| 03:43 | <jruderman> | annevk: typo on http://www.w3.org/TR/css3-mediaqueries/ where "with" should be "width": "Since the exact with of an ‘em’ cannot be determined" |
| 07:07 | <Hixie> | shepazu: did you ever get a reply to http://www.w3.org/mid/48B26024.8090403⊙wo ? |
| 09:07 | <annevk> | jruderman, thx, fixed in offline copy |
| 09:14 | hsivonen | notices that CSS Mobile Profile 2.0 is edited by an OOo guy from Sun. |
| 09:14 | <hsivonen> | what's the connection between Mobile and OOo? |
| 09:17 | jgraham | wwonders how he pissed off mrlastweek |
| 09:22 | <annevk> | I think hanging out on this IRC channel is generally enough |
| 09:22 | <annevk> | hsivonen, I should know that, but I forgot |
| 09:23 | <hsivonen> | "Very few people followed instructions. The instructions were to specify what was wanted and why. Most people left out the second (and more important) part, which makes their feedback much less useful and their comments much less convincing." -- http://fantasai.inkedblade.net/style/discuss/wasp-feedback-2008 |
| 09:23 | <annevk> | hsivonen, I don't believe there was a direct relation |
| 09:23 | <hsivonen> | annevk: ok |
| 09:46 | <annevk> | http://twitter.com/Kroc/statuses/1123297201 |
| 09:51 | <Philip`> | http://fantasai.inkedblade.net/style/discuss/wasp-feedback-2008 - search for "clear-after" - Opera 9.6 parses it badly and destroys the rest of the page, but it looks like HTML5 actually works there, so that's nice |
| 09:54 | <Philip`> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cdiv%3E%20foo%20%3Ca%20href%3E%3Ccode%3Ebar%3C%2Fa%3E%3C%2Fa%3E%20baz%20%3C%2Fdiv%3E%0D%0A%3Cdiv%3E%20quux%20%3C%2Fdiv%3E |
| 09:54 | <Philip`> | Oddly enough, Opera parses it differently if there's no href attribute on the <a> |
| 09:54 | <annevk> | known :/ |
| 10:02 | <hsivonen> | can someone explain to me, how J. Random User tells apart Class 1 and Class 2 SSL certs? |
| 10:06 | <deane> | hsivonen: What are the "icu" files? In the /dependencies directory I only have commons-codec-1.3, commons-httpclient-3.1 and commons-logging-1.1.1 and the .zip equivalents. Did you mean delete everything in /dependencies and try again? I tried installing it again and got that same error, might try again tomorrow |
| 10:08 | <hsivonen> | dependencies/icu4j-4_0.jar and dependencies/icu4j-charsets-4_0.jar |
| 10:09 | <hsivonen> | deane: ^ |
| 10:09 | <deane> | hsivonen: nah, they aren't there at all :| |
| 10:09 | <hsivonen> | ok, then they didn't download even partially |
| 10:10 | <deane> | I'll try again later, perhaps I might get a better connection |
| 10:13 | <Philip`> | hsivonen: Why would J. ever want to tell apart Class 1 and Class 2 SSL certs? |
| 10:15 | <hsivonen> | Philip`: well, if Class 1 can be obtained without human inspection, how do I know a site pretending to be my bank isn't run by a scammer with a Class 1 cert? |
| 10:15 | <hsivonen> | Philip`: also, what's the point of having Class 2 if the user can't tell? |
| 10:16 | <jgraham> | hsivonen: I guess security UI is such a disaster anyway that it hardly makes a difference |
| 10:17 | <Philip`> | hsivonen: You look in the address bar and if it says "www.thenameofyourbank.com" then it's not a scammer |
| 10:17 | <hsivonen> | it did bother me a bit when I noticed that my bank site is run by (unknown) according to Firefox |
| 10:18 | <hsivonen> | also, it's annoying that I've already encountered two useful sites that had certs from Comodo or their resellers |
| 10:24 | <MikeSmith> | banks should all be required to have EV certs |
| 10:24 | <Philip`> | MikeSmith: Why? |
| 10:25 | <MikeSmith> | cannot be obtained without human inspection |
| 10:25 | <hsivonen> | who uses EV except Paypal, Vidoop and addons.mozilla.org? |
| 10:25 | <Philip`> | MikeSmith: In particular, why would that make any harder for scammers, who can just obtain a non-EV cert (or not even bother with one at all)? |
| 10:25 | <hsivonen> | in the case of Paypal and Vidoop, I really do appreciate the EV cert |
| 10:25 | <MikeSmith> | educate people to not trust non-EV certs at all |
| 10:26 | <MikeSmith> | because we are long ago reached they point where they clearly can't and shouldn't |
| 10:27 | <Philip`> | Educate people to notice the difference between two shades of rectangle in a weird corner of their browser? |
| 10:27 | <hsivonen> | Philip`: the EV green bar is pretty much in your face |
| 10:27 | <hsivonen> | although one has to wonder when someone manages to spoil EV, too |
| 10:27 | <Philip`> | hsivonen: I thought it was the same as the non-EV yellow bar, except green |
| 10:28 | <MikeSmith> | Philip`: EV bar also shows a name |
| 10:28 | <MikeSmith> | a real business/organization name |
| 10:28 | <MikeSmith> | not just a color |
| 10:28 | <Philip`> | Ah |
| 10:28 | <hsivonen> | Philip`: non-EV has no bar, and yellow is *so* last Firefox cycle |
| 10:30 | <hsivonen> | even Google isn't using EV |
| 10:31 | <hsivonen> | Yahoo! isn't, either |
| 10:31 | <hsivonen> | nor Amazon |
| 10:32 | Philip` | was speaking to someone a while ago who said something along the lines of "I don't trust computers enough to do online banking, because there's all these stories of hackers, but I've heard that if it says something like HPTS in the thing then it's meant to be safe" |
| 10:32 | <Philip`> | which I would guess is more educated than average |
| 10:38 | <MikeSmith> | average is closer to somebody getting an e-mail message saying, "Urgent! Your credit-card information may have been stolen. To check, immediately go to the following page and type in your name and credit-card number and expiration date and secret pin code." |
| 10:38 | <MikeSmith> | ..and they do it |
| 10:39 | <annevk> | wtf, http://isitbaconfriday.com/ |
| 10:39 | <annevk> | uses HTML5 though :) |
| 10:39 | <annevk> | hsivonen, EV is pretty expensive I believe |
| 10:41 | <zcorpan> | annevk: some people have a lot of money on their banks I believe |
| 10:41 | annevk | isn't sure he likes it that "trust" needs to bought |
| 10:41 | <hsivonen> | annevk: I don't like having to pay for protection, either |
| 10:41 | <MikeSmith> | annevk: yeahg, relative to normal certs, they're more expensive |
| 10:42 | <MikeSmith> | the reason being the CAs must incur real costs in order to be able to issue them |
| 10:42 | <MikeSmith> | they have to do actual work |
| 10:42 | <annevk> | zcorpan, sure :) |
| 10:42 | <annevk> | MikeSmith, six figures or so I heard |
| 10:42 | <Philip`> | MikeSmith: Then there's the people receiving legitimate emails saying "WARNING: a scam email requesting your password has been circulating. WE WILL NEVER ASK FOR YOUR PASSWORD. For reference, here is the original scam email: "... Please reply with: Username: ******* Password: ********"", and then replying with their username and password filled in |
| 10:42 | <hsivonen> | in the military, I had to explain to my superiors why the Finnish army needs to purchase its trustworthiness from abroad if they want it to work and why purchasing it from a CA owned by the state didn't work in practice |
| 10:42 | <MikeSmith> | unlike regular certs, where they do zero work and get people to pay for them anyway |
| 10:42 | <MikeSmith> | annevk: market prices |
| 10:43 | <hsivonen> | wikipedia claims someone sells EV certs for $500 |
| 10:43 | <MikeSmith> | the CA market is extremely, cut-throatingly competitive |
| 10:43 | <hsivonen> | I guess {citation needed} |
| 10:44 | <MikeSmith> | Philip`: hey, I think I responded to one of those |
| 10:44 | <annevk> | {citation needed} reminds of someone telling me they made {citation needed} stickers and put placed them over advertizements throughout the city |
| 10:51 | <MikeSmith> | anyway, all that EV certs do that normal certs don't is provide an identify component, where the identity has been verified according to rules in a publicly available specification. |
| 10:52 | <MikeSmith> | EV certs cost real money because it costs real money to do real identify verification. |
| 10:52 | <MikeSmith> | there is no standard for identity verification for non-EV certs |
| 10:53 | <MikeSmith> | it's left up to the individual CAs to follow whatever process they like |
| 10:53 | <MikeSmith> | so it of course became a race to the bottom |
| 10:59 | <Philip`> | MikeSmith: Do you expect us to believe that anyone actually follows specifications correctly? :-) |
| 11:00 | <MikeSmith> | heh |
| 11:03 | <hsivonen> | I don't care about bank EV for banking |
| 11:03 | <hsivonen> | I care about it when the bank acts as a Paypal substitute or as an identity broken with redirects from other sites |
| 11:05 | <hsivonen> | s/broken/broker/ |
| 11:07 | <Philip`> | Some UK online shopping sites ask you for credit card details then redirect to www.securesuite.co.uk which asks you to type in some characters from your bank's PIN number and password, which doesn't seem a good way to train people to only enter their details into the bank's own site |
| 11:22 | <othermaciej> | EV certs are snake oil |
| 11:24 | <hsivonen> | othermaciej: does that generalize to certs being snake oil? |
| 11:25 | <hsivonen> | othermaciej: why does Safari support EV? |
| 11:26 | <othermaciej> | it would have been my preference to instead publicize the studies showing that special UI indicators for EV certs do not improve security, and may make it worse |
| 11:26 | <othermaciej> | but I am not the decision-making authority on Safari UI features |
| 11:28 | <othermaciej> | SSL itself at least provides encryption, which protects against man-in-the-middle attacks |
| 11:28 | <othermaciej> | and the lock indicator might be a tiny bit remotely helpful to conscientious users |
| 11:28 | <Philip`> | othermaciej: I thought the problem was that unverified SSL only protected against passive eavesdroppers, not against men-in-the-middle |
| 11:28 | <othermaciej> | (although the danger warning being lack of an indicator is pretty poor design, and EV certs just pile on top more of the same) |
| 11:29 | <hsivonen> | I find PayPal EV helpful when I'm redirected to PayPal from another site of from email |
| 11:29 | <othermaciej> | but you are probably not at risk of being phished in the first place |
| 11:29 | <othermaciej> | thus a feature that protects users of your technical expertise level from phishing is kind of pointless |
| 11:30 | <hsivonen> | of course, I'm required to remember which sites have EV |
| 11:30 | <hsivonen> | I wonder if Firefox would alert me if a site that previously had EV stopped having it |
| 11:30 | <othermaciej> | really all certs should always have had the level of validation that EV certs have |
| 11:31 | <othermaciej> | I think the likely phshing attack would be the "same site" with a subtly different or confusing URL that has a non-EV cert |
| 11:31 | <othermaciej> | and Firefox can't know that this site without an EV cert is trying to look like paypal.com, even though it is served from elsewhere |
| 11:32 | <hsivonen> | good point |
| 11:39 | <Philip`> | Firefox can know that you've never logged into https://paypal.com.zzzzzzzzzzzzzzz.example.com before, and tell you to check it more carefully |
| 11:47 | <othermaciej> | that is true, but of course has nothing to do with EV certs either way |
| 12:11 | annevk | wonders if he should start using rel="feed" |
| 12:15 | annevk | replies to rubys, in case anyone else was doing it too |
| 12:16 | jgraham | wonders what rubys's motivations for missing some of the obvious other choices from his list of four were |
| 12:17 | <annevk> | well, his arguments around 3 characters are also false |
| 12:17 | <annevk> | which is all I tried to point out |
| 12:25 | <Lachy> | I think the fact that ruby incorrectly used <!DOCTYPE HTML "">, that DanC made exactly the same mistake in the telcon yesterday, is evidence that we should discourage people from using the alternative syntax. http://krijnhoetmer.nl/irc-logs/html-wg/20090115#l-676 |
| 12:26 | <Lachy> | even I have sometimes accidentally omitted the PUBLIC keyword when typing the HTML4 DOCTYPE |
| 12:27 | <jgraham> | Lachy: That is probably a point worth making to the list |
| 12:29 | <annevk> | indeed |
| 12:30 | <Lachy> | I'm replying to ruby now |
| 12:35 | <hsivonen> | I think I'm going chill for at least 30 minutes before I reply |
| 12:36 | <hsivonen> | but using "can live with" criteria with bikesheds is a race to an aesthetic bottom |
| 12:41 | <Philip`> | "can be produced by all known tools" sounds like a bad condition, because it's impossible to have a list of all known tools in order to validate the claim, and because I'm sure there are plenty of tools that e.g. always emit the XHTML 1.0 doctype and cannot produce anything else |
| 12:42 | <hsivonen> | Indeed, I've previously made a tool that hard-coded one of the HTML 4.01 Strict doctypes |
| 12:45 | <Philip`> | It seems more practical to give a list of specific tools and describe what they can each produce, and to limit that list to a small number of the most common tools that have a sensible intersection of what they can all output |
| 12:46 | <Philip`> | ...which is actually what the spec currently does, just using a list of about one XSLT tool and perhaps some Java API |
| 12:47 | <Lachy> | does anyone know which other known tools besides XSLT have problems outputting <!DOCTYPE html>? |
| 12:49 | <hsivonen> | Lachy: Julian cited the TrAX serializer, but that's essentially the XSLT serializer |
| 12:49 | <jgraham> | iirc the libxml2 serializer but I may be mistaken |
| 12:50 | <Philip`> | http://www.xom.nu/apidocs/nu/xom/DocType.html - looks like XOM can't do doctypes with a pub-id and no sys-id |
| 12:50 | <Philip`> | so it couldn't do <!DOCTYPE html PUBLIC ""> |
| 12:50 | <Philip`> | ("Sets the public ID for the external DTD subset. This can only be set after a system ID has been set, because XML requires that all document type declarations with public IDs have system IDs.") |
| 12:50 | <hsivonen> | Philip`: but can XOM serialize to text/html anyway? |
| 12:51 | <annevk> | that was not a criteria |
| 12:51 | <Lachy> | Philip`, it looks like it can: http://www.xom.nu/apidocs/nu/xom/DocType.html#DocType(java.lang.String) |
| 12:51 | <hsivonen> | annevk: surely this whole issue is only about the text/html serialization? |
| 12:52 | <Lachy> | Philip`, oops, I misread what you wrote |
| 12:53 | <annevk> | hsivonen, sure, but most XHTML is pretty HTML-compatible |
| 12:54 | <jgraham> | annevk: BTW can you send some testcases for things that are broken in the html5lib serializer/specgen/whatever to me? |
| 12:54 | <Philip`> | hsivonen: Seemingly not, so I guess that's not so relevant |
| 12:55 | <annevk> | jgraham, hopefully later |
| 12:55 | <annevk> | today |
| 12:56 | <Philip`> | Judging by the source code, TagSoup's XMLWriter can't output <!DOCTYPE html PUBLIC ""> |
| 12:57 | <Philip`> | (nor <!DOCTYPE html>) |
| 12:57 | <Philip`> | because it always outputs at least the sys-id |
| 12:58 | <Lachy> | Philip`, is XMLWriter designed for outputting HTML too? |
| 12:58 | <Philip`> | Not sure |
| 12:58 | <Philip`> | Hmm, looks like it is |
| 12:58 | <Philip`> | because it's got special cases for <script> and <br> and all that stuff |
| 13:00 | <Philip`> | Seems it's not even outputting HTML-compatible XML - it just e.g. skips the end tags for <br> entirely, and doesn't write <br /> or anything |
| 13:02 | <jgraham> | annevk: Well if you do, I am much more likely to try and fix the right things :) |
| 13:04 | Philip` | thanks Google for making the code search's source code viewer quite utterly broken in Opera, such that all the text is invisible |
| 13:08 | <jgraham> | Philip`: But if they wanted to support multiple brosers, just think how much effort they'd have to put into hiring the brightest people |
| 13:09 | <Philip`> | To be fair, making a web page that displays some text is really quite challenging |
| 13:19 | <annevk> | maybe it should be <!DOCTYPE html PUBLIC "xml-compat"> |
| 13:19 | <annevk> | than only the XHTML nazis would be affected |
| 13:19 | <annevk> | s/than/then/ |
| 13:22 | <jgraham> | s/xml/sgml/ for less cool value |
| 13:22 | <jgraham> | and s/compat/legacy/ |
| 13:22 | <annevk> | we're getting somewhere |
| 13:23 | <annevk> | :) |
| 13:25 | jgraham | doesn't think anyone will take <!DOCTYPE PUBLIC "sgml-legacy" ""> very seriously even though it is accurate, will presumably work in most tools and encourages <!doctype html> |
| 13:26 | <jgraham> | er. s/ / html / |
| 13:29 | <annevk> | maybe it should be a URL that explains the issue |
| 13:30 | <annevk> | it would please the RDF crowd, it would be long and silly and it clearly tells you at the other end that you should not use it |
| 13:30 | <annevk> | unless you have to |
| 13:34 | <Lachy> | annevk, using a URL as a public identifier would be confusing because people may think it's supposed to be a SYSTEM identifier |
| 13:35 | <Lachy> | although we could use <!DOCTYPE html SYSTEM "http://example.org/dont-use-this-if-you-dont-have-to">; |
| 13:35 | <annevk> | yeah |
| 13:36 | <Lachy> | then people might expect it to resolve to a DTD :-( |
| 13:36 | annevk | needs to stop wasting his time on this discussion |
| 13:49 | <annevk> | hmm, Google Reader is not smart enough to realize my two feeds are now one |
| 13:50 | <annevk> | it also hasn't updated the URL despite the 301, but it did respect the 301 |
| 13:50 | <annevk> | silly |
| 13:50 | <annevk> | and feed readers are supposed to be simple compared to browsers... |
| 13:51 | <Philip`> | What is "Java's SAXTransformer"? |
| 13:51 | <Philip`> | I can't seem to find it anywhere |
| 13:51 | <othermaciej> | I found a Google reference: http://www.jfrn.gov.br:83/javadoc/com/caucho/transform/SAXTransformer.html |
| 13:51 | <Lachy> | yeah, I don't think Google really respects the permanent aspects of 301 and 410 response codes at all |
| 13:51 | <othermaciej> | surprisingly there is no higher ranked JavaDoc site in the results |
| 13:53 | <Philip`> | Oh, maybe it's like http://java.sun.com/j2se/1.5.0/docs/api/javax/xml/transform/sax/package-summary.html |
| 13:53 | <Philip`> | (not actually called SAXTransformer) |
| 13:53 | <Philip`> | othermaciej: That seems to be an unrelated class that just happens to have that name |
| 13:53 | <othermaciej> | awesome |
| 13:54 | jgraham | brings some paint to the bikeshed |
| 13:54 | <hsivonen> | Philip`: it has a totally different *qualified* name. Namespaces FTW! |
| 13:54 | <othermaciej> | I still have no idea what it is though |
| 13:55 | <Philip`> | It's sad that I can read http://java.sun.com/j2se/1.5.0/docs/api/javax/xml/transform/sax/package-summary.html and not acquire any clue at all about what it's talking about |
| 13:55 | <hsivonen> | could the Caucho thing be a pre-TrAX interface for doing what the TrAX SAXTransformer does? |
| 13:55 | <Lachy> | I don't understand what Sam means here: "Given that attaining consensus on preference on this issue is apparently not attainable, I am looking for answers in the form of "I could not live with..."" -- http://lists.w3.org/Archives/Public/public-html/2009Jan/0157.html |
| 13:56 | <Philip`> | It's all handlers and sources and factories, and I can't see the bit where it talks about any actual code being actually executed and actually doing anything |
| 13:56 | <Lachy> | It seems to mean that statments like "I could not live with..." are based on preference, and so go against what he said about attaining concensus on preference not being possible |
| 13:56 | <Lachy> | s/to mean/to me/ |
| 13:57 | <Philip`> | Lachy: Preference is an ordering, "could (not) live with" is a binary function |
| 13:58 | <Philip`> | Lachy: and agreeing on the highest ordered choice is likely to be harder than finding a non-zero intersection of what everyone could live with |
| 13:59 | <annevk> | so my plan of not wasting more time on it obviously fails if everyone starts bombarding the list o_O |
| 14:00 | <Philip`> | You could simply not read the posts :-) |
| 14:00 | Philip` | attempted to add new facts, which is hopefully a positive step |
| 14:00 | <hsivonen> | <!DOCTYPE html SYSTEM "x-cargo:sgml-compat"> |
| 14:02 | <othermaciej> | I was only goaded into responding by Sam's apparent irritation that Ian hasn't immediately jumped on this issue as the most exciting thing to spend his time on |
| 14:02 | <Philip`> | I hope no tools check that the sys-id is a valid non-relative URI |
| 14:03 | <annevk> | oh <!DOCTYPE html SYSTEM "tag:w3.org,01-01-2022:/dead-end"> |
| 14:03 | <othermaciej> | for any possible choice of doctype, there is surely at least one serializer out there somewhere that cannot output it |
| 14:03 | takkaria | suspects bikeshed |
| 14:04 | <annevk> | othermaciej, the funny thing is that none of those serializers can deal with <source> either... |
| 14:04 | <othermaciej> | perhaps the most legacy-compatible thing would be to allow any doctype that would trigger standards mode as per the HTML5 algorithm for determining it |
| 14:04 | <annevk> | othermaciej, but nobody cares |
| 14:06 | <jgraham> | annevk: Sshh! for all our sakes... |
| 14:06 | <Philip`> | othermaciej: There must be two serializers each with a different single hard-coded doctype, so there's trivially no doctype that could work in all serializers |
| 14:06 | <hsivonen> | othermaciej: Didn't Bert Bos already suggest that? |
| 14:07 | <othermaciej> | hsivonen: I don't know - I was blissfully ignorant of this issue until about an hour ago |
| 14:07 | <annevk> | note that <!DOCTYPE html PUBLIC ""> would be introducing another difference between HTML and XHTML syntax |
| 14:07 | <annevk> | I didn't know, but XML requires at least <!DOCTYPE html PUBLIC "" ""> |
| 14:08 | <Lachy> | Philip`, thanks for the evidence. I now accept that XSLT-compat isn't ideal |
| 14:08 | <annevk> | or <!DOCTYPE html SYSTEM ""> |
| 14:08 | <othermaciej> | it seems the call for aesthetic consensus is premature given the apparently incomplete technical information at hand |
| 14:11 | <Philip`> | I guess it's debating the colour of the bikeshed before having checked whether it's large enough to fit a bike into, and before even knowing whether anyone has any bikes they want to store in it |
| 14:11 | <Lachy> | othermaciej, I think this discussion just shows that a concensus based approach is fundamentally flawed. Though I'm glad there have been some technical arguments raised |
| 14:11 | <othermaciej> | Lachy: consensus of a large group is not a great way to make aesthetic decisions |
| 14:12 | <othermaciej> | technical, maybe, since more facts may come to light |
| 14:12 | <othermaciej> | I was tempted to propose "any DOCTYPE that would trigger standards mode" on the list but a sense of guilt stopped me |
| 14:14 | <Lachy> | I like hsivonen's proposal <!DOCTYPE html SYSTEM "about:sgml-compat"> |
| 14:14 | <Philip`> | hsivonen: Why does it matter if people can't spell "compatibility"? The worst that would happen is they'd have the validator complain and tell them that should use the nice easy-to-type <!doctype html> or in rare cases they might want to copy-and-paste <!doctype html system "about:sgml-compatibility"> |
| 14:15 | <hsivonen> | Philip`: it would be pointless to have to debug a spelling error like that |
| 14:15 | <hsivonen> | I don't want validator support requests on English spelling |
| 14:15 | <Philip`> | hsivonen: It wouldn't be pointless, because it would encourage them to use the simpler version and save themselves some effort in the long term |
| 14:16 | <hsivonen> | Philip`: I still foresee a support and documentation problem for me |
| 14:16 | <Lachy> | I also like the idea of giving up on the whole thing and going with just <!DOCTYPE html> :-) http://lists.w3.org/Archives/Public/public-html/2009Jan/0163.html |
| 14:17 | <hsivonen> | are the html5lib meta prescan test cases up-to-date? |
| 14:18 | <jgraham> | hsivonen: I believe so but I woudn't bet my life on them being right or anything |
| 14:19 | <hsivonen> | jgraham: ok |
| 14:21 | Philip` | notes as a pointless technicality that while <!DOCTYPE html SYSTEM "x"> would work for any non-zero-length value of x, it actually won't work in XML::Handler::HTMLWriter if x is "0" |
| 14:23 | zcorpan | notes that about:sgml-compat becomes opera:sgml-compat if typed into opera's address bar |
| 14:24 | <annevk> | same for any value of sgml-compat |
| 14:24 | <Philip`> | I think in really old version of IE it showed a page containing the text "sgml-compat" |
| 14:24 | <Philip`> | *versions |
| 14:25 | <Philip`> | (so about: worked like data:text/html,) |
| 14:28 | <Lachy> | even though about: redirects to opera:, the advantage of using an about: scheme is that people are more familiar with it than, e.g., the tag: URI scheme, and it's non-retrievable, and so there's no expectation of a DTD |
| 14:28 | <Philip`> | What's the practical advantage of using a URI? |
| 14:29 | <Philip`> | Oh, I suppose it's if someone uses an XML tool that downloads DTDs |
| 14:29 | <Lachy> | SGML and XML nazis who point the spec and say that the SYSTEM identifier should be a URI are silenced |
| 14:30 | <Philip`> | I suggest using <!doctype html system "file:///dev/urandom">, just to mess with tools that download DTDs |
| 14:30 | <Lachy> | as long as there are no consumers that choke on an unretrievable DTD |
| 14:30 | <Lachy> | LOL |
| 14:32 | <Philip`> | Unretrievable absolute URIs would be better than unretrievable relative ones so that servers don't get hit with spurious 404ing GETs |
| 14:32 | <hsivonen> | Philip`: that was my point |
| 14:33 | <Philip`> | hsivonen: It just took me some thought to work out that that was your point, since I don't think you actually said it :-) |
| 14:42 | <hsivonen> | jgraham: it seems to me that test with data <meta http-equiv="Content-Type" content="text/html; charset='utf-8"> is wrong |
| 14:42 | <hsivonen> | as far as I can tell, the single quote should disappear in comparison and the test should resolve to UTF-8 |
| 14:44 | <hsivonen> | fix checked in |
| 14:45 | <jgraham> | hsivonen: That seems sensible |
| 14:48 | <zcorpan> | annevk: why not minimized 'required'? |
| 15:00 | <annevk> | i like quotes |
| 15:00 | <annevk> | that was what the "I know" was for |
| 15:00 | <annevk> | well, I sometimes like quotes :) |
| 15:07 | <Philip`> | Hmm, tbroyer's Java code could result in output like <!DOCTYPE html><<!DOCTYPE html>h<!DOCTYPE html>t<!DOCTYPE html>m<!DOCTYPE html>l... which probably isn't quite right |
| 15:23 | <beowulf> | hi, autocomplete="off" is not allowed on element form, am i being lazy for asking that it should be? |
| 15:25 | <Philip`> | beowulf: http://lists.whatwg.org/pipermail/commit-watchers-whatwg.org/2008/001739.html |
| 15:27 | <Philip`> | http://www.w3.org/XML/1998/namespace - "Error: The stylesheet http://rddl.org/xrd.css was not loaded because its MIME type, "text/plain", is not "text/css"." |
| 15:27 | <beowulf> | Philip`: ta |
| 15:27 | <Philip`> | Silly browsers, trusting the server to get the right content-type |
| 15:27 | <zcorpan> | Philip`: blame Hixie |
| 15:27 | <Philip`> | Hixie: It's all your fault |
| 15:31 | <beowulf> | Philip`: so, the reason is because IE doesn't deal with textarea's when the @autocomplete is on the form? |
| 15:33 | <annevk> | beowulf, it also doesn't work in any other browser |
| 15:33 | <Philip`> | beowulf: No, the reason is that not many people use <form autocomplete> and so it's not worth adding support to the non-IE browsers that presumably don't support it already |
| 15:34 | <beowulf> | oh, ok |
| 15:34 | <Philip`> | beowulf: (The <textarea> thing is an independent issue, about <textarea autocomplete> not being added since it's not even in IE) |
| 15:34 | <Philip`> | or at least that's what I gather from the related emails |
| 15:35 | <beowulf> | gotcha |
| 15:37 | Philip` | notes that pretty much the entire XSLT output section is using the word "should", so implementors aren't very highly constrained by it |
| 15:37 | <beowulf> | annevk: are you sure <form autocomplete> doesn't work on any other browser? or were you saying some thing else? |
| 15:38 | <annevk> | beowulf, not sure |
| 15:42 | <beowulf> | i am pretty sure it works, i use it to ensure an event is fired on a input change |
| 15:42 | <zcorpan> | writing an output mode for html5 for xslt 1.0 seems like a useful exercise |
| 15:48 | <annevk> | "CURIE Syntax 1.0" is CR, fail |
| 15:51 | <hsivonen> | annevk: whose FAIL? |
| 15:53 | <annevk> | not sure, but hopefully it doesn't affect me |
| 15:57 | <Lachy> | My banking site uses <form autocomplete=off> and both Firefox and IE support it, AFAIK, so it shouldn't have been dropped |
| 15:58 | <Lachy> | was there a mail from Hixie explaining why autocomplete on <form> was dropped? |
| 15:58 | gsnedders | doesn't think banking sites are a good example of markup |
| 15:59 | <Philip`> | Lachy: It was never dropped, since it was never there in the first place |
| 15:59 | <Philip`> | gsnedders: They're a good example of being required to deal with bad markup :-) |
| 16:00 | <Philip`> | Lachy: http://www.w3.org/mid/Pine.LNX.4.62.0812240739030.24109⊙hdc |
| 16:00 | <Philip`> | "That's very rare. I've removed the XXX in the spec saying we might add this. (IE supports it.)" |
| 16:01 | <Lachy> | I'd better respond to that. It seems that all the major banks in Australia use the autocomplete on <form> |
| 16:05 | jgraham | is dissapointed to find that the population density of Australia is > 1/km^2 so it is harder to claim that Australian's don't represent a significant use case :) |
| 16:05 | <jgraham> | s/'// |
| 16:06 | <Philip`> | It's only > 1/km^2 if you count the dingos too |
| 16:07 | <Lachy> | LOL |
| 16:11 | beowulf | will endeavour to read the archives next time |
| 16:12 | <Philip`> | beowulf: The problem is there's quite a lot of them, and no particularly good search tools |
| 16:13 | <beowulf> | though @autocomplete on a form is handy and it stops my users getting around safari not triggering a change event on with autocompleted values |
| 16:13 | <Lachy> | wow, Australia's population density is only 2.6/km^2. I knew it was low, but didn't realise it was that low |
| 16:13 | <beowulf> | but that's not a problem for here |
| 16:15 | <Lachy> | woah, the ANZ bank website is extremely slow. It's the last of the 4 major banking sites in Australia I need to check |
| 16:18 | <rubys1> | Wyoming — Population - Density: 5.4/sq MI (2.08/km2) |
| 16:21 | <Philip`> | Greenland: 0.026/km^2 |
| 16:22 | <Philip`> | The old marketing trick of calling a frozen wasteland "green" in an attempt to attract settlers has clearly not been extremely successful |
| 16:30 | <Dashiva> | Philip`: Apparently it was a little (but not much) greener back in those days :) |
| 16:37 | <Lachy> | Does that mean the people of Iceland wanted to use the opposite marketing trick of driving people away from their rather green countryside? |
| 16:43 | <gsnedders> | Lachy: No, Iceland just sounds cooler. |
| 16:46 | <Philip`> | Lachy: If they were choosing the name now instead of back whenever it was, they'd probably call it Tesco |
| 16:47 | <Philip`> | Or: Iceland was going to be followed by Fireland, Desertland, Spaceland and Underwaterland, and they were going to be hired out for computer platform gaming enthusiasts |
| 17:34 | <annevk> | http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#attr-textarea-wrap what happens when it has a value that is not valid? |
| 17:34 | <annevk> | would that default to "soft" or to neither? |
| 17:35 | <annevk> | and what is the difference between "soft" and neither? |
| 17:42 | <annevk> | rubys1, before forking, you can always appeal to "WHATWG Members" to get Hixie removed as editor in case he isn't acting properly |
| 17:42 | <annevk> | to the* |
| 17:43 | <annevk> | rubys1, but I guess you have read that already |
| 17:47 | <krijnh> | Did that ever happen? |
| 17:47 | <annevk> | I haven't heard anything |
| 17:48 | <annevk> | but I've only been a member for half a year or so |
| 17:48 | <krijnh> | Yeah |
| 17:58 | <rubys> | I don't think Hixie can be replaced. Nor do I believe I can change Hixie. I therefore simply want to create an environment where others can also contribute constructively. |
| 18:04 | <annevk> | rubys, fair enough, since I've nothing better to do I started updating bits of the FAQ |
| 18:04 | <annevk> | rubys, the ones that are simple, that is :) |
| 18:11 | <rubys> | cool. And by the way, based on experience with open source, providing the option to fork is a great way to call the bluff of people who complain incessantly. |
| 18:11 | <annevk> | I think we have the ability to fork with the current license, no? |
| 18:11 | <annevk> | I believe the W3C copy doesn't |
| 18:12 | <rubys> | If that becomes an issue, that's what I'll work on next. |
| 18:17 | <smedero> | I think the language in the W3C doc license is something like "No right to create modifications or derivatives of W3C documents is granted pursuant to this license." |
| 18:18 | <smedero> | if this is still up-to-date: http://www.w3.org/Consortium/Legal/2002/copyright-documents-20021231 |
| 18:19 | <rubys> | following the links, the original author holds the copyrights. That's the same at the ASF, which I'm familiar with. |
| 18:20 | <rubys> | The W3C license may (or may not) provide additional rights, but it doesn't take away any rights. |
| 18:21 | <rubys> | In any case, I've learned to avoid pursuing hypothetical questions (one of my roles is Vice President of Legal Affairs at the ASF). If this becomes an issue, I'll address the specific issue. |
| 18:22 | <annevk> | I'm not sure I understand "What’s more important, and largely unaddressed, is the fact that the domain of documents which can be serialized as HTML5 and as XHTML5 do not observe a proper super/subset relationship." as XHTML5 would never use text/html |
| 18:22 | <rubys> | There are documents that I can only express in XHTML. There are documents that I can only express in HTML. |
| 18:23 | <annevk> | the point of the question was that from now on you either use HTML or XHTML and that which you use depends on the media type you declare |
| 18:23 | <rubys> | in other words, I get to pick something and later find out that there are things I can't do. |
| 18:24 | <annevk> | yeah, both have limits |
| 18:24 | <rubys> | if one was a proper subset of the other, I would be fine with that. But that is not the case. |
| 18:25 | <annevk> | phrasing your objection to XML well-formedness madness as an objection to that question makes it hard for me to address feedback |
| 18:26 | <rubys> | let me take a look |
| 18:26 | <annevk> | this does not appear to be a problem with the FAQ, but rather with intrinsic properties of HTML and XHTML as things stand today |
| 18:27 | <rubys> | why "stand today"? HTML5 defines some forward thinking things... |
| 18:27 | <annevk> | but we can't tackle the XML issue |
| 18:28 | <rubys> | then, IMO, (X)HTML5 will not "finally put an end to the XHTML as text/html" debate. |
| 18:28 | <rubys> | yes, it would seem that the issues are separable, but I suggest that they are not. |
| 18:28 | <annevk> | though, to be honest, some have suggested we can, by defining an "XML Application" spec that basically defines XML5 and says you have to follow XML 1.0 until you hit a well-formedness error and then you follow XML5... |
| 18:29 | <rubys> | XML5, as you have started to define it, would put an end to the HTML/XHTML debate. Of course it would create an XML1/XML5 debate... but that might not be bad. |
| 18:29 | <annevk> | hmm, I guess I'll just leave that unchanged because I'm not sure what to say instead |
| 18:30 | <rubys> | But first we would need to find someone with the interest, ability, and cycles to pursue XML5. I doubt that such a person exists. |
| 18:32 | <rubys> | done with the wiki page for now? I want to link to your updates... |
| 18:32 | <rubys> | http://realtech.burningbird.net/web/page-markups/why-i-will-never-support-html5 |
| 18:32 | <annevk> | rubys, I updated the browser vendors thing as well to just say "implementors" |
| 18:33 | <rubys> | i'll wait until you are at a stopping point |
| 18:33 | <annevk> | rubys, it was meant to say that anyway, but Hixie chose "browser vendors" because he thought that might be easier for people |
| 18:33 | <annevk> | i reversed it (put the browser vendor thing in the note rather than the implementor thing) as it apparently "upset" you |
| 18:34 | <rubys> | I've had a number of occasions where I have gotten the distinct impression that Ian takes the input of browser vendors over pretty much everybody else. I am not alone in that opinion. And I don't think changing the wiki page will affect that impression. |
| 18:35 | <annevk> | the people implementing standalone parsers or accessibility tools can also ask for priority and some allocated time |
| 18:35 | <annevk> | at least, in my experience |
| 18:36 | <annevk> | non-implementor feedback is another matter, agreed |
| 18:36 | <rubys> | I was thinking more in terms of producers |
| 18:37 | <takkaria> | whatwg is invite-only and w3c has a high barrier to entry? |
| 18:38 | takkaria | is a bit confused at Shelley thinking that |
| 18:40 | <annevk> | rubys, ok, I suppose that's true |
| 18:40 | <annevk> | rubys, done for now |
| 18:40 | <annevk> | rubys, with the FAQ |
| 18:40 | <rubys> | so s/implementors/user agents/? |
| 18:40 | <rubys> | or consumers? |
| 18:41 | <annevk> | oh, I guess I can change it to "user agents" |
| 18:41 | <annevk> | or do you mean something else? |
| 18:41 | <rubys> | that would be great |
| 18:43 | <annevk> | done |
| 18:45 | <gsnedders> | rubys: I think all conforming HTML 5 documents have XHTML5 representations |
| 18:46 | <rubys> | section 1.6 disagrees with you |
| 18:48 | <gsnedders> | Hmm, noscript. |
| 18:49 | <gsnedders> | I think if DOM is conforming it has an XML representation |
| 18:49 | <rubys> | Can comments contain "--" in HTML5? |
| 18:50 | <annevk> | currently a parse error IIRC |
| 18:50 | <annevk> | but they will end up there |
| 18:50 | <gsnedders> | rubys: Non-conforming |
| 18:50 | <gsnedders> | rubys: (But yes, they can) |
| 18:51 | <gsnedders> | (I did deliberately say conforming HTML 5 and the DOM) |
| 18:51 | <rubys> | yea, I see that |
| 18:51 | <rubys> | 8.1.6 |
| 18:51 | <annevk> | <noscript> cannot be in XHTML currently |
| 18:51 | <gsnedders> | And it doesn't get into DOM |
| 18:52 | <annevk> | it doesn't? |
| 18:52 | <annevk> | <meta charset=utf-8> can't go in XHTML |
| 18:52 | <rubys> | that I (personally) would like to see fixed. |
| 18:52 | <rubys> | not a high priority |
| 18:52 | <gsnedders> | Oh, noscript does |
| 18:53 | <gsnedders> | rubys: Should it have any meaning? |
| 18:53 | <annevk> | rubys, as a talisman thing? |
| 18:53 | <gsnedders> | Otherwise we end up having different rules to find character encoding in XML and XHTML |
| 18:53 | <rubys> | basically, but the document would be non-conforming if meta-charset does not match the encoding an XML parser would associate with the document. |
| 18:54 | <rubys> | (my response was to annevk's question) |
| 18:54 | <gsnedders> | And would it do anything if it was different? |
| 18:54 | <rubys> | nope, it would be ignored (which is pretty much what browsers today do with it) |
| 18:55 | <gsnedders> | The only at all possible issue is authors thinking it does something |
| 18:55 | <rubys> | it would allow me to sent the exact same bytes on my weblog as text/html as application/xhtml+xml. |
| 18:55 | <gsnedders> | Can't you just use the charset parameter on the Content-Type header and not have meta@charset at all? |
| 18:56 | <rubys> | the short answer is yes, but the long answer is that content-type is unreliable. |
| 18:57 | <gsnedders> | How so? When is it ignored except when it looks like a feed? |
| 18:57 | <rubys> | by "helpful" caches which are buggy. |
| 18:58 | <gsnedders> | Ah |
| 18:58 | <rubys> | Periodically, I have people complain about my weblog, and it turns out that they are viewing a cached xhtml page served to them as text/html. |
| 18:59 | <rubys> | As opera and firefox move towards the parsing that html5 defines (and is implemented by webkit), my pages look reasonably fine when served as text/html (minus, for the moment, the display of svg). |
| 19:02 | <rubys> | And I actually have versions of my page that deal with browsers (including IE8) that don't yet support html5's parsing rules. |
| 19:03 | <annevk> | rubys, in the comment on your own blog the & got eaten |
| 19:04 | <rubys> | fixed. Thanks! |
| 19:09 | <webben_> | rubys: "turns out that they are viewing a cached xhtml page" ... cached by what? |
| 19:10 | webben_ | is considering creating a bliki serving HTML5 to supporting browsers only, so is interested in the practical problems involved. |
| 19:10 | <rubys> | intermediaries between the client and the server. Could be an ISP. Could be something your employer provides. |
| 19:10 | <webben_> | I see. |
| 19:13 | <rubys> | example: http://intertwingly.net/blog/2008/10/16/Popping-Pie-Partial#c1224229904 |
| 19:15 | <webben_> | hmm |
| 19:20 | <Philip`> | Is there any way to use <pre> or <textarea> with arbitrary element contents, such that the same bytes will work the same in text/html as in application/xhtml+xml? |
| 19:20 | <Philip`> | (what with the leading-newline-stripping behaviour in the text/html parser) |
| 19:27 | <BenMillard> | othermaciej, I wrote about allowing other doctypes here (specifically, W3C ones which trigger standards mode): http://lists.w3.org/Archives/Public/public-html/2007Jun/0318.html |
| 19:30 | <rubys> | Phillip`: In general, no. But a lot of content doesn't pose a problem. |
| 19:30 | <rubys> | For example, my code samples don't typically start with a newline. |
| 19:31 | <rubys> | It actually is legal to put the text on the same line as the <pre> in html :-) |
| 20:56 | <annevk> | http://www.reddit.com/r/reddit.com/comments/7q7l1/this_must_be_the_most_embarrassing_book_an_author/ lol |
| 21:01 | <Dashiva> | "This book has a very interesting premise, which is that common sense and experienced reality are both false." |
| 21:20 | <olliej> | Dashiva: hehe |
| 22:34 | <Philip`> | http://youtube.com/watch?v=4sZKlKEU2do - now with "click to download" link, returning MPEG4 |
| 22:34 | <Philip`> | (under a rubbish filename of "video.mp4") |
| 22:34 | <Philip`> | Claims to be MPEG-4 AAC LC, 240.024 secs, 107 kbps, 44100 Hz, and H264 Baseline⊙2, 239.572 secs, 351 kbps, 480x270 @ 29.970113 fps |
| 22:36 | <Hixie> | oh we finally added the download link? |
| 22:37 | <Philip`> | Not to all videos |
| 22:37 | <Hixie> | i think it's user opt-in |
| 22:37 | <Philip`> | Maybe only to Obama ones, because people were complaining about that a lot |
| 22:38 | <Hixie> | so on a totally different front |
| 22:39 | <Hixie> | am i the last person to find out that IE doesn't expose window.toolbar and company? |
| 22:43 | annevk | wonders what .toolbar is |
| 22:45 | <Hixie> | opera doesn't have it either it seems |
| 22:45 | <Hixie> | though i got the weirdest error trying to find out |
| 22:45 | <Hixie> | var x = {}; for (i in window) if (window[i] == "[object BarInfo]") x[i] = 1; x |
| 22:45 | <Hixie> | => Statement on line 1: The Object does not implement [[Call]] |
| 22:45 | <Hixie> | there are no method calls in that! |
| 22:46 | <Hixie> | wtf |
| 22:47 | <Dashiva> | window[i].toString for some i, maybe |
| 22:48 | <Hixie> | i guess |
| 22:50 | <Dashiva> | [[Call]] is pretty low-level in the first place. I'd expect 'not a function' or something for a direct misplaced call |
| 22:51 | <Hixie> | window.java, window.sun, window.netscape, and window.Packages |
| 22:52 | <Hixie> | those are the ones that don't implement [[Call]] |
| 22:52 | <Hixie> | i wonder whether to bother supporting these properties |
| 22:52 | <Hixie> | IE doesn't |
| 22:52 | <Hixie> | (the BarProp ones) |
| 22:52 | <Hixie> | (or BarInfo in Safari) |
| 22:56 | <Hixie> | hm, my opera might just be out of date |
| 22:56 | <Hixie> | how does one update to the latest opera build again? |
| 22:57 | <Hixie> | http://snapshot.opera.com/ ? |
| 22:57 | <annevk> | http://my.opera.com/desktopteam/blog/ |
| 22:58 | <annevk> | in opera:config#AutoUpdate you can set a setting to make sure it always updates to the latest snapshot |
| 22:58 | <Hixie> | oh good |
| 22:58 | <Hixie> | that must be new since the version i have |
| 22:59 | <Hixie> | i have some ancient build |
| 23:02 | <Hixie> | well that doesn't seem to have changed the result of my test |
| 23:05 | <annevk> | sweet: http://www.filmroster.com/article/60000_piece_star_wars_lego_diorama |
| 23:05 | annevk | is just watching that movie |
| 23:06 | <Hixie> | i love teh footprints in the snow |
| 23:31 | gsnedders | will probably annoy everyone in here this weekend constantly asking for help with maths :P |
| 23:32 | Philip` | likes people asking maths questions iff he can easily remember the answer, but otherwise it makes him realise how much he's forgotten |
| 23:33 | <gsnedders> | Philip`: How much can you remember complex numbers? |
| 23:33 | <gsnedders> | (Or rather, how to do anything with them?) |
| 23:34 | <Philip`> | Complex numbers are easy, they're just like normal numbers but there's two of them |
| 23:34 | <gsnedders> | :) |
| 23:34 | <gsnedders> | Then hopefully it satisfies the iff |
| 23:35 | <Philip`> | Alternatively: Complex numbers are easy, they're just geometry - you simply imagine each number being a line, and then the answers are obvious |
| 23:35 | gsnedders | doesn't like making imaginary things into real things to plot them on a graph though |
| 23:36 | <gsnedders> | it makes them seem real. |
| 23:36 | <Hixie> | complex numbers are easy, just s/i/sqrt(-1)/ and then it all makes sense |
| 23:36 | <gsnedders> | Hixie: Surely s/i/sqrt(-1)/g? |
| 23:37 | <Hixie> | i thought you didn't like me calling you g :-P |
| 23:37 | <gsnedders> | Hixie: :P |
| 23:37 | Philip` | is currently busy failing to remember how to do the maths for a Breakout game |
| 23:37 | gsnedders | wonders if sed has an h flag |
| 23:38 | <Hixie> | Philip`: do it on a text terminal, the collision detection algorithm becomes much easier |
| 23:39 | <Philip`> | Actually I think the problem is not the maths, it's that I'm modelling the world as if it contains several independently-moving objects with infinite mass, which causes problems when the ball gets stuck between them |
| 23:40 | <Hixie> | if they can both move and they can collide, that's gonna give you issues, yes |
| 23:41 | <Hixie> | you should do breakout using box2d's engine |
| 23:42 | <Hixie> | complete overkill, but you could do some pretty funky things |
| 23:42 | <Hixie> | like having the boxes knock each others over or something |
| 23:42 | <Philip`> | Using a library would be cheating :-) |
| 23:43 | <Hixie> | using a library is sensible :-P |
| 23:44 | <Philip`> | Being sensible is cheating |
| 23:46 | <Lachy> | gsnedders, I'm very surprised that your working with complex numbers. I didn't that would be covered in high school. It wasn't for me. |
| 23:47 | <Philip`> | You didn't do complex numbers? :-o |
| 23:50 | <Lachy> | That might have been covered in the 3 unit maths in years 11 and 12 that I didn't do |
| 23:52 | <Lachy> | I saw a video once last year that explained how they work really clearly. I can't remember where though |
| 23:52 | <Hixie> | shockingly, complex numbers were not covered for me until advanced maths (the math level beyond the normal A-level maths) |
| 23:59 | <annevk> | Hixie, seems XXX-Origin can be named Origin again |