super master bug) HTML5 implematations". |
| 14:24 | <kennyluck> | And this is blocked by the Mozilla HTML5 implemenation master bug and Webkit HTML5 implementation master bug. |
| 14:26 | <kennyluck> | As long as the Mozilla/Webkit HTML5 implementaion bugs are closed. We can close the "HTML5 implemetations" bug and release the W3C CR (correct me if I am wrong) |
| 14:26 | <Philip`> | He can do that without any RDF or semanticity, just by adding himself to the CC list on the Mozilla and WebKit bugzillas and noticing updates |
| 14:28 | <kennyluck> | Well, what I suggested was the feature -> "Allow an issue in a issue tracking system to block other issue in another domain". RDF is not required, indeed. |
| 14:28 | <kennyluck> | Any format that allows Linking *externally* is good enough. Microdata could be a choice. |
| 14:30 | <Philip`> | Why is <p>This bug is blocked by <a href="http://bugzilla.example.org/12345">this other bug</a></p> insufficient? |
| 14:30 | <kennyluck> | It is good, but it might not be machine readable. |
| 14:30 | <Philip`> | Why does it need to be machine readable? |
| 14:31 | <kennyluck> | micrioformat, microdata, JSON, RDF, CSV, XML |
| 14:31 | <kennyluck> | Oh, sorry. I thought you asked about the formats. |
| 14:32 | <kennyluck> | Because then we can do more things with the data. Microdata is machine readable. |
| 14:32 | <Philip`> | What "more things"? |
| 14:32 | <mahound_> | well, the only advantage i see is that it can be closed automatically or something... |
| 14:32 | <kennyluck> | Like rendering dependency graph, etc. |
| 14:33 | <mahound_> | i mean... the blocked issue can crawl the other two and close itself when they're closed |
| 14:33 | <kennyluck> | mahound_, It's a good advantage, isn't it. Just think about why do we the the field "Blocking" |
| 14:33 | <Philip`> | Seems like working backwards from the solution (providing machine-readable data) to come up with hypothetical problems that it could help with |
| 14:34 | <kennyluck> | You don't like Microdata, Philip` ? |
| 14:34 | <mahound_> | Philip`, well, it's still useful :) |
| 14:35 | <Philip`> | In the concrete example that was given (moving the HTML5 spec to CR when implementors have closed their HTML5-implementation bugs) you wouldn't actually want to tie the closed status, because browsers say closed when fixed in source control but the spec needs to wait for them to ship |
| 14:35 | <annevk> | except it's not magically solved by adding some RDF |
| 14:36 | <mahound_> | of course it's discussible whether it would be worth implementing such a thing... |
| 14:36 | <Philip`> | so automatically responding to status changes won't help that use case |
| 14:36 | <kennyluck> | annevk, no it's not. Lots of hard work needs to be done to Bugzilla, Redmine, etc. |
| 14:37 | <kennyluck> | I can imagine,as long as we have such system, "Google Todos". |
| 14:37 | <mahound_> | i think microformats would be more than enough |
| 14:37 | <kennyluck> | Then you can ask "Oh God Google, could you give me something worth doing. Please no sudoko, which does not have a dependency". |
| 14:38 | <kennyluck> | mahound_, maybe. |
| 14:38 | <mahound_> | anyway, i think a simple link or CC is still the best solution ;) |
| 14:39 | <mahound_> | as Philip` says, the workflow is sometimes too complex |
| 14:39 | <mahound_> | and requires human intervention |
| 14:40 | <kennyluck> | Well, the goal is to make issue trackers go global. |
| 14:40 | <kennyluck> | Of course issue trackers don't understand human. |
| 14:42 | <mahound_> | speaking of microformats... when will i see native hCalendar support in major web browsers? |
| 14:43 | <kennyluck> | :) that I am interested in, too. |
| 14:43 | <mahound_> | i really don't get it, it's such an interesting, useful use case |
| 14:43 | <mahound_> | how many times do i have to copy-paste dates and times, and places to my google calendar... |
| 14:53 | <variable> | mahound_, Firefox has an addon that does that |
| 14:54 | <variable> | I duno about chrome |
| 14:54 | <kennyluck> | Operator you mean? |
| 14:55 | <variable> | kennyluck, I think so |
| 14:57 | <variable> | kennyluck, http://microformats.org/wiki/hcalendar-implementations#Firefox_Greasemonkey_Plugins |
| 15:01 | <boblet> | mahound_: I think FireFox was planning to release native support a while back, but decided to pull it and work on interface more. check azzazza’s blog |
| 15:03 | <variable> | boblet, link? |
| 15:04 | <kennyluck> | Hmm... I have to thank them then. I can't live without the Firefox smart bar. |
| 15:04 | <zcorpan_> | <mo><xsl:text disable-output-escaping="yes"><![CDATA[⁢]]></xsl:text></mo> |
| 15:05 | <boblet> | variable: wetware only orz |
| 15:05 | <zcorpan_> | http://www.google.com/codesearch/p?hl=en#FQzGEBTCSY4/soft/OMathML/releases/omathml-0.1.tar.gz%7CDuTpMw-4EDI/xsl/mathmlc2p.xsl&q=%3C(%5B%5E%5Cs/%5D+:)?math%20%3C(%5B%5E%5Cs/%5D+:)?mo%3E%5B%5E%3C%5D*%3C%5B%5E/%5D |
| 15:05 | <variable> | wetware ? |
| 15:06 | <boblet> | the “I think…” part |
| 15:06 | <variable> | boblet, I meant link to azzazza's blog |
| 15:06 | <boblet> | looking… i spelled it wrong. wrong number of zs? |
| 15:06 | <zcorpan_> | Hixie: i see <mo><mrow> here http://www.google.com/codesearch/p?hl=en#KGV8vs9gsGI/content/abacus/content/XNS/docs/BigDecimal.html&q=%3C(%5B%5E%5Cs/%5D%2B:)%3Fmath%20%3C(%5B%5E%5Cs/%5D%2B:)%3Fmo%3E%5B%5E%3C%5D*%3C%5B%5E/%5D&sa=N&cd=25&ct=rc |
| 15:07 | <zcorpan_> | Hixie: what made you special-case "mglyph" and "malignmark"? |
| 15:07 | <boblet> | http://www.azarask.in/blog/ |
| 15:07 | <variable> | boblet, thanks |
| 15:08 | <zcorpan_> | Hixie: that also has <mi><msub> |
| 15:09 | <boblet> | variable: sorry if it’s not Aza, it was a Firefox UI designer’s blog tho. I remember thinking just do it like Apple Mail |
| 15:10 | <variable> | boblet, if you find the link let me know please |
| 15:10 | <variable> | I'm goolging as well |
| 15:10 | <variable> | but my google fu isn't as good as it used to be. I think google is getting worse :\ |
| 15:11 | <boblet> | heh, the web is getting larger |
| 15:11 | <boblet> | will do |
| 15:13 | <boblet> | oh dear, it seems the English defeat has had an effect on my fellow doctors: http://isbrucepartofhtml5.com/ |
| 15:13 | <boblet> | they’re obviously hitting the gin still a little too hard |
| 15:14 | <Dashiva^2> | All these domain purchases are helping rejuvenate the world economy, if nothing else |
| 15:15 | <jgraham> | Also, knob has a k in it |
| 15:15 | <variable> | Everyone says domain purchases are cheap. but if you have 50 of them thats $500/yr :\ |
| 15:16 | <Rik`> | if you're going to do a lot jokes about html5, just buy partofhtml5.com |
| 15:22 | <Lachy> | Rik`, annevk already suggested that |
| 15:22 | <annevk> | variable, but with all the banner income you'll be rich! |
| 15:23 | <Lachy> | haha. That only works if people actually visit your site, and the best way to to do that is by domain squatting |
| 15:23 | <MikeSmith> | haha http://isbrucepartofhtml5.com/ is great |
| 15:24 | <variable> | annevk, the owner of com.com (IIRC its cnet) should be rich. Everyone else has to hope that people visit the websites. That is why all the domain parking sites take a monthly fee instead of a percentage |
| 15:26 | <Lachy> | anyone know who set up the bruce one? |
| 15:26 | <variable> | I don't get it. |
| 15:27 | <Philip`> | Lachy: View source |
| 15:27 | <Lachy> | from a comment in the source "All of these elements should really be <bruce>'s but those damn standards folks just won't let it happen! Give the man some love: @brucel" |
| 15:27 | <Lachy> | :-) |
| 15:27 | <MikeSmith> | zcorpan_: mathml schema has "token.content = mglyph|malignmark|text" |
| 15:28 | <variable> | who is bruce ? |
| 15:28 | <MikeSmith> | zcorpan_: "With the exception of the empty mspace element, token elements can contain any sequence of zero or more Unicode characters, or mglyph or malignmark elements." |
| 15:28 | <MikeSmith> | http://www.w3.org/TR/MathML3/chapter3.html#presm.tokenchars |
| 15:29 | <annevk> | variable, that sounds somewhat naive |
| 15:29 | <MikeSmith> | zcorpan_: so in standard mathml, those examples you found are not valid, fwiw |
| 15:30 | <annevk> | variable, I made quite a bit of money for a while, but I stopped doing advertising |
| 15:30 | <variable> | annevk, with content? or with regular parking? |
| 15:31 | <annevk> | variable, also, http://www.brucelawson.co.uk/ |
| 15:31 | <variable> | annevk, oh - I've been reading that blog for a while. I need to remember names better :\ |
| 15:33 | <MikeSmith> | zcorpan_: I don't know what to make of the one that has <xsl:text> .. but the standard mathml schema does not allow foreign-namespace content like that anywhere except with (effing) annotation-xml |
| 15:38 | <davidc> | mike: wasn't that a fragment of xslt generating mathml, rather than mathml? |
| 15:38 | <MikeSmith> | davidc! |
| 15:38 | <zcorpan_> | MikeSmith: ok |
| 15:38 | <zcorpan_> | i just filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=10047 |
| 15:39 | <MikeSmith> | davidc: you mean as part of an XSL stylesheet? |
| 15:39 | <MikeSmith> | I guess it must be |
| 15:39 | <MikeSmith> | if so, that'd certainly make a lot more sense |
| 15:39 | <zcorpan_> | MikeSmith: the xsl one was just a WTF (how many levels of escaping can you get for a single character?) |
| 15:39 | <davidc> | yes, although I missed the initial conext but <mo><xsl:text disable-output-escaping="yes"><![CDATA[⁢]]></xsl:text></mo> is xslt not mathml |
| 15:40 | <MikeSmith> | zcorpan_: ah |
| 15:41 | <MikeSmith> | davidc: yeah, I just misunderstood zcorpan_ reason for mentioning it |
| 15:42 | <MikeSmith> | davidc: the spec for annotation-xml is giving me heartburn |
| 15:43 | <MikeSmith> | I would be curious to hear details about how non-browser mathml UAs implement support for that |
| 15:43 | <jgraham> | Argh. How the f— did Foreign Content mode get so complex |
| 15:47 | <davidc> | MikeSmith: on the heartburn, sorry about that, but there must be something that I'm missing, really the only thing the browser needs to do is to correctly step past the content. Many MathML editing tools stick on all sorts of annotations (eg original source formats, or fallback display formats or whatever) and once we start telling people they can use mathml in html, it would be good if you... |
| 15:47 | <davidc> | ...could just drop these iannotated expressions in and have the annotations not mess things up. having more support is a bonus, but just ignoring the annotations for rendering is a good start. |
| 15:49 | <zcorpan_> | davidc: "there are essentially no real world cases of html in math in a text/html document as this (as you know) is all new." - this is not the case, there *are* text/html pages with bogus <math> elements in them (and expect it to not affect anything) |
| 15:49 | <zcorpan_> | davidc: which existed before the html5 parser had any of this |
| 15:49 | <davidc> | zcorpan_: |
| 15:51 | <davidc> | zcorpan_: yes but that's my point, if some weird pages misused some syntax for some other use, too bad. they could never have worked cross browser, they were always in error, how can you consider disabling mathml forever because of these pages? |
| 15:52 | <boblet> | oh dear. we appear to have started … something : http://ishtml5goingtokillflash.com/ |
| 15:52 | <zcorpan_> | davidc: why would they not work cross-browser? all browsers do pretty much the same for unknown tags |
| 15:54 | <davidc> | zcorpan: yes well it depends what's inside the elements, i was thinking you meant some pseudo-mathml, but i guess you mean <math><i>x</i>=3</math> or something? |
| 15:54 | <zcorpan_> | davidc: my question about html in annotation-xml was supposed to be read as xhtml in annotation-xml (mozilla has supported xhtml and mathml for a long time) |
| 15:54 | <zcorpan_> | davidc: yep. or just <math><!doctype html> (rest of page) |
| 15:55 | <davidc> | zcorpan: but if you found <video><i>x</i>=3</video> on some page somewhere would you have not added video? |
| 15:56 | <annevk> | XMLHttpRequest now has class=domintro boxes too |
| 15:56 | <annevk> | so you can quickly tell what each method does and whether it throws (and what) |
| 15:57 | <zcorpan_> | davidc: we would have called video something else, yes, if it was considered too much breakage |
| 15:58 | <davidc> | zcorpan: on xhtml in annotation-xm, I really don't know we have a lot of mathml here (at nag.co.uk) but i strip it down as much as I can so it works as cross browser as possible, other people are in same boat. putting mathml into text/html is supposed to make it easier, having undocumentable parsing rules for annotations rather works against that |
| 15:58 | <zcorpan_> | davidc: (for instance we renamed <input action> to <input formaction> for web compat that had bogus <input action>, and we're now considering renaming <input required> since pages have bogus <input required>) |
| 15:58 | <annevk> | davidb, undocumentable? |
| 15:59 | <davidb_> | heh too many davids |
| 15:59 | <annevk> | oops |
| 16:00 | <davidc> | zcorpan: yes I would guess that there's orders of magnitude more pages with <input... than <math in current text/html, and you were naming new features, not features specified a decade ago which are used by the main authoring tools |
| 16:00 | <zcorpan_> | davidc: i guess we could make <annotation-xml> kind of scoping so that it doesn't break out so easily |
| 16:01 | <MikeSmith> | yeah |
| 16:01 | <MikeSmith> | the problem is determining whether that is worth doing |
| 16:02 | <davidc> | MikeSmith: making me happy must be worth doing |
| 16:03 | <MikeSmith> | heh :) |
| 16:03 | <MikeSmith> | yeah, it's worth something |
| 16:03 | <zcorpan_> | davidc: there might be other ways to make you happy :) |
| 16:03 | <davidc> | zcorpan_: sadly not (speaks the xml geek at heart) |
| 16:04 | <MikeSmith> | davidc: about the spec and HTML content within MathML content: the current handling is such that a conformant parser does what I described in my message, except within token elements |
| 16:05 | <MikeSmith> | that is, it basically ends up moving the HTML content out of document order to follow the mathml content |
| 16:05 | <MikeSmith> | so it does not do that only for the annotation-xml case |
| 16:06 | jgraham | has not been following but |
| 16:06 | <davidc> | MikeSmith: well in the other cases the input is in error so I don't really care what fixup the html parser does., |
| 16:06 | <jgraham> | erm, no |
| 16:07 | <jgraham> | can someone give me an example of the problem markup |
| 16:07 | <MikeSmith> | davidc: we can make it an error in annotation-xml also, within HTML documents |
| 16:07 | <MikeSmith> | jgraham: <annotation-xml><img src=foo> |
| 16:08 | <MikeSmith> | jgraham: <mathml><semantics><mi></mi><annotation-xml><img src=foo> |
| 16:08 | <jgraham> | Oh, I thought that was fine |
| 16:08 | <davidc> | MikeSmith: You could except it will look a bit odd to say that annotation-xml can have as child any markup from any vocabulary except in html it cant have html. Wouldn't it? |
| 16:08 | <jgraham> | but I was being dumb |
| 16:09 | <jgraham> | we should just s/whose tag name is "svg"// in the spec |
| 16:09 | <jgraham> | I think that would solve the problem |
| 16:09 | <zcorpan_> | jgraham: you need to support <annotation-xml><somemathmlelement> also |
| 16:09 | <MikeSmith> | davidc: well, we could say that annotation-xml is not allowed at all in text/html documents |
| 16:09 | <MikeSmith> | jgraham: what zcorpan_ said |
| 16:10 | <MikeSmith> | that is the big complication |
| 16:10 | <jgraham> | Hmm? |
| 16:10 | <jgraham> | In what way support? |
| 16:10 | <MikeSmith> | that is what makes it different from the mathml token content elements |
| 16:10 | <MikeSmith> | *token elements |
| 16:10 | <MikeSmith> | the token elements can't contain mathml |
| 16:11 | <zcorpan_> | jgraham: <annotation-xml cd="mathmlkeys" name="contentequiv" encoding="MathML-Content"> <apply> |
| 16:11 | <davidc> | MikeSmith: yes but as I said somewhere if the mathml output from editors and CA systems etc is anotated then saying you can't use annotations in html means the editors have to change or users have to strip out the content or both |
| 16:11 | <MikeSmith> | jgraham: mathml within annotation-xml needs to be parsed as mathml, while html within mathml would need to be parsed as html |
| 16:12 | <MikeSmith> | davidc: well, somebody has to deal with costs here |
| 16:12 | <jgraham> | MikeSmith: Why? I thought it was fine to just ignore anything in annotation-xml? |
| 16:13 | <MikeSmith> | browser projects should not be expected to incur costs of implementing something with additional complexity unless the use cases for it are clear |
| 16:13 | <MikeSmith> | jgraham: it is display:none by *default* |
| 16:14 | <MikeSmith> | but what if an author sets its display property to something other than "none"? |
| 16:14 | <jgraham> | MikeSmith: It seems like we could just change the rules to put everything in annotation-xml in the html namespace. What use case requires something else? |
| 16:14 | <MikeSmith> | svg |
| 16:14 | <MikeSmith> | and mathml child content |
| 16:14 | <jgraham> | What does it mean if you do those things? |
| 16:14 | <jgraham> | What's the use case? |
| 16:15 | <davidc> | MikeSmith: yes but in my mental model just getting to the end of /annotation-xml without falling over is the simplest of options, not the most complicated, but perhaps 10 minutes on this irc isn't long enough to get my mental model aligned with the collective... |
| 16:15 | <MikeSmith> | jgraham: see http://golem.ph.utexas.edu/~distler/blog/archives/001475.html for some examples |
| 16:15 | <zcorpan_> | jgraham: you can annotate presentation-mathml with content-mathml in annotation-xml and vice versa |
| 16:15 | <MikeSmith> | jgraham: we are trying to determine what the use cases are |
| 16:16 | <MikeSmith> | davidc: it is not the simplest of options, due to the fact that the spec chooses to allow both mathml and non-mathml content in annotation-xml |
| 16:16 | <jgraham> | MikeSmith: I have seen Jacques example before, but if the svg is actually a representation of an operator or something, it is hardly an annotation |
| 16:17 | <jgraham> | isn't it possible to put svg in other places in MathML? |
| 16:17 | jgraham | guesses he should read the spec |
| 16:17 | <MikeSmith> | davidc: the reasons is because the parsing algorithm necessarily must handle HTML content within MathML in the way I've described |
| 16:17 | <davidc> | MikeSmith: well it was written for a simpler universe where every language used the same parsing model to avoid these problems. |
| 16:17 | <MikeSmith> | davidc: except for the special case of the token elements |
| 16:18 | <MikeSmith> | davidc: that is not the world we seem to find ourselves in |
| 16:18 | <davidc> | MikeSmith: oh? :-) |
| 16:18 | <jgraham> | zcorpan_: I think I would like to pretend that content mathml doesn't exist for the purpose of the web |
| 16:19 | <MikeSmith> | jgraham: it is not possible if you limit yourself to the standard/default MathML schema |
| 16:19 | <MikeSmith> | the standard schema only allows other namespaced content in annotation-xml |
| 16:19 | <jgraham> | Hmm |
| 16:19 | <MikeSmith> | that is why Jacques did it that way |
| 16:19 | <jgraham> | This seems like a lot of trouble |
| 16:20 | <MikeSmith> | but HTML5 parser allows it in the mathml "token elements" -- mi, mn, mn, mo, mtext |
| 16:20 | <MikeSmith> | so if Jacques were to use HTML5, he could/should just put it in mtext |
| 16:20 | <davidc> | jgraham: It's no trouble at all in the xhtml+mathml world for which it was designed, you can reliably parse past any content in annotation-xml, and by default you can always ignore it. |
| 16:20 | <jgraham> | I think I would prefer that approach in the long term |
| 16:20 | <MikeSmith> | what Jacques has in his examples is not even valid |
| 16:21 | <MikeSmith> | Jacques examples have <semantics><annotation-xml> (annotation-xml element as 1st child of semantics, which is not valid) |
| 16:22 | <davidc> | MikeSmith: if the use case is to make the svg appear in the math, putting in a token element rather than the annotation makes perfect sense |
| 16:22 | <MikeSmith> | Jacques it just relying there on some rendering behavior that just happens to work in Firefox but that is not really conformant to the MathML spec |
| 16:23 | <MikeSmith> | davidc: that is exactly what Jacques is using it for |
| 16:23 | <davidc> | If annotation-xml is proving difficult in text/html then just parsing it and leaving it in the dom unrendered would do, or if that's not possible, even just throwing the content away would do. |
| 16:23 | <jgraham> | MikeSmith: Based on what davidc is saying it seems like we really don't want to make annotation-xml insanely complex because it is mostly just for tool-specific junk that is not relevant to end users |
| 16:23 | <MikeSmith> | it is exactly what every page I have seen on the Web is using <annotation-xml> for |
| 16:23 | <MikeSmith> | davidc: understood |
| 16:23 | <davidc> | Yes but like the rest of us he tries to tread a line between what's specified and what works. He would change if what was specified also worked. |
| 16:25 | <MikeSmith> | davidc: I suggested earlier that we have UAs rely on the @encoding attribute of annotation-xml to address this issue |
| 16:25 | <MikeSmith> | but zcorpan_ pooh-poohed that for some reason |
| 16:25 | <MikeSmith> | I can't remember why |
| 16:26 | <davidc> | jgraham: well it might be relevant to end users, but it isn't relevant to the browser except for just being a data format. If I cut and paste an expression out of the browser, it would be nice (and works by the way in IE) if the annotations went along with it, then if I paste into say (maple0 it may well be able to evaluate the content mathml, even if teh presentation mathml is showing some... |
| 16:26 | <davidc> | ...strange notation only used in one textbook |
| 16:26 | <MikeSmith> | e.g, in order to have some annotation-xml content parsed as HTML, you'd need to do annotation-xml/@encoding=text/html |
| 16:26 | <zcorpan_> | MikeSmith: because it complicates the parser and we've generally tried to avoid inspecting attributes in the parser |
| 16:27 | <MikeSmith> | zcorpan_: OK |
| 16:27 | <MikeSmith> | ah yeah |
| 16:27 | <MikeSmith> | that makes sense, yeah |
| 16:28 | <MikeSmith> | davidc: what zcorpan_ said .. inspecting the values of particular attributes during DOM tree-building is something to be avoided |
| 16:28 | <davidc> | MikeSmith: agreed |
| 16:30 | <MikeSmith> | davidc: stepping back, let me make clear: as currently specced, everything essentially works as users might expect -- with the one exception of the case of HTML content in annotation-xml |
| 16:31 | <davidc> | All I really care about is that the parser 9using whatever parsing model is convenient for you to specify) gets as far as </annotation-xml> and then carries on the rendering as if it wasn't there. It would be nice if it stuffed the annotation somewhere usable by script but that isn't absolutely essential. |
| 16:31 | <davidc> | MikeSmith (how do you folks keep up with teh bandwidth on irc conversations:-) perhaps that is not as bad as it originally seemed |
| 16:32 | <MikeSmith> | davidc: so it will essentially to that for everything except HTML in annotation-xml |
| 16:32 | <davidc> | It would be good if there was some way to get an html annotation someone mentioned using a dov containere as a possibility? |
| 16:32 | <MikeSmith> | davidc: yeah, that is possibility |
| 16:32 | <MikeSmith> | but that is why I asked for use cases on www-html |
| 16:32 | <davidc> | modulo my spelling |
| 16:33 | <MikeSmith> | I can't personally help to make a compelling case to implementors for changing the spec unless I understand myself what the use cases are |
| 16:33 | <zcorpan_> | davidc: do you want <annotation> to similarly survive-but-not-render? |
| 16:34 | <davidc> | yes |
| 16:35 | <davidc> | zcorpan: basically (in my mental model of the universe) they should be treated pretty much as unknown attributes, carrying data but by default not doing anything |
| 16:35 | <jgraham> | We could treat annotation-xml as RCDATA or something. I wonder if that would break sites |
| 16:36 | <zcorpan_> | CDATA in that case |
| 16:37 | <davidc> | jgraham; I would parse the annotation using normal text parsing: If people put elements in there do whatever html fixup you find convenient, but don't go to trouble to make it "work" |
| 16:38 | <davidc> | jgraham: sorry scratch that you were talking about annotation-xml, I read it as ,annotation> sorry |
| 16:39 | <davidc> | apparently, i have a day job, got to go |
| 16:40 | <MikeSmith> | a possible use case for HTML in annotation-xml that I can think of have been but reluctant to bring up is audio content |
| 16:40 | <MikeSmith> | I don't know if the math wg has thought of that one yet |
| 16:40 | <MikeSmith> | but I am sure somebody else will eventually |
| 16:42 | <MikeSmith> | I would think that one possible use case for annotation-xml is to provide alternative representations for accessibility purposes |
| 16:42 | <MikeSmith> | e.g., a recording of an actual human being reading out a mathml expression |
| 16:43 | <jgraham> | MikeSmith: Hmm, that would work with the "just parse it all as HTML" model but not the CDATA model |
| 16:43 | <MikeSmith> | jgraham: yeah |
| 16:44 | <MikeSmith> | which is why I think the HTML case might be important |
| 16:44 | <MikeSmith> | but I'd rather not do their use-case documenting for them |
| 16:45 | <MikeSmith> | I don't know what the current accessibility story is with MathML |
| 16:45 | <MikeSmith> | but I would like to figure it out |
| 16:46 | <MikeSmith> | preemptively |
| 16:46 | <MikeSmith> | before it blossoms into a contentious Issue for the HTML WG |
| 16:46 | <MikeSmith> | maybe they already have all kinds of accessibility hooks in there, for all I know |
| 16:48 | <zcorpan_> | did Hixie publish a list of pages with bogus <math> and <svg> from his research? |
| 16:51 | <zcorpan_> | http://twitter.com/ibmkhd/statuses/17341570393 - disappointing his link is a 404 |
| 16:55 | <zcorpan_> | http://lazeroids.com/ |
| 16:55 | <zcorpan_> | get an error in firefox 'console not defined' |
| 16:56 | <jgraham> | http://www.bennadel.com/blog/1955-Pushing-Base64-Encoded-Images-Over-HTML5-WebSockets-With-Pusher-And-ColdFusion.htm |
| 16:56 | <zcorpan_> | thanks |
| 16:56 | <jgraham> | zcorpan_: Open firebug :) |
| 16:56 | <jgraham> | (for the "console not defined" thing) |
| 17:01 | <zcorpan_> | don't have firebug |
| 17:03 | <MikeSmith> | zcorpan_: he meant Web Inspector |
| 17:04 | <zcorpan_> | in opera i get TypeError: 'console.dir' is not a function |
| 17:05 | <jgraham> | zcorpan_: It does indeed work with firebug open |
| 17:05 | <jgraham> | report a bug? |
| 17:07 | <gsnedders> | http://iscss3partofhtml5.com/ |
| 17:07 | <zcorpan_> | reported |
| 17:07 | <zcorpan_> | gsnedders: old news |
| 17:07 | <gsnedders> | Meh. |
| 17:36 | <hsivonen> | whoa. a bug about the axis attribute! |
| 17:38 | <jgraham> | oh no not again |
| 17:39 | <zcorpan_> | where? |
| 17:39 | <jgraham> | or do you mean a Gecko bug? |
| 17:42 | <hsivonen> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=10045 |
| 17:46 | <jgraham> | Seems easy to close: "no axis is not going to be supported" |
| 17:50 | <Rik`> | video::before {content : "a"} should work, right ? |
| 18:00 | <zcorpan_> | Rik`: seems to work in opera (same with img) |
| 18:11 | <Rik`> | zcorpan_: well, Opera was the only one I haven't tested |
| 18:37 | <TabAtkins> | Oh wow, that lazeroids game *really* needs to put a cap on how fast your ship can go. |
| 18:38 | <TabAtkins> | It's too easy to just jam on the up key and quickly be moving fast enough that you advance past each screen before it can scroll to show you. |
| 18:43 | <zcorpan_> | TabAtkins: file a bug :) |
| 18:45 | <TabAtkins> | zcorpan_: Too late, already bored! |
| 18:47 | <TabAtkins> | Also, heh, hitting a bullet with another bullet destroys both for 0 points. |
| 18:49 | <paul_irish> | hsivonen: yahoo wanted to distribute modernizer as a YUI module and apparently needed it licensed BSD to satify something something. |
| 18:50 | <paul_irish> | it didn't make much sense to me. |
| 18:51 | <AryehGregor> | TabAtkins, it's physically unrealistic for spaceships to be unable to accelerate without limit. |
| 18:51 | <AryehGregor> | (although once they reach relativistic velocities, they'll start seeing diminishing returns from the same fuel investment) |
| 18:52 | <TabAtkins> | AryehGregor: In space, constant thrust = constant velocity. TV teaches us this. |
| 18:52 | <AryehGregor> | I said "physically unrealistic", not "televisionally unrealistic". |
| 18:53 | <TabAtkins> | Computers are like TVs, so clearly they should behave similarly. |
| 18:53 | <Philip`> | AryehGregor: Acceleration requires fuel, and fuel is limited |
| 18:53 | <AryehGregor> | Whoa, massively-multiplayer Asteroids? |
| 18:53 | <AryehGregor> | Philip`, yes, but that's not the constraint we're discussing here. |
| 18:54 | <AryehGregor> | If your fuel ran out and your acceleration was thereby limited, that would be different. |
| 18:54 | <AryehGregor> | That would be entirely realistic, since fuel is likely to be a serious limiting factor even for advanced spacecraft. |
| 18:54 | <AryehGregor> | Depending on how fast you want to travel. |
| 18:54 | <Philip`> | Perhaps you could say the spaceship has very limited fuel(/energy) storage (to save mass) and so it has to generate energy as it's required, and it can use that energy for either acceleration or shielding against micrometeorites |
| 18:54 | <TabAtkins> | In any case, I'm talking game-fun. If you go fast enough, you literally hover off the edge of the screen. |
| 18:55 | <Philip`> | and you're in a region of space where there's a lot of very slow-moving micrometeorites |
| 18:55 | <Philip`> | so if you're stationary relative to them, then you don't need much shielding and can accelerate fast |
| 18:55 | <TabAtkins> | You're just reintroducing drag with more technobabble. |
| 18:55 | <Philip`> | but as you move faster you have to slow down acceleration, until you reach a top speed where all your energy is shielding you from collisions |
| 18:55 | <TabAtkins> | Just say that space is filled with luminiferous ether and be done with it. |
| 18:56 | <Philip`> | It's not drag since you're not transferring momentum to your surroundings |
| 18:56 | <TabAtkins> | The effect is the same - as you go faster, an outside force gradually saps more of your acceleration. |
| 18:57 | <AryehGregor> | Philip`, your technobabble is fairly impressive. |
| 18:57 | <AryehGregor> | Now just tell me what the heck these "shields" are that can deflect micrometeorites if you put energy into them. |
| 18:57 | <AryehGregor> | (also, we aren't talking about light, what does the luminiferous ether have to do with it?) |
| 18:57 | <Philip`> | TabAtkins: Sure - the point is to try to have a more physically plausible explanation for the same effects, so you don't have to implement stupid impossible gameplay |
| 18:58 | <TabAtkins> | AryehGregor: The ether also produces a drag on mass, clearly. |
| 18:58 | <Philip`> | (Realistically you'd attack a planet by accelerating an asteroid into an orbit that will collide with it a few years later, but it's much more fun to have an explanation for why you'd want to build a Death Star) |
| 18:59 | <TabAtkins> | AryehGregor: The shields are magnetically-contained plasma envelope that vaporize micrometeorites before they touch the hull. |
| 18:59 | <AryehGregor> | Philip`, then your target could decelerate it. They'd be able to spot you accelerating such a large object that much from billions of miles away. |
| 19:00 | <Philip`> | AryehGregor: The 'shields' are just lasers that vapourise micrometeorites, so that the collision energy is dispersed and easily handled by the hull |
| 19:00 | <AryehGregor> | TabAtkins, if they're magnetically contained, why do you need any energy input? Also, vaporizing them doesn't get rid of the kinetic energy. |
| 19:00 | <TabAtkins> | AryehGregor: You start the acceleration out pretty far, preferably off the ecliptic as well so it's harder to notice. |
| 19:00 | <AryehGregor> | Philip`, okay, I'm beginning to buy this. |
| 19:01 | <AryehGregor> | TabAtkins, the further away you accelerate it from, the longer it takes to get there. Unless you accelerate it more, but then that cancels out the greater distance, so it doesn't help you. |
| 19:01 | <TabAtkins> | AryehGregor: Electromagnets. Also, vaporizing them allows the plasma to absorb their momentum, thus spreading the impact over the entire magnetic field rather than focused on a point. |
| 19:01 | <Philip`> | AryehGregor: You find a rock that's going to go close to the sun, secretly attach some rockets to it, then activate them when it's got close so the light from the sun masks the rocket activity |
| 19:01 | <TabAtkins> | AryehGregor: The point is that the further you start it out, the bigger head start you have. |
| 19:01 | <AryehGregor> | Electromagnets don't consume energy to stay on, if you're smart and use a superconductor. |
| 19:02 | <AryehGregor> | Philip`, but it will be clearly visible from a different angle. |
| 19:02 | <TabAtkins> | AryehGregor: These people aren't smart enough to realize that the ether isn't real, you think they'd use a superconductor? |
| 19:02 | <Philip`> | or you just stick a giant screen around the rock before you boost it so that nobody sees |
| 19:02 | <AryehGregor> | TabAtkins, well, since space is only 4 K anyway, it seems like a pretty obvious idea. |
| 19:02 | <AryehGregor> | Philip`, the heat will leak out through the screen, you can't prevent that. |
| 19:02 | <Philip`> | AryehGregor: Space is big and you can't watch all of it all the time from all directions with highly accurate instruments |
| 19:03 | <TabAtkins> | Thus why I suggested launching it from out of the ecliptic. |
| 19:03 | <AryehGregor> | You don't need highly accurate instruments to spot that much acceleration being applied. The thermal radiation would be greater than your typical star. |
| 19:03 | <TabAtkins> | Nobody looks out there. |
| 19:03 | AryehGregor | goes to look up his website |
| 19:03 | <Philip`> | and there's stars and rocks and ice and supernovas and spaceships everywhere, so it's easy to hide |
| 19:03 | <Philip`> | AryehGregor: You apply the acceleration very very slowly |
| 19:04 | <TabAtkins> | True - you need very little acceleration to make it work. |
| 19:04 | <AryehGregor> | http://www.projectrho.com/rocket/index.html |
| 19:04 | <TabAtkins> | Hell, use a magdriver rather than an engine. |
| 19:05 | <TabAtkins> | A couple of robots, a railgun attached to a nuclear pile throwing off mass from the rock, the whole thing burrowed in so the rock itself acts as thermal shielding... |
| 19:05 | <AryehGregor> | http://www.projectrho.com/rocket/rocket3w.html#nostealth |
| 19:05 | <TabAtkins> | No stealth for rocket engines, sure. Those things are crazy. |
| 19:05 | <AryehGregor> | The rock won't act as thermal shielding over long periods of time. The heat has to escape somehow. |
| 19:05 | <AryehGregor> | No, no stealth for anything reasonably warm. |
| 19:06 | <TabAtkins> | Hm, true. Probably better to just put it on the surface and keep it low-power so it can radiate as much as possible while facing away from detection. |
| 19:06 | <Philip`> | It doesn't need to work for long periods of time, just long enough that your target doesn't have time to reach it and accelerate it in a different direction |
| 19:06 | <AryehGregor> | Anything that's, say, human body temperature will be visible from insane distances. Although maybe if it were a robot, with a very cool engine . . . how cool can engines get? |
| 19:06 | <TabAtkins> | Very cool? |
| 19:06 | <AryehGregor> | You've inevitably got thermodynamic inefficiencies here, no? |
| 19:06 | <AryehGregor> | Railguns get hot. |
| 19:07 | <Philip`> | and you can stick a load of laser defence towers on the rock to discourage your target's population from trying to touch it |
| 19:07 | <TabAtkins> | Fire it once every week, or whatever you need to dissipate heat sufficiently. |
| 19:07 | <AryehGregor> | "even a puny ship using ion drive to thrust at a measly 1/1000 of a g could be spotted at one astronomical unit." |
| 19:07 | <TabAtkins> | 1/1000 of a g? Such luxury! |
| 19:07 | <AryehGregor> | Philip`, then they accelerate a rock to blow up your rock from out of laser range? |
| 19:07 | <hsivonen> | paul_irish: I see. |
| 19:07 | <AryehGregor> | TabAtkins, then it will never actually get anywhere until the people you're annoyed at have been dead for millennia anyway . . . |
| 19:08 | <TabAtkins> | When you're planning to kill a planet, a century or ten here or there isn't that important. |
| 19:08 | <Philip`> | AryehGregor: It's much harder for them to hit your little rock than for your little rock to hit their giant planet, particularly when you have years longer to prepare for it |
| 19:08 | <TabAtkins> | Philip`: To be fair, planets are pretty freaking tiny. |
| 19:08 | <AryehGregor> | Do we need to start doing back-of-the-envelope calculations now? |
| 19:09 | <TabAtkins> | No. |
| 19:09 | <AryehGregor> | Okay. |
| 19:09 | <AryehGregor> | Good. |
| 19:09 | <TabAtkins> | I'll just pretend that warp drives are possible and the whole thing is done in hyperspace. |
| 19:11 | <AryehGregor> | That's one approach. |
| 19:11 | <AryehGregor> | Rather the most common one, I'd think. |
| 19:12 | <zcorpan_> | ok that's the last time i link to a space ship game |
| 19:13 | <TabAtkins> | Other solutions involve doing all the acceleration in a nearby star system so you *can* crank it up nice and fast, and then just waiting a century or two for it to hit at .01c or something. |
| 19:13 | <AryehGregor> | TabAtkins, would be clearly visible. |
| 19:14 | <AryehGregor> | If you accelerate it more, at greater distance so that it hits in the same time, they cancel out exactly. |
| 19:14 | <AryehGregor> | My link said something like that. |
| 19:15 | <TabAtkins> | Just use an engine with power low enough to be hidden in the star's signature. Or even start it *behind* the star, timing things so that it just passes near the star on the way to the target. |
| 19:15 | <AryehGregor> | Then all the locals will see it, and spies will report back. Anyone on the other side of the star. |
| 19:16 | <TabAtkins> | That's fine. |
| 19:16 | <AryehGregor> | The general strategy is clearly plausible, but the details would need a lot of thought. |
| 19:16 | <TabAtkins> | Once it gets going, it's going to be hard to stop. |
| 19:16 | <AryehGregor> | Very easy to deflect it enough to miss, if you're sending it from light-years away. |
| 19:16 | <TabAtkins> | Most plausible "asteroid deflection" plans involve putting something on the asteroid and accelerating it to the side. |
| 19:16 | <AryehGregor> | You only need to deflect it a minute amount. |
| 19:16 | <AryehGregor> | Yep. |
| 19:17 | <TabAtkins> | If it's already going quite fast, you'd need to match velocities to land on it, which may be difficult. |
| 19:17 | <AryehGregor> | So it's an attack that's far easier to deflect than to conduct. |
| 19:17 | <Philip`> | TabAtkins: It's no good if your attack takes so long that the political situation has changed and you're no longer enemies with the people you're sending an unstoppable asteroid to |
| 19:17 | <AryehGregor> | Still much, much easier than accelerating the thing to start with. |
| 19:17 | <TabAtkins> | Philip`: It's good enough if you're sure that your enemies are infidels whos descendants will be forever hated by your alien god. |
| 19:18 | <AryehGregor> | You don't need to match velocities, either. |
| 19:18 | <AryehGregor> | Get a few really big nukes, that's enough. |
| 19:18 | <AryehGregor> | Especially at a large distance. |
| 19:18 | <TabAtkins> | Well, maybe. You dont' want to just fracture it. |
| 19:18 | <TabAtkins> | Multiple small nukes is better. |
| 19:18 | <AryehGregor> | Possibly. |
| 19:18 | <AryehGregor> | Also less chance of failure. |
| 19:19 | <AryehGregor> | This attack would be more plausible if you found an asteroid that was passing very close to the planet anyway, and just deflected it slightly. |
| 19:19 | <AryehGregor> | But then your enemy could do the same. They'd want to know about random asteroids passing nearby anyway. |
| 19:20 | <AryehGregor> | So even if they don't know you did it, their asteroid defense team would notice it and go to stop it. |
| 19:20 | <AryehGregor> | 0.01c is not very fast, plenty of time to spot it if it's large enough to blow up the world at that speed. |
| 19:20 | <TabAtkins> | Assuming you haven't already blown up their asteroid defense, and aren't just dropping a rock on them to finish them off. |
| 19:21 | <AryehGregor> | If you're just finishing them off, you don't need to get fancy. Nuke 'em from orbit. |
| 19:21 | <TabAtkins> | I assume putting a cheap thruster on a nearby asteroid is less expensive than that. |
| 19:22 | <AryehGregor> | I really doubt that. |
| 19:22 | <TabAtkins> | I'd assume that "drop a big rock on them" is the least fancy you can get. ^_^ |
| 19:22 | <AryehGregor> | Nukes are not very expensive at all. |
| 19:22 | <AryehGregor> | (relative to space travel) |
| 19:22 | Philip` | notes that when he last read about the asteroid idea, the aim wasn't to hit the target but to pass by close enough to pull it out of its stable Lagrange point orbit |
| 19:22 | <AryehGregor> | Also much faster. |
| 19:22 | <TabAtkins> | Nukes are expensive when you have to carry them in space, though. |
| 19:22 | <AryehGregor> | Maybe. It depends. |
| 19:23 | <TabAtkins> | Just use a couple of those nukes to deflect a rock rather than decimate a single city, and the rest of the work is done by Newton himself. |
| 19:26 | <AryehGregor> | After a very long time. |
| 19:27 | <TabAtkins> | That might be enough. Bashing apart their industrial capacity is easy, and won't take much actual destructive power. The rock is for sterilizing purposes. |
| 19:28 | AryehGregor | notes that he happened to be reading: "It is also true that if the attacker's objective is sheer devastation, nuclear bombardment from high orbit will do the job. (So will the ever popular asteroid toss, but this is just a Rube Goldberg way to get the same effect.)" http://www.rocketpunk-manifesto.com/2009/06/space-warfare-i-gravity-well.html |
| 19:28 | <AryehGregor> | So it's two against one now. You're outvoted. |
| 19:28 | <TabAtkins> | Who cares if it takes 50 years for it to hit the planet, if they can't scrounge up enough industrial capacity in that time to do anything about it, and so you can immediately leave for the next planet in your life-killing spree. |
| 19:29 | <TabAtkins> | Rube Goldberg devices can sometimes be more energy-efficient. ^_^] |
| 19:29 | <zcorpan_> | AryehGregor: i thought whatwg didn't do voting |
| 19:29 | <AryehGregor> | zcorpan_, no, but we're discussing physics, and physical reality is decided by vote. Didn't you know? |
| 19:29 | <Philip`> | In Spore I usually just terraform the planet until all life is extinguished, and then terraform it back again |
| 19:30 | <Philip`> | and then paint the seas purple |
| 19:30 | <TabAtkins> | Insult to injury, I see. |
| 19:46 | <AryehGregor> | Am I the only one who would gladly take a cell phone that's three times as thick and heavy if it means twice the battery life? |
| 19:46 | <AryehGregor> | (since I clearly haven't spammed enough off-topic conversation here today) |
| 19:46 | <Philip`> | The markets suggests you are |
| 19:46 | <Philip`> | *market |
| 19:47 | <AryehGregor> | Maybe that's because you can immediately tell how thin and light a device is, but you don't know how long its battery lasts until you've actually bought it. |
| 19:47 | <AryehGregor> | First impressions and all. |
| 19:47 | <AryehGregor> | (since obviously no one in their right mind believes manufacturers' battery life claims) |
| 19:48 | <Philip`> | Do people buy phones without reading any independent reviews? |
| 19:48 | <Philip`> | and/or do reviews not discuss battery life? |
| 19:48 | <AryehGregor> | Yes in both cases. :) |
| 19:49 | Philip` | would be happy with any battery life longer than the time between waking and sleeping, since he can recharge it overnight |
| 19:49 | <AryehGregor> | I could do that, but isn't it supposed to be bad for the battery if you recharge it before it gets low? |
| 19:50 | <Philip`> | (My phone seems to last at least several days before the battery indicator goes non-green, anyway) |
| 19:51 | <Philip`> | I don't think it's bad enough that you'll notice, particularly if you buy a new phone every couple of years, and if it's a problem you can just buy a new battery |
| 19:51 | <AryehGregor> | Reasonable points. |
| 19:51 | <AryehGregor> | What phone do you have? |
| 19:52 | <Philip`> | It appears to be a K530i |
| 19:52 | <zcorpan_> | aren't phone batteries pretty expensive? |
| 19:52 | <AryehGregor> | See, I got such a fancy phone (Nexus One) that its battery barely lasts 24 hours even if I don't actually use it. |
| 19:53 | <AryehGregor> | Google suggests $49.99 new. |
| 19:54 | Philip` | has never seen any references to any actual experiments on battery life and the effects of recharging, just a load of (often contradictory) apocryphal claims about the dangers of charging too much or too little or too fast or too slow |
| 19:54 | <Philip`> | (though maybe I just haven't looked in the right places) |
| 19:54 | <AryehGregor> | Philip`, well, your phone weighs significantly less than mine but its battery lasts longer, so there goes my hypothesis. It's also shorter and narrower, also a few millimeters thicker. |
| 19:55 | <Philip`> | AryehGregor: Your problem is that you have a phone with features |
| 19:55 | <AryehGregor> | But it doesn't last as long as yours even when it's *off*. |
| 19:56 | <Hixie> | ok i've finished describing the websrt format authoring rules |
| 19:56 | <Hixie> | now i just have the parser to finish, and the renderer, and the DOM API, and the examples, and the.... |
| 19:57 | <Philip`> | Then I guess you merely need to convince the world to adopt the format |
| 20:02 | <Hixie> | details |
| 20:02 | <zcorpan_> | Hixie: surely the cue identifier can't contain linebreaks? |
| 20:02 | <zcorpan_> | A WebSRT cue identifier is any sequence of one or more characters not containing the substring "-->" (U+002D HYPHEN-MINUS, U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN). |
| 20:03 | <Hixie> | good catch |
| 20:04 | <zcorpan_> | also, shouldn't there be a linebreak separating the cue identifier and the cue timings? |
| 20:09 | <jgraham> | AryehGregor: I belive with lithium ion batteries you are supposed to recharge little and often rather than running them right down |
| 20:09 | <jgraham> | That was some other type of battery |
| 20:09 | <jgraham> | Hixie: You are working on the parser now? Awesome :) |
| 20:10 | <Hixie> | been working on the websrt parser for a while now |
| 20:10 | <Hixie> | it's mostly done except for the cue text |
| 20:10 | <Hixie> | (which of course is the hard part) |
| 20:10 | <jgraham> | (I meant the HTML parser) |
| 20:10 | <AryehGregor> | jgraham, Wikipedia says they have no memory effect, and that sounds like what I was remembering, so you're probably right. Thanks. |
| 20:10 | <jgraham> | (but I misunderstood you) |
| 20:10 | <jgraham> | (so ignore me) |
| 20:13 | <gsnedders> | (I often ignore him, it seems quite effective) |
| 20:13 | <jgraham> | (I know) |
| 20:13 | <gsnedders> | (Then he throws things at me, or gets people to throw things at me) |
| 20:13 | <gsnedders> | (that kinda sucks) |
| 20:14 | <jgraham> | (that hardly ever happens) |
| 20:15 | <jgraham> | (mostly other people throw thingat you of their own volition) |
| 20:18 | <TabAtkins> | (I'm whispering in parentheses) |
| 20:19 | <jgraham> | (I'm shouting in parenthesis) |
| 20:19 | <jgraham> | (but you can' tell) |
| 20:19 | <TabAtkins> | (NO YOU'RE NOT, I AM) |
| 20:19 | <jgraham> | (*can't) |
| 20:23 | <dglazkov> | () |
| 20:23 | <jgraham> | Hmm surely we have tests for the £ vs &prod thing |
| 20:24 | <gsnedders> | That's what you think. |
| 20:24 | <jgraham> | I would think that. I wonder if I would be right |
| 20:24 | <jgraham> | I also wonder if webkit are using those tests |
| 20:24 | <jgraham> | assuming they exit |
| 20:24 | <jgraham> | *exist |
| 20:25 | <zcorpan_> | + <p>Each <span>WebSRT timestamp</span> in a <span>WebSRT file</span> |
| 20:25 | <zcorpan_> | + must represent a time greater than than all the times represented by |
| 20:25 | <zcorpan_> | + <span title="WebSRT timestamp">WebSRT timestamps</span> earlier in |
| 20:25 | <zcorpan_> | + the file.</p> |
| 20:25 | <zcorpan_> | Hixie: i thought we wanted to support overlapping cues? |
| 20:26 | <Hixie> | oh, right |
| 20:26 | <Hixie> | good point |
| 20:26 | <Hixie> | hmm |
| 20:26 | <Hixie> | i need a way to say that the karaoke timestamps have to be in order and in the timeframe of the cue |
| 20:26 | <Hixie> | and that the end stamp has to be later than the start stamp |
| 20:26 | <zcorpan_> | start > earlier starts; end > current start |
| 20:26 | <variable> | Hixie, do you recall any bugs or mailing list statments where you said that we won't create a advert tag..... |
| 20:26 | <Hixie> | and that the start stamp has to be later than the previous start stamp |
| 20:26 | <TabAtkins> | variable, I sent you links yesterday about that |
| 20:27 | <jgraham> | variable: TabAtkins answered that yeserday, no? |
| 20:27 | <jgraham> | heh |
| 20:27 | <variable> | I recall TabAtkins saying he will but I don't recall seeing the links |
| 20:27 | variable | greps logs |
| 20:27 | <Hixie> | variable: not off-hand, but i can tell you the rationale is just that such an element wouldn't work, since as soon as people used it, other people would use it to remove their ads, and then the people would stop using it |
| 20:27 | <variable> | Hixie, yep - I just want to add a source ;) |
| 20:28 | <TabAtkins> | One sec, I'm looking up the line in the logs. |
| 20:28 | <TabAtkins> | http://krijnhoetmer.nl/irc-logs/whatwg/20100628#l-1013 |
| 20:28 | <Hixie> | :-) |
| 20:28 | <variable> | TabAtkins, thanks. I don't see that in my logs :\ |
| 20:29 | <variable> | % grep -R "TabAtkins" \#whatwg.log|grep "two me"|wc -l |
| 20:29 | <variable> | 0 |
| 20:29 | <variable> | :\ |
| 20:30 | <variable> | but thanks for the link |
| 20:31 | <annevk> | zcorpan_, I filed a bug on that missing line break separator |
| 20:32 | jgraham | wonders if Julian is just expecting us to return the originally quoted text, verbatim |
| 20:33 | <annevk> | what do you mean? |
| 20:34 | <Hixie> | can julian not read english? |
| 20:36 | <annevk> | oh, I didn't realize he had a reply in that thread |
| 20:37 | <annevk> | guess it makes sense for his anti-non-BNF campaign |
| 20:38 | <annevk> | he missed out on a "for the record" note there |
| 20:38 | <jgraham> | Is anti-non- the same as pro-? |
| 20:39 | <Hixie> | julian isn't pro anything |
| 20:39 | <jgraham> | It doesn't really make sense because BNF is even less like English than English is |
| 20:39 | <jgraham> | Most things are in fact |
| 20:39 | <annevk> | he never really appears pro, he just complaints |
| 20:39 | <AryehGregor> | jgraham, an anti-non-BNF person might accept having no grammar at all, since that wouldn't include anything non-BNF. |
| 20:40 | <AryehGregor> | Well, why should he bother endorsing the parts of the spec he agrees with? They're unlikely to change unless someone happens to object. |
| 20:41 | <AryehGregor> | (I guess he could oppose someone else's objection, and thereby be pro- something, but then he'd be anti-the objection) |
| 20:41 | <jgraham> | AryehGregor: Bits of the spec I agree with kep getting changed |
| 20:41 | <jgraham> | Although that typically happens via people objecting |
| 20:42 | <jgraham> | often with the word "Formally" dangling just out of sight |
| 20:42 | <annevk> | saying formally is just a formality |
| 20:43 | <jgraham> | So, as a distraction from one of the several useful things I should be doing, does anyone know how one deals with random numbers in Haskell in a sane way |
| 20:44 | <TabAtkins> | "sane" meaning "without monads"? |
| 20:44 | <annevk> | I was planning on replacing cookies with localStorage on html5.org earlier today, but I guess I'll leave that for tomorrow |
| 20:44 | <jgraham> | TabAtkins: I think sane might mean "with monads but without making the whole program non-pure" |
| 20:44 | <Hixie> | AryehGregor: he doesn't seem to endorse anything, he just says things are bad and doesn't say what he'd prefer |
| 20:44 | <TabAtkins> | jgraham: http://en.wikibooks.org/wiki/Haskell/Hierarchical_libraries/Randoms#The_Standard_Random_Number_Generator |
| 20:45 | <TabAtkins> | That not good enough for you? |
| 20:45 | <jgraham> | TabAtkins: I am a total n00b though, this is just a learning project |
| 20:45 | <TabAtkins> | Essentially doing random numbers as a sequence. |
| 20:45 | <AryehGregor> | He wrote the pro-multiple-languages-in-Content-Language Change Proposal. That counts as saying what he'd prefer. |
| 20:45 | <annevk> | in theory it should simplify matters, and reduce traffic I suppose |
| 20:45 | <hsivonen> | speaking of objections, the objector turnout is pretty low even though the polls close tomorrow |
| 20:45 | <TabAtkins> | I did my objecting. |
| 20:46 | <jgraham> | The polls close tomorrow? |
| 20:46 | <jgraham> | Oh |
| 20:48 | <zcorpan_> | Hixie: have you seen http://www.storiesinflight.com/js_videosub/ ? |
| 20:48 | <AryehGregor> | Apparently some people don't understand what "objections" means. Two of the five people objecting to making the Content-Language pragma non-conforming actually support it. |
| 20:48 | hsivonen | wonders if metacity and compiz allow apps to draw in the title bar while the WM draws the buttons |
| 20:48 | <hsivonen> | as opposed to getting rid of the title bar altogether as in Chrome and Wine |
| 20:49 | <hsivonen> | or if apps can set a per-window synthetic theme |
| 20:52 | <Philip`> | jgraham: I'm sure we have tokeniser tests for character references, but they're probably not trivial to run in a browser-hosted parser implementation |
| 20:52 | <jgraham> | Philip`: Good point |
| 20:52 | <Philip`> | and I wouldn't be surprised if the html5lib tree constructor tests didn't attempt to test that part of the tokeniser |
| 20:52 | <jgraham> | Philip`: I am pretty sure they don't |
| 20:54 | <hsivonen> | Philip`: Jonathan Griffin made a harness that turned tokenizer tests into tree builder tests, but the harness hasn't landed and is rotting as bugzilla attachment :-( |
| 20:54 | <jgraham> | hsivonen: Bug number? |
| 20:54 | <jgraham> | I might resurrect that |
| 20:55 | <hsivonen> | jgraham: https://bugzilla.mozilla.org/show_bug.cgi?id=html5-parsing |
| 20:55 | <hsivonen> | jgraham: you'll need a freash build of the JS version of the parser |
| 20:56 | <hsivonen> | *fresh |
| 20:56 | <jgraham> | hsivonen: Thanks |
| 20:56 | hsivonen | tries to remember to recompile it tomorrow |
| 21:02 | <annevk> | Hixie, http://www.w3.org/Bugs/Public/show_bug.cgi?id=10024 is another WebSRT syntax issue |
| 21:03 | <zcorpan_> | annevk: wasn't it fixed in the last checkin? |
| 21:11 | <hober> | hsivonen: by "the JS version of the parser," do you mean http://github.com/aredridel/html5 ? |
| 21:12 | <gsnedders> | hober: I presume he means the GWT compiled version of the Java one |
| 21:12 | <gsnedders> | (Instead of the C++ compiled version of the Java one) |
| 21:12 | <gsnedders> | (Or the Java one itself) |
| 21:12 | <hober> | ahh, indeed. |
| 21:14 | <gsnedders> | (Because all ES QA love GWT) |
| 21:15 | <Philip`> | gsnedders: You should consider the ability to view the original source code of everything you're debugging as a privilege, not a right |
| 21:28 | <hsivonen> | hober: I mean the parser from http://livedom.validator.nu/ |
| 21:29 | <TabAtkins> | Man, fuck clearance. Just defining it as "partially-collapsed margins" would have been simpler. |
| 21:37 | <zcorpan_> | Hixie: "WebSRT cue text line" is a confusing term since it's not actually a line |
| 21:38 | <annevk> | zcorpan_, oh, missed that first change |
| 21:39 | <zcorpan_> | Hixie: maybe "WebSRT cue markup" |
| 21:39 | <TabAtkins> | cue item? |
| 21:40 | <zcorpan_> | is "&" an item? |
| 21:41 | <annevk> | cue body? |
| 21:41 | <TabAtkins> | Ooh, I like that one. |
| 21:41 | <zcorpan_> | seems weird to nest bodies :) |
| 21:41 | <TabAtkins> | "A WebSRT cue body contains one or more of the following components, in any order:" |
| 21:44 | <annevk> | zcorpan_, well, you'd have to change the other bits to match I suppose |
| 21:45 | <zcorpan_> | annevk: how would you change http://www.whatwg.org/specs/web-apps/current-work/complete/video.html#websrt-cue-italics-span ? |
| 21:45 | <TabAtkins> | Actually, I think I'd be okay with it saying that you nest cue bodies. |
| 21:48 | <annevk> | zcorpan_, so how is it not a line? |
| 21:49 | <zcorpan_> | annevk: a line to me implies a line break after it |
| 21:49 | <annevk> | zcorpan_, the way it is defined is that you have a lines separated by line terminators; makes sense to me |
| 21:49 | <annevk> | zcorpan_, and line terminators cannot be part of the line |
| 21:50 | <zcorpan_> | annevk: see <i>, <b> and <ruby> |
| 21:51 | <annevk> | zcorpan_, you mean that it refers back to cue text line? |
| 21:51 | <zcorpan_> | yeah |
| 21:51 | <TabAtkins> | lines embedded in lines without line separators involved does sound kinda weird. |
| 21:52 | <annevk> | fair enough |
| 21:52 | <annevk> | I wonder about the parsing details |
| 21:53 | <annevk> | and also why we do not allow all entities |
| 21:54 | <annevk> | for instance, VLC supports © |
| 21:56 | <annevk> | also ™ etc. |
| 21:56 | <annevk> | GStreamer doesn't by default though |
| 21:56 | <annevk> | but GStreamer is pretty crappy with text/srt in general |
| 22:11 | <zcorpan_> | i wonder if it's possible to download a whole bunch of srt subtitles from somewhere for research puposes |
| 22:12 | <zcorpan_> | other than crawling a site and downloading them one-by-one |
| 22:24 | <zcorpan_> | yay lazeroids works in opera now :) |
| 22:24 | <zcorpan_> | (and firefox without firebug) |
| 22:27 | <zcorpan_> | except it doesn't seem to work so good in opera :/ |
| 22:29 | <AryehGregor> | Lazeroids is buggy for me in Chrome. Movement doesn't work right. |
| 22:29 | AryehGregor | tries again |
| 22:30 | <AryehGregor> | Also, I thought you said you'd |
| 22:30 | <AryehGregor> | Also, I thought you said you wouldn't mention space ship games here again. |
| 22:33 | <AryehGregor> | I think the problem is that holding down arrow keys doesn't work for moving. |
| 22:33 | <AryehGregor> | Or something. |
| 22:34 | <zcorpan_> | well i lied |
| 22:36 | <annevk> | haha -- http://twitter.com/jreschke/status/17364260158 |
| 22:37 | <annevk> | that email did only make sense in context of the rambling earlier |
| 22:45 | <Philip`> | On the subject of test result formats: I used YAML for canvas tests, which is nice for humans to write and hard for computers to parse |
| 22:45 | <Philip`> | which is completely not what's needed for test results |
| 22:45 | <Philip`> | so don't use that |
| 22:47 | <AryehGregor> | How about JSON? |
| 22:49 | <Philip`> | That was suggested already |
| 22:49 | <Philip`> | (according to the minutes) |
| 22:59 | <jgraham> | I suggested JSON |
| 23:00 | <jgraham> | Since XML is massive amounts of overkill and json is only small amounts of overkill |
| 23:01 | <jgraham> | Oh I just thought of another issue that wasn't discussed |
| 23:01 | <jgraham> | We need a way to deal with multiple tests per file |
| 23:01 | <jgraham> | Hmm |
| 23:02 | <jgraham> | Oh well, bedtime |
| 23:13 | <Hixie> | zcorpan_: sweet |
| 23:13 | <Hixie> | i haven't even half finished the spec yet and we're already got implementations |
| 23:14 | <TabAtkins> | Haha, I just made several people look through multiple rulebooks before they realized I was making a rule-34 joke. |
| 23:15 | <TabAtkins> | http://forums.xkcd.com/viewtopic.php?p=2215558#p2215558 |
| 23:16 | TabAtkins | is a dork. |
| 23:19 | <Philip`> | zcorpan_: http://urd.let.rug.nl/tiedeman/OPUS/OpenSubtitles.php ? |
| 23:20 | <zcorpan_> | Philip`: thanks |
| 23:21 | <Philip`> | zcorpan_: It's not clear how many are SRT but it sounds like some are |
| 23:22 | <MikeSmith> | TabAtkins: nice |
| 23:22 | <MikeSmith> | "I don't see it. What edition?" |
| 23:22 | <TabAtkins> | I love being a D&D nerd. The thought of D&D defining that is actually plausible. |
| 23:23 | <MikeSmith> | D&D defining anything is plausible |
| 23:23 | <Philip`> | zcorpan_: ...unless they've only got converted-to-XML versions in there |
| 23:23 | <gsnedders> | TabAtkins: Rule 34? |
| 23:23 | <Philip`> | zcorpan_: but I don't fancy downloading 1.6GB just to check |
| 23:24 | <TabAtkins> | gsnedders: http://lmgtfy.com/?q=rule+34 |
| 23:24 | <gsnedders> | Oh. |
| 23:24 | <AryehGregor> | TabAtkins, they would only define that in the BoVD or something. |
| 23:24 | <AryehGregor> | No way would that be in the DMG. |
| 23:24 | <TabAtkins> | BoEF, more than likely, but that's not an actual wotc book. |
| 23:25 | <AryehGregor> | I know of BoVD and BoED, what's BoEF? |
| 23:25 | <AryehGregor> | I probably don't want to know about it. |
| 23:25 | <TabAtkins> | Erotic Fantasy. |
| 23:25 | <AryehGregor> | Yeah. |
| 23:26 | <MikeSmith> | reminds me of the way to give annoying people driving directions: Give them such that they'll head off in a completely wrong direction, in the boondocks, and when you are done explaining most of it, say, "...once you get around there, you might feel like you've gone too far in that direction, but don't worry, just keep going a little further and you'll find it" |
| 23:27 | <TabAtkins> | I actually had to always append that to directions to my old house, because otherwise I'd get calls from people driving to my place wondering where they missed their turn. |
| 23:28 | <AryehGregor> | Why? |
| 23:28 | <AryehGregor> | Also, I guess this was in the pre-Google Maps/GPS era. |
| 23:28 | <TabAtkins> | No, it was just a few months ago. |
| 23:28 | <AryehGregor> | A Luddite, then? |
| 23:28 | <AryehGregor> | Or do people not use those things as much as I'd think? |
| 23:29 | <AryehGregor> | (I don't drive, of course, since I live in Manhattan.) |
| 23:29 | <TabAtkins> | Houston highways are almost completely developed, so you never go any distance without seeing stores or strip malls or whatnot. My house was just far enough out from the city that there was a stretch of forest and a creek before the strip malls started again. |
| 23:29 | <TabAtkins> | It confused people, because usually seeing that means you're on your way to Dallas or Austin. |
| 23:31 | <AryehGregor> | I see. |
| 23:34 | <MikeSmith> | http://www.w3.org/TR/1998/NOTE-xml-names-0119 "error on line 2 at column 30: Encoding error" |
| 23:34 | MikeSmith | resorts to lynx |
| 23:35 | <MikeSmith> | let's hope the lynx developers add an XML parser |
| 23:35 | <MikeSmith> | that'd be fun |
| 23:37 | <TabAtkins> | lynx developers just need to add <video> support via one of the video-to-ascii converters. |
| 23:38 | <TabAtkins> | Then I will finally be able to fulfill my dream: browsing youtube in lynx. |
| 23:41 | <TabAtkins> | re: xml error, hoisted by their own doctype, it appears. |
| 23:45 | <MikeSmith> | Murata-san is 1998 - http://lists.xml.org/archives/xml-dev/199808/msg00164.html |
| 23:46 | <MikeSmith> | [[ |
| 23:46 | <MikeSmith> | As a member of the WG, I have been involved. I have agreed on colonization, |
| 23:46 | <MikeSmith> | and have always believed that colonization provides a good basis for |
| 23:46 | <MikeSmith> | the namespace extension. Now that we have local scoping and declaration by |
| 23:46 | <MikeSmith> | attributes, I start to wonder. |
| 23:46 | <MikeSmith> | Since the same prefix can be bound to |
| 23:46 | <MikeSmith> | different namespaces, it is no longer possible to construct an equivalent |
| 23:46 | <MikeSmith> | XML 1.0 DTD from a collection of namespace-schema pairs. Then, what is |
| 23:46 | <MikeSmith> | the point of using prefixes? |
| 23:46 | <MikeSmith> | It would have |
| 23:46 | <MikeSmith> | been a lot simpler if we had introduced a reserved attribute for specifying |
| 23:46 | <MikeSmith> | the namespace of the element. |
| 23:46 | <MikeSmith> | ]] |
| 23:46 | <Hixie> | man that would have been hella verbose |
| 23:47 | <TabAtkins> | <svg namespace="foobar">? |
| 23:47 | <MikeSmith> | yeah, sure |
| 23:47 | <MikeSmith> | more <svg xmlns="foobar> I guess |
| 23:48 | <MikeSmith> | oops |
| 23:48 | <Hixie> | if it was scoping that might work for svg, html, mathml, etc; wouldn't be very pretty for rdf |
| 23:49 | <zcorpan_> | Philip`: seems it was xml with sentence-for-sentence translations |
| 23:49 | <TabAtkins> | No, rdf would have sucked with that. |
| 23:49 | <TabAtkins> | But it feels like it would have been saner for normal elements. |
| 23:50 | <MikeSmith> | I just point it out as a reminder that there was nothing close to real consensus in the WG at that time |
| 23:50 | <MikeSmith> | no consensus that prefix mechanism was the right way |
| 23:51 | <MikeSmith> | just as there was also no consensus then about draconian error handling being the right way |
| 23:51 | <Hixie> | just as there's no consensus in the htmlwg that html5 is the right way :-) |
| 23:51 | <Hixie> | consensus isn't a good language design mechanism |
| 23:53 | <MikeSmith> | well it's especially not a good design mechanism when claims are made that it exists for a particular issue when it really does not |
| 23:53 | <MikeSmith> | decision about draconian error handling was made by simple majority vote |
| 23:53 | <MikeSmith> | which is not consensus anyway |
