| 00:02 | <tantek> | Lachy - regarding the page, I've added questions to the one Pro item of the sort that I think should be asked about Pro and Con items of any particular proposed solution. |
| 00:04 | <Lachy> | tantek, yeah, I intend to provide that info. There's some notes about it in the esw wiki and others I've seen in the mail archives that I will have to find |
| 01:37 | <Hixie> | sicking: do you think we should require SQL support? |
| 01:57 | <Philip`> | Hixie: philip.html5.org is more recent, and less of a random meaningless domain |
| 01:59 | <Hixie> | k |
| 02:04 | <Philip`> | jgraham: The cost of the ifs seems pretty irrelevant in overall html5lib performance, so it doesn't really seem like an interesting thing to optimise, if I remember correctly |
| 02:05 | <Philip`> | Changing e.g. the way it looks at content model flag doesn't seem to affect much either |
| 02:06 | <Philip`> | (I think I tried making an extremely stripped down tokeniser that basically parses < and tag names and > and data, or something like that, and it was quite similarly slow to html5lib) |
| 02:18 | Philip` | notes that Alton Towers has nice rollercoasters |
| 03:07 | <roc> | Hixie: are you comparing requiring SQL with requiring Theora? |
| 03:39 | jwalden | would be fine with sql if there were general agreement upon some rigorous specification of exact syntax and behavior |
| 03:40 | <jwalden> | which doesn't seem likely to happen given the enormity of the prerequisite tasks, but is not entirely unlikely |
| 04:01 | <Hixie> | roc: well, a lot of the arguments sicking makes seem to apply equally to ignoring mozilla's preference and requiring support for the sql database api |
| 04:04 | <roc> | The arguments seem completely different to me |
| 04:05 | <othermaciej> | in each case, the argument for inclusion would be to promote interoperability by exerting market pressure on holdouts |
| 04:05 | <roc> | but the reasons for people holding out are completely different |
| 04:06 | <othermaciej> | so your Mozilla's reasons for holding out are valid but Apple's are not? |
| 04:06 | <roc> | Modulo my limited understanding of Apple's reasons --- yes. |
| 04:08 | <roc> | if Web SQL was the only plausible solution for structured client-side storage, I'm sure we'd just bite the bullet and do it. |
| 04:12 | <othermaciej> | so existence of plausible (though not yet materialized) alternatives justifies holding out, in a way that specs should accomodate? |
| 04:13 | <roc> | depends on what you mean by "materialized" |
| 04:13 | <roc> | http://hg.toolness.com/browser-couch/raw-file/blog-post/index.html |
| 04:14 | <roc> | there are many other differences between the two situations |
| 04:14 | <Hixie> | fundamentally, i don't think i should be in the business of saying that apple's reasons are less good than mozilla's reasons |
| 04:14 | <Hixie> | in particular, not when it won't make any difference anyway |
| 04:14 | <Hixie> | as in this case |
| 04:15 | <roc> | e.g., only one browser implements Web SQL currently, so there's no interop to speak of, whereas we have three browsers with demoed or shipping Theora implementations, one of which is actually different code |
| 04:15 | <roc> | there's the issue of Web SQL not having a real spec |
| 04:15 | <othermaciej> | Right now we have 1 browser shipping it and 2 that have publicly committed to doing so |
| 04:15 | <roc> | Has Opera committed to it? |
| 04:15 | <othermaciej> | vs for Theora, 2 browsers shipping and 1 publicly committed |
| 04:16 | <othermaciej> | Opera has said they will do it |
| 04:16 | <Hixie> | (websql will have a real spec after html5 is in last call, assuming no better alternatives surface -- i'll write it myself if i have to :-) ) |
| 04:16 | <roc> | othermaciej: OK, that strengthens the case for Web SQL a bit. |
| 04:17 | <othermaciej> | I think Mozilla is entitled to wait and see what happens, even if I would prefer you did implement Web SQL |
| 04:17 | <othermaciej> | that's why I proposed splitting it from the rest of the Web Storage spec |
| 04:17 | <roc> | I doubt we'll see an independent implementation of SQLite's SQL dialect, though, which remains a worry |
| 04:17 | <Hixie> | i think mozilla's concerns with websql are perfectly valid |
| 04:17 | <Hixie> | i think apple's concerns with theora are valid too |
| 04:18 | <Hixie> | insofar as mozilla thinks their concerns are valid, and apple thinks their concerns are valid |
| 04:18 | <othermaciej> | fwiw Apple is doing a new round of research on the matter |
| 04:18 | <roc> | I don't think we should shy away from making value judgements about peoples' concerns |
| 04:18 | <roc> | people have unreasonable concerns sometimes |
| 04:18 | <roc> | I know it can sound arrogant |
| 04:18 | <Hixie> | that's a very dangerous path to go down, imho |
| 04:19 | <Hixie> | and not one that leads to productive cooperation |
| 04:19 | <Hixie> | what would you do if i said that mozilla's concerns re websql weren't valid? |
| 04:19 | <Hixie> | i doubt you would really appreciate it |
| 04:19 | <roc> | We'd argue about it |
| 04:19 | <roc> | in the past I've had a lot of concerns that you have assuaged in one way or another |
| 04:20 | <roc> | I don't think I personally, or Mozilla in general, am immune to unrealistic or overblown concerns |
| 04:21 | <roc> | putting them on an untouchable pedestal is itself dangerous |
| 04:21 | <Hixie> | i'm talking specifically about "we won't implement X" feedback |
| 04:21 | <Hixie> | for more nuanced feedback, things are more complex |
| 04:21 | <othermaciej> | I think the only thing anyone has done to assuage Apple's concerns (or Google's) about Theora is to say they shouldn't count or imply that they are not on the level |
| 04:21 | <roc> | Mozilla has not said "we won't implement SQL" |
| 04:22 | <roc> | at least, I don't think we have |
| 04:22 | <Hixie> | not quite as explicitly as apple has said it about theora, i'll grant you |
| 04:23 | <Hixie> | anyway, my point was just that sicking's comments seemed to be dangerously close to saying "ignore apple", which i don't think it would be appropriate for me to do, as i think it would set a precedent for ignoring any vendor whose feedback isn't convenient. |
| 04:23 | <Hixie> | which could be mozilla, e.g. as with websql |
| 04:24 | <Hixie> | i'm just trying to treat everyone fairly :-) |
| 04:24 | <roc> | othermaciej: there's an interesting thing. Google's got concerns about Theora not being a great codec. Fine. But they're implementing it anyway. I don't understand why Apple can't do the same. |
| 04:24 | <roc> | othermaciej: or at least, patents are the only reason I can understand. |
| 04:24 | <othermaciej> | our concerns are that we could get sued, and that it might not be viable on the phone |
| 04:24 | <othermaciej> | we're going to investigate both these issues much more in depth |
| 04:25 | <sayrer_> | roc, btw, here's the message from MS on SQL dialects: |
| 04:25 | <sayrer_> | http://groups.google.com/group/mozilla.community.web-standards/msg/d6a92db27bd52bcb |
| 04:25 | <roc> | "viable" as in "consumes more power than your H.264 hardware decoder" or as in "Will Not Work At All"? |
| 04:25 | <roc> | the former is a red herring, I think. |
| 04:26 | <othermaciej> | I wouldn't phrase it in as absolute a way as you just did |
| 04:26 | <roc> | the latter would be a shock |
| 04:27 | <doublec> | vlc plays theora videos on jailbroken iphones so it's certainly possible |
| 04:27 | <roc> | One question in the mega-thread that I asked that never got answered is "how would supporting Theora make the iPhone *worse*"? |
| 04:27 | <othermaciej> | "viable" as in "consumes more power than seems reasonable" and "renders at adequate performance relative to the quality of experience we expect on iPhone OS" |
| 04:28 | <othermaciej> | er |
| 04:28 | <othermaciej> | I phrased one of those backwards |
| 04:28 | <roc> | does Apple reject applications from the iPhone store because they run the CPU at 100%? |
| 04:28 | <othermaciej> | "doesn't consume more power than seems reasonable" |
| 04:28 | <roc> | does Safari refuse to render Web pages that run too much JS? |
| 04:28 | <othermaciej> | we do limit background apps in part to avoid killing the battery |
| 04:28 | <othermaciej> | and yes, MobileSafari has a JS execution limit |
| 04:28 | <roc> | what sort of limit? |
| 04:29 | <othermaciej> | run JS for too long within a given time period and it gets shut off for that page |
| 04:29 | <roc> | ok |
| 04:30 | <roc> | I wonder what Opera thinks about this mobile power issue |
| 04:30 | <roc> | since that's most of their business |
| 04:30 | <othermaciej> | anyway, to clarify the parameters, if it took 10x as much battery that would probably be unacceptable, if it took 1.5x, that probably wouldn't be |
| 04:30 | <doublec> | the camera app on the iphone consumes an enormous amount of power - so does skype. Ban them imho. |
| 04:31 | <othermaciej> | anyway I think it needs to be put to the test whether a software-only implementation could be acceptable |
| 04:32 | <othermaciej> | I suspect Opera cares less, because it's the device vendor, not the app vendor, who faces the users' ire over poor battery life |
| 04:33 | <roc> | I suspect device vendors are able to forward that feedback |
| 04:34 | <sayrer_> | roc, anyway, I think the SQL comparison is not such a good. I don't think the storage draft has a dangling dependency. |
| 04:34 | <roc> | sayrer_: it's worse, it has a dependency on something that doesn't exist, a spec for SQLite |
| 04:34 | <sayrer_> | well, if the storage draft contains SQL |
| 04:35 | <sayrer_> | if it's moved into a new Web SQL API draft, that document still has it |
| 04:35 | <roc> | right, that could be fixed. I believe Ian when he says he'll fix it if necessary. But we can't be blamed for pushing back against the draft that exists today |
| 04:36 | <roc> | no-one really knows how hard it will be to spec out the SQL bit. I have faith in Ian's supreme speccing skills, that's all |
| 04:36 | <sayrer_> | roc, I read the response as saying the splitting of the storage spec was due to our lack of implementation |
| 04:36 | <sayrer_> | but it seems to me that splitting the spec is not really the same as leaving something unspecified |
| 04:39 | <Hixie> | roc: :-) |
| 04:55 | jwalden | wishes hsivonen were not in a time zone that's still somewhat asleep right now |
| 04:56 | jwalden | reads the .web-standards newsgroup posts |
| 04:58 | jwalden | hates Hates HATES how threading is utterly destroyed by the unholy mailing list-newsgroup chimera the mirroring software creates |
| 06:12 | <MikeSmith> | anybody know of GUI alternative to ssh-keygen? |
| 06:14 | <kinetik> | MikeSmith: PuTTYGen, on Windows. |
| 06:14 | <MikeSmith> | kinetik: thanks |
| 06:16 | <MikeSmith> | kinetik: know if there's any equivalent for OSX? |
| 06:17 | <MikeSmith> | I got an request by e-mail for info about generating a keypair, and don't know if he's on Windows or Mac |
| 06:18 | <MikeSmith> | and no clue personally because I've never used anything but ssh-keygen myself |
| 06:18 | <jwalden> | SSHKeychain, I *think* |
| 06:18 | <gavin_> | send him http://github.com/guides/providing-your-ssh-key ? |
| 06:18 | <gavin_> | it has instructions for win/mac/linux |
| 06:20 | <MikeSmith> | jwalden: I find something called "SSH Agent" (not so great a name, since there's the ssh-agent CLI tool) |
| 06:20 | <MikeSmith> | gavin_: great, thanks |
| 06:21 | <jwalden> | http://www.sshkeychain.org/ was what I referred to |
| 06:24 | <MikeSmith> | jwalden: OK |
| 06:24 | <jwalden> | comment enough in <https://bugzilla.mozilla.org/show_bug.cgi?id=472529>;, Hixie? :-P |
| 07:36 | jgraham | wonders what happens if Microsoft announce that they will never support <canvas> or that they will never replace their non-tree-like DOM |
| 07:39 | <hober> | I think we're OK in the latter case |
| 07:40 | <jgraham> | hober: afaict Hixie's arguments about requiring vendor consensus means that we would have to either a) get everyone else to agree to the IE model or b) leave a big chunk of the parsing mechanism unspecified |
| 07:44 | <hober> | My understanding is that the cases in which IE's DOM is a non-tree aren't important for web compat |
| 07:45 | <hober> | which is how the other browsers have gotten away with not reverse-engineering it |
| 07:48 | <jgraham> | hober: Indeed. But that seems to me to be a more nuanced position than Hixie has taken. |
| 07:49 | <jgraham> | (possibly because the examples that have actually come up have not required it) |
| 07:53 | <hsivonen> | jwalden: I'm awake now |
| 07:54 | <jwalden> | hsivonen: aha! |
| 07:54 | <jwalden> | hsivonen: so most people on planet have full-text feeds there |
| 07:55 | <jwalden> | dunno if that was intentional or not, but it's a rather marked difference from most of the other posts |
| 07:56 | <hsivonen> | jwalden: My full text feed has the full text for the latest entry and the most recent entries from 24 hours |
| 07:56 | <hsivonen> | jwalden: unless there's a bug, of course |
| 07:56 | <jwalden> | hm |
| 07:57 | <hsivonen> | jwalden: the older entries only get summaries and typically don't show up on planets anyway after initial setup |
| 07:57 | <hsivonen> | jwalden: though the other planets I'm on are actually Venus |
| 07:57 | <hsivonen> | in case that makes a difference |
| 07:57 | <jwalden> | well, at least validation 2.0 and large-amount-of-finnish-I-can't read are showing up in google reader without a body |
| 07:57 | <jwalden> | ;-) |
| 07:58 | <hsivonen> | jwalden: but the latest entry in the Atom feed has full text, right? |
| 07:59 | <hsivonen> | (I should rewrite my feed generator anyway. the current one uses severely legacy software and I can't touch it without breaking it) |
| 08:00 | <jwalden> | hm, I think so; I have auto-send-me-elsewhere in all the browsers I can test directly now, view-source: suggests yes but there's no formatting, so it's hard to say (wonder if mime type is wrong maybe) |
| 08:01 | <hsivonen> | works on Planet Intertwingly |
| 08:03 | <hsivonen> | maybe the Planet Mozilla view becomes normal once I update when planet is already subscribing |
| 08:03 | <hsivonen> | unless the version of planet at Planet Mozilla is so old that it can't deal with XHTML-in-Atom |
| 08:03 | <jgraham> | hsivonen: I see the full text of " |
| 08:04 | <jgraham> | An Unofficial Q&A about the Discontinuation of the XHTML2 WG" on planet.mozilla.org |
| 08:04 | <hsivonen> | oh. right. now I see it too. |
| 08:04 | <hsivonen> | I have no idea why the old posts appear after it |
| 08:04 | <hsivonen> | after in time |
| 08:04 | <hsivonen> | before in page order |
| 08:05 | <hsivonen> | oh, and planet.mozilla.org says it's running Venus |
| 08:05 | <hsivonen> | so things should go normal with the next update |
| 08:05 | <hsivonen> | hopefully |
| 08:06 | <jwalden> | heh |
| 08:06 | <jgraham> | hsivonen: Write a short article about using the hml5 parser in nightlies so we can find out :) |
| 08:07 | <hsivonen> | jgraham: I will once comment end bang is on m-c |
| 08:07 | <hsivonen> | jgraham: since not having it known to be a compat issue, so there's no point in inviting reports about not having it breaking pages |
| 08:10 | <jgraham> | me things that "comment end bang" is a nice name for something |
| 08:11 | <hsivonen> | jgraham: like "spurious triples" |
| 08:22 | <Hixie> | jgraham: i don't know what we would do if we couldn't get consensus on something as core as parsing |
| 09:37 | <hsivonen> | http://www.sandisk.com/Products/Catalog%281433%29-Players_bundles.aspx is interesting with the HTML5 parsing algorithm |
| 09:37 | <Hixie> | hsivonen: the squares at the bottom of your pages always confuse me |
| 09:37 | <Hixie> | i always think they're an image that failed to load |
| 09:37 | <Hixie> | and spent 10 seconds each time interacting with them |
| 09:38 | <hsivonen> | Hixie: do you mean the dingbat that I use for <hr>? |
| 09:38 | <Hixie> | yeah |
| 09:38 | <hsivonen> | Hixie: it's a PUA glyph. you need @font-face supporting UA :-) |
| 09:38 | <Hixie> | i have one |
| 09:38 | <Hixie> | two even |
| 09:38 | <Hixie> | still looks like a square :-) |
| 09:39 | <hsivonen> | Hixie: interesting. which UA and which platform? |
| 09:39 | <Hixie> | mac os x, safari and firefox nightlies |
| 09:39 | <Hixie> | actually i haven't tried firefox |
| 09:39 | <Hixie> | works in firefox |
| 09:40 | <Hixie> | that sandisk page is sad |
| 09:40 | <Hixie> | any idea what's wrong yet? |
| 09:40 | <hsivonen> | Hixie: I think it's <ul></li> |
| 09:40 | <hsivonen> | Hixie: but further than that, I don't know yet |
| 09:41 | <Hixie> | k |
| 09:41 | <hsivonen> | Hixie: Safari WFM with the font installed locally. Apple makes it too hard to test Safari with @font-face if you have the font file anywhere on disk |
| 09:42 | <hsivonen> | Hixie: and Parallels is too scared of Apple to let me run a clean VM |
| 09:43 | <Hixie> | i wish safari would show your content before the ttf file downloaded |
| 09:43 | <Hixie> | firefox's flickering experience is actually nicer, imho |
| 09:43 | <hsivonen> | I suppose I could uninstall the font *and* go zipping all instances of the font files around all mounted volumes so that Safari couldn't detect the presence of the font on disk |
| 09:44 | <hsivonen> | Hixie: I wish hyatt agreed with you |
| 09:44 | <hsivonen> | maybe I'll just take a look the next time I visit another Mac user who doesn't have the font locally |
| 09:50 | <jgraham> | hsivonen: WFM in firefox / opera on Linux |
| 09:54 | <MikeSmith> | jgraham: which page? http://www.sandisk.com/Products/Catalog%281433%29-Players_bundles.aspx ? |
| 09:54 | <MikeSmith> | what would I see in the browser if that page not being rendered as expected? |
| 09:54 | <MikeSmith> | replacement characters or artifacts on the page somewhere? |
| 09:56 | <jgraham> | MikeSmith: No hsivonen's funky pua thing |
| 09:56 | <MikeSmith> | ah |
| 09:56 | <jgraham> | MikeSmith: The sandisk page breaks for me in firefox with html5.enable |
| 09:56 | <jgraham> | (look at the top menu) |
| 09:57 | <MikeSmith> | jgraham: oh. re-reading the scrollback now, I see that's what hsivonen mentioned |
| 09:57 | MikeSmith | got confused |
| 09:58 | <hsivonen> | The sandisk issue minimizes to <!DOCTYPE html><ul><li>A<ul></li>B |
| 10:00 | <Lachy> | Hixie, re http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-July/020783.html - you mentioned you're going to be taking out the SQL stuff from webstorage. Will it be replaced by something else, or just progress in its own spec? |
| 10:00 | <Hixie> | Lachy: separate spec for now |
| 10:01 | <Hixie> | Lachy: maybe other specs will be introduced to compete with it |
| 10:01 | <Hixie> | we'll see |
| 10:01 | <Creap> | Is it ok to ask short html5 questions here, or is this channel about the spec process? |
| 10:01 | <Hixie> | Creap: both |
| 10:02 | <Creap> | ok |
| 10:02 | <Hixie> | though this is often a bitch channel so you might not get useful replies :-) |
| 10:02 | <gsnedders> | Also, it's just a social channel damnit :P |
| 10:02 | <Creap> | isn't that just IRC overall :) |
| 10:03 | <MikeSmith> | hsivonen: weird. The PUA wfm on Arora on Linux (built against Webkit trunk) but not in latest Chrome (191) |
| 10:04 | <Creap> | well, I'm considering having a fieldset with a meter measuring how secure a password a user has entered is, and some text on how to improve it. technically it's not a form element.. |
| 10:04 | <Creap> | maybe if I combine it with an output? |
| 10:04 | <MikeSmith> | hsivonen: maybe beause Arora is QtWebkit, Chrome is GTK+ |
| 10:04 | <Lachy> | Hixie, have you explained what will happen with the SQL stuff in more detail, elsewhere? |
| 10:04 | <Lachy> | like in any mail I may have skipped in public-webapps? |
| 10:05 | <Hixie> | no |
| 10:05 | <Hixie> | bu tthere's no much to explain |
| 10:06 | <Hixie> | i just plan to split web storage into two, web storage and web database |
| 10:06 | <Lachy> | ok |
| 10:07 | <hsivonen> | Hixie: for now, I'm treating the PUA thing as a bug that should be fixed in WebKit rather than fixing it in my CSS |
| 10:07 | <Hixie> | hsivonen: wow, weird. i thought IE ignored </li> altogether. send mail, we should be able to fix that. |
| 10:08 | <hsivonen> | Hixie: this is interesting: http://msdn.microsoft.com/en-us/library/ms533020%28VS.85%29.aspx#Close_Your_Tags |
| 10:08 | <Lachy> | what's the reason for splitting it? Is it because it's not clear whether the SQL stuff has an uncertain future? |
| 10:08 | <Hixie> | hsivonen: btw if you could comment on mjs' idea of introducing a third level of conformance (summary being conforming but showing a warning) that would be useful |
| 10:08 | <Lachy> | let me rephrase... |
| 10:08 | <MikeSmith> | hsivonen: hmm, even http://sandisk.com page looks equally porked with html5.enable |
| 10:08 | <Hixie> | hsivonen: (on the list, about to go to bed) |
| 10:08 | <Lachy> | ... Is it because the SQL stuff has an uncertain future? |
| 10:09 | <hsivonen> | Hixie: assuming that the MSDN article has been written by someone who actually knows IE internals |
| 10:09 | <Hixie> | Lachy: main reason is because they are independent features, and might have different lifecycles |
| 10:09 | <Lachy> | ok |
| 10:09 | <hsivonen> | Hixie: I haven't read list email today. I'll get to it. |
| 10:09 | <Hixie> | Lachy: but the uncertain future and the weirdness from the oracle guy were further motivating factors |
| 10:10 | <hsivonen> | I'll get to it some time, not now that is. |
| 10:10 | <Hixie> | hsivonen: no rush, but i'm holding off on replying to mjs pending your input |
| 10:10 | <Hixie> | looking at msdn page... |
| 10:10 | <Hixie> | Creap: <meter> would be appropriate for that in html5 |
| 10:10 | <Hixie> | Creap: not supported anywhere yet of course |
| 10:11 | <Hixie> | Creap: <output> would be fine too |
| 10:13 | <Creap> | nope, but since javascript will be required anyway, I might as well use meter with attributes for visual hints, for instance, if value is lower than 'low' then make the text red |
| 10:14 | <Hixie> | hsivonen: if that's written by someone who knows their parser, that's a frightening insight into what their code does |
| 10:23 | <jgraham> | hsivonen: There seems to be a typo in your bug report |
| 10:23 | <Philip`> | hsivonen: I guess http://www.w3.org/Bugs/Public/show_bug.cgi?id=7084 should say <ul>, not <iul> |
| 10:23 | Philip` | shakes his fist at jgraham |
| 10:24 | jgraham | laughs, manically |
| 10:25 | <Philip`> | Oh, and there's a </l> |
| 10:27 | <hsivonen> | I guess some accidental drag&drop moved the i around |
| 10:27 | <gsnedders> | jgraham: Also amazingly silently for manic laughter |
| 10:27 | <Philip`> | http://slashdot.org/pollBooth.pl?qid=1813&aid=-1 |
| 10:28 | <hsivonen> | jd found stackoverflow and got downvoted. |
| 10:29 | <gsnedders> | hsivonen: link? |
| 10:29 | <hsivonen> | gsnedders: http://stackoverflow.com/users/133893/john-dowdell |
| 10:30 | <gsnedders> | http://stackoverflow.com/questions/1086886/html-5-video-tag-vs-flash-video-what-are-the-pros-and-cons/1089156#1089156 is his one and only comment |
| 11:12 | <Dashiva> | krijnh: More logs? |
| 11:12 | <krijnh> | For #webapps, dont' know |
| 11:13 | <krijnh> | Still gotta put yours online, but they're on a different machine atm :) |
| 11:13 | <Dashiva> | And reformat them, I bet :) |
| 11:14 | <krijnh> | Nah, why? :) |
| 11:14 | <Dashiva> | You don't seem to be on w3c-irc right now, so I can hold off sending them until you're back stable |
| 11:14 | <krijnh> | I'm not? |
| 11:14 | <Dashiva> | Well, the logs I sent you include the control characters for colors and stuff |
| 11:14 | <krijnh> | Ah, I'm not :p |
| 11:15 | <krijnh> | http://krijnhoetmer.nl/irc-logs/webapps/20090707#l-25 |
| 11:16 | <krijnh> | I don't think those characters are a problem |
| 11:17 | <Dashiva> | Well, if you want user modes you have to translate the nick colors :) |
| 11:18 | <Dashiva> | Since I use color coding instead of mode prefixes |
| 12:01 | <hsivonen> | Hixie: do Microsoft and Nokia count as browser vendors who have veto on spec sections? |
| 12:05 | <hsivonen> | wow. spammers have developed networks of hundreds of inter-following spam twitter accounts |
| 12:05 | <Lachy> | hsivonen, link? |
| 12:06 | <Lachy> | I wonder how that's useful for spammers |
| 12:07 | <hsivonen> | pm'ed link to avoid driving traffic to spammers |
| 12:07 | <hsivonen> | from krijn's logs, that is |
| 12:07 | <Dashiva> | Doesn't this channel have the comment feature? |
| 12:07 | <hsivonen> | ah. right. what was it? |
| 12:08 | <Dashiva> | I don't know |
| 12:08 | <Lachy> | Dashiva, what feature are you referring to? |
| 12:08 | <Dashiva> | We could check the logs - oh wait :) |
| 12:08 | <hsivonen> | Lachy: the one developed for webapps WG |
| 12:08 | <krijnh> | #off, but it doesn't work here |
| 12:08 | <Philip`> | Dashiva: I don't think it does |
| 12:08 | <Lachy> | it's [off] but it doesn't work here |
| 12:08 | <hsivonen> | krijnh: ah. ok |
| 12:08 | <krijnh> | Of [off] |
| 12:09 | <virtuelv> | hsivonen: is it really surprising? |
| 12:09 | <hsivonen> | virtuelv: well, not really |
| 12:10 | <hsivonen> | virtuelv: I'm not surprised that there are spam accounts |
| 12:10 | <hsivonen> | virtuelv: I'm mildly surprised about them taking the time to arrange follower numbers |
| 12:11 | <virtuelv> | I'd call it "two degrees of Google" |
| 12:12 | <virtuelv> | while twitter uses nofollow, I think the ultimate purpose is to get people to create the links elsewhere |
| 12:12 | <virtuelv> | and twitter drives significant traffic to a site on its own |
| 12:14 | <hsivonen> | http://trac.webkit.org/changeset/43972 is a great example of why hasFeature() kind of APIs don't work on the Web |
| 12:15 | <Dashiva> | Isn't hasFeature pretty much relegated to Java use? |
| 12:15 | <hsivonen> | I don't know |
| 12:16 | <hsivonen> | maybe the return values should be "maybe" and "maybeNot" |
| 12:16 | <Philip`> | Has anyone ever actually used the HTML DOM in Java, and do they still? |
| 12:16 | <virtuelv> | Dashiva: hasFeature has been considered for widgets |
| 12:16 | <Philip`> | (where by "HTML DOM" I mean the parts that aren't used for generic XML) |
| 12:17 | Philip` | is unaware of what contexts it would be used in |
| 12:17 | <hsivonen> | oh. great. now Dr. Nielsen's expertise *can* be questioned :-) |
| 12:17 | <virtuelv> | because let's say you specify <feature name="http://example.com/directbraininterface">;, and configuration disallows it, you might just want to know |
| 12:17 | <Dashiva> | Philip`: Isn't hasFeature a generic part of DOM? |
| 12:18 | <virtuelv> | Dashiva: document.implementation.hasFeature, yes |
| 12:19 | <virtuelv> | DOM1 |
| 12:22 | <jgraham> | hsivonen: Maybe the return value should just be "maybe" |
| 12:23 | <jgraham> | http://grack.com/blog/2009/07/07/video-support-for-older-browsers/ |
| 12:38 | <virtuelv> | jgraham: "maybe" and "false" |
| 12:38 | <virtuelv> | or "no way in hell" |
| 12:42 | <jgraham> | virtuelv: "buggy" and "false" seems more accurate |
| 12:56 | <MikeSmith> | opera friends, please help me out |
| 12:56 | <MikeSmith> | I see http://my.opera.com/desktopteam/blog/2009/07/03/crash-logging-on-the-mac says there is a new snapshot build |
| 12:56 | <Lachy> | MikeSmith, with what? |
| 12:57 | <jgraham> | MikeSmith: ... |
| 12:57 | <MikeSmith> | Lachy, jgraham - I have opera:config#AutoUpdate|DownloadAllSnapshots set to on |
| 12:58 | <MikeSmith> | but when I do "Check for updates" |
| 12:58 | jgraham | can see where this is going and the answer is "I don't know" |
| 12:58 | <MikeSmith> | it tells me I'm running the latest version |
| 12:58 | <MikeSmith> | heh |
| 12:58 | <Lachy> | MikeSmith, you'll want to ask someone from Desktop, I think |
| 12:58 | <jgraham> | I can of course ask if Lachy or someone doesn't know |
| 12:59 | <MikeSmith> | well, if you know who does know, feel free to PM me with a name, and I'll bug them instead |
| 13:00 | <jgraham> | Lachy: Can I leave this to you? All the desktop people are in Oslo... |
| 13:00 | <Lachy> | MikeSmith, I will relay your question to #staff |
| 13:04 | <Lachy> | MikeSmith, I was given this as the response: |
| 13:04 | <Lachy> | http://my.opera.com/desktopteam/blog/2009/07/03/crash-logging-on-the-mac?startidx=150#comment9311695 |
| 13:08 | <MikeSmith> | Lachy: I see. well, please convey to Rijk (or Olli or Claudio or somebody) that perhaps it's time they might want to get around to spending some time making it a "push one button" operation |
| 13:26 | <Lachy> | MikeSmith, feel free to email them or comment on the desktop team blog about that |
| 13:27 | <Lachy> | http://www.appleinsider.com/articles/09/07/06/ogg_theora_h_264_and_the_html_5_browser_squabble.html |
| 13:29 | <hsivonen> | Lachy: oh great. Apple enthusiasts joining the 'debate' |
| 13:30 | <Lachy> | for the most part, it's a reasonable article, except for a few bogus claims about Opera near the end |
| 13:30 | <Rik|work> | I like the historical point of view |
| 13:31 | <hsivonen> | Lachy: well, up front they say " Ogg Theora, commercially abandoned nearly a decade ago, doesn't have what it takes to deliver video on the increasingly mobile web" |
| 13:31 | <Lachy> | and it seems to reject Ogg Theora as "commercially abandoned nearly a decade ago," |
| 13:31 | <Lachy> | yeah |
| 13:32 | <Lachy> | oh, and they also repeat Google's claim about it's inability to use Ogg Theora due to bandwidth issues, without mentioning that claim itself is highly questionable |
| 13:33 | <Rik|work> | they also claim that Apple can't bet on Theora because of submarine patents but H264 may have the same problems |
| 13:34 | <hsivonen> | it has an Apple press flavor where you assume that whatever Apple is doing with the iPhone is right and then construct arguments not to contradict that assumption |
| 13:34 | <hsivonen> | or explaining how great move it is to go glossy with displays |
| 13:35 | <Lachy> | Rik|work, sure, there's always a chance that h.264 has some submarine patents applying to it that haven't yet surfaced, but given that it's been out for so long with so much industry support, and that none have surfaced yet, it's becoming more and more unlikely |
| 13:36 | <hsivonen> | also, the article totally misses the point regarding patents and GPLed software produced outside the U.S. and not shipped commercially |
| 13:36 | <Lachy> | glossy displays with sufficient anti-reflective coating may not be too bad. The glossy display on my iMac has no problems |
| 13:37 | <jgraham> | Lachy: That article is so far from npov it cannot be described as "reasonable" |
| 13:37 | <Lachy> | Though, I haven't yet used a laptop with a glossy display, in enough places to judge |
| 13:42 | <hsivonen> | Lachy: my data points are that the built-in display of my MacBook sucks horribly. (fortunately, I only use it while traveling) |
| 13:43 | <hsivonen> | Lachy: and my in-store experience with the latest Cinema Display is that it is unusably reflective |
| 13:43 | <hsivonen> | which is bad, because having to shop for an HP or Lenovo display is a pain |
| 13:43 | <hsivonen> | having to read technical details and stuff |
| 13:44 | <hsivonen> | and having to make compromises on how the enclosure fits into other decor |
| 13:45 | <Lachy> | the problem with in-store experience is that it doesn't quite match real world conditions |
| 13:45 | <Philip`> | They should let you borrow monitors and take them home to test them in real world conditions |
| 13:46 | <hsivonen> | the store had a window |
| 13:46 | <hsivonen> | the real world has a window, too |
| 13:46 | <jgraham> | The real world has many windows |
| 13:46 | <hsivonen> | meanwhile, the old 30" Cinema Display was great in the same store under the same harsh conditions |
| 13:47 | <hsivonen> | harsh as in having a window! |
| 13:47 | <hsivonen> | with sunlight |
| 13:47 | <Lachy> | next time I'm in an apple store, I will take a look at their displays |
| 13:47 | <Lachy> | which should, hopefully, be some time this week, assuming they finally repair my iMac |
| 13:48 | <hsivonen> | Exception in thread "main" java.lang.UnsupportedClassVersionError: Bad version number in .class file |
| 13:48 | <hsivonen> | haven't had that problem with JDK 5 before |
| 13:48 | <hsivonen> | maybe the world has started to move on to 6 now |
| 14:11 | <Philip`> | hsivonen: Is hsivonen.iki.fi down? |
| 14:12 | <Philip`> | hsivonen: Oh, it worked now |
| 14:12 | <hsivonen> | Philip`: just really slow |
| 14:12 | <Philip`> | hsivonen: Okay |
| 14:12 | <hsivonen> | (I don't know why. It's a shared host.) |
| 14:13 | <hsivonen> | My legacy scripts depending on PyGenx are inconvenient to move to another host |
| 14:14 | <hsivonen> | lesson: never depend of C-backed python modules that aren't available via apt |
| 14:21 | <hsivonen> | http://www.ibiblio.org/jwrobie/blog/?p=120 (via Julian) |
| 14:25 | <Philip`> | "I would like to see the XHTML Working Group finish their work, and I would like to see the result used as the XML serialization of HTML5" |
| 14:26 | <hsivonen> | http://www.zeldman.com/2009/07/07/in-defense-of-web-developers/ |
| 14:29 | <Rik|work> | I don't understand why zeldman loves xhtml so much |
| 14:29 | <Rik|work> | what does it provide over html4 ? |
| 14:30 | <Philip`> | It encourages prettier syntax |
| 14:32 | <Philip`> | Is there some web server or proxy that can be configured to return 503 if there are more than n simultaneous requests to the set of URIs matching some pattern? |
| 14:33 | <hsivonen> | right now I have trouble saying anything that wouldn't be snarky, and I've already been blamed for being snarky about XHTML-as-text/html |
| 14:34 | <hsivonen> | but it seems it's more in defense of gurus who promoted XHTML than in defense of the developers who followed them |
| 14:37 | <Philip`> | It seems in defense of the tactics of picking a technology as the basis of a marketing platform that has little to do with the technology itself |
| 14:39 | <karlcow> | welcome to the real world, aka not browsers, but people writing xhtml everywhere :) |
| 14:39 | <Philip`> | i.e. using XHTML to promote clean semantic markup, similar to using AJAX to promote asynchronous updates, and similar to using HTML5 to promote web application APIs |
| 14:41 | <Philip`> | "switching from HTML to XHTML was the perfect catharsis to leave behind not only the old syntax, but more important the old habits and practices of building a webpage" |
| 14:41 | <karlcow> | Philip`: it is a discussion I already had with hsivonen in october 2007. XHTML exists. It is used. It is not served properly and not parsed by browsers as such, but people are writing xhtml in their editors. and it's what matters to them. |
| 14:41 | <karlcow> | Philip`: yep, it helped people to look at the language in a different way. |
| 14:42 | <hsivonen> | the popularity of XHTML 1.0 Transitional doctype suggests a different thing |
| 14:42 | <karlcow> | xhtml 1.0 transitional suggests: target ;) |
| 14:43 | <karlcow> | But working again in a Web agency after a few years, I'm listening the people around me. |
| 14:44 | <Philip`> | "[XHTML] taught a generation of developers what good code should or could look like. It was something that was always possible in HTML4 but XHTML1 helped train people to be more thoughtful and careful with their code. And in that sense, I think it was a success." |
| 14:44 | <Lachy> | It seems that Zeldman's point is that promoting XHTML was good because he used it as a hook for the idea of promoting better semantics and structure, but that unfortuately gave rise to the myth that XHTML is somehow more semantic than HTML. |
| 14:44 | <Philip`> | Lachy: Myths aren't necessarily bad |
| 14:44 | <karlcow> | Philip`: indeed |
| 14:44 | <Lachy> | they are when people use those myths to then promote XHTML as being better than HTML |
| 14:45 | <Philip`> | Lachy: Why is that bad? |
| 14:45 | <jgraham> | That argument sounds remarkably close to "people are too stupid to understand the full story so we have to present half truths instead" |
| 14:46 | <Lachy> | because it attaches a stigma to HTML portraying it as being less semantic, and thus not a good choice to use |
| 14:46 | <hsivonen> | fwiw, my much more snarky "wannabe" post was set in motion by an XHTML believer complaining about OpenOffice.org's HTML output having upper-case tags |
| 14:46 | <hsivonen> | if you look at my latest, you can tell which tags are generated by OO.o and which ones are hand-written in TextWrangler |
| 14:46 | <hsivonen> | which is probably horribly untidy |
| 14:46 | karlcow | wonders in the polyglot section if I should write down the requirements to have real documents working in the two world. It will be a very tiny subset but that would be interesting at least to set that. |
| 14:47 | <Lachy> | using uppercase tag names is so last century. Lowercase tags are obviously superior |
| 14:47 | <Lachy> | karlcow, in which document are you writing that? |
| 14:48 | <hsivonen> | Lachy: it's probably also very uncool to use OO.o to write HTML |
| 14:48 | <Philip`> | jgraham: That sounds like marketing, which is what this is all about |
| 14:48 | <karlcow> | Lachy: http://www.la-grange.net/2009/07/05/html5-xhtml5/ I have not sent it to the WG, but MikeSmith suggested a few fixes. |
| 14:49 | <hsivonen> | karlcow: I suggest not having a polyglot section |
| 14:49 | <Lachy> | Using OOo (or MS Word) as a tool to write documents intended to be HTML is bad. But being able to export documents written in those formats to HTML is a good thing, if it gets implemented well |
| 14:49 | <karlcow> | hsivonen: why? |
| 14:50 | <Philip`> | Lachy: I think I'm interested in what the actual real harmful effects of promoting XHTML over HTML are, when it's used as part of the semantic markup marketing platform, because hopefully we could try to minimise similar effects when using HTML5 as part of the rich web application marketing platform |
| 14:50 | <hsivonen> | karlcow: if a polyglot section goes up on w3.org, we'll have a generation of gurus telling developers to carefully write polyglot documents without the developers quite understanding why |
| 14:50 | <karlcow> | I already had one comment by mail asking me to develop it further. |
| 14:50 | <jgraham> | Philip`: It seems like people ho are paid to work with HTML should have a better-than-marketing level knowledge of the technology. Being more informed allows them to make more sensible choices about their code |
| 14:50 | <jgraham> | *who |
| 14:50 | <karlcow> | hsivonen: head in the sand? |
| 14:51 | <hsivonen> | karlcow: I didn't put my head in sand but instead suggested outright that people not bother with polyglot documents |
| 14:51 | <jgraham> | karlcow: Taking private feedback :) |
| 14:51 | <karlcow> | jgraham: your "should" is not aligned with real world. What I said at the beginning, welcome back to real world |
| 14:52 | <Philip`> | karlcow: In your definition of polyglot documents, should it include a requirement that the syntax is processed in equivalent ways in both HTML and XHTML? e.g. <textarea>[newline]stuff</textarea> is very different in HTML vs XHTML, so it can't be safely used in polyglot documents |
| 14:52 | <Lachy> | oh, nice. I have plans to discuss a lot of those issues in the HTML 5 Reference. Maybe, when you're done, I could incorporate some of that into it |
| 14:52 | <karlcow> | hsivonen: the fact that you can write polyglot documents, makes it a possibility that people will explore. :) |
| 14:53 | <jgraham> | karlcow: So in your real world, the gurus should make up whatever story they think will placate the masses of web developers without worrying about the side effects of people being underinformed? |
| 14:53 | <karlcow> | mouahaha |
| 14:54 | <hsivonen> | karlcow: I can already foresee new myths being developed on top of the polyglot concept |
| 14:54 | <karlcow> | we turn into rhetorical manipulation. :) going back to work ;) sorry chaps |
| 14:54 | <Lachy> | karlcow, http://wiki.whatwg.org/wiki/HTML_vs._XHTML lists most of the syntactic differences. I think there was a few small things missing, but I can't recall what they were |
| 14:54 | <karlcow> | Lachy: look at the top of the document |
| 14:54 | <karlcow> | I cite this document |
| 14:54 | <karlcow> | :) |
| 14:54 | <Lachy> | oh |
| 14:55 | <Lachy> | I just saw the table listing syntax differences, and it seemed a little short |
| 14:55 | <karlcow> | Previous Version: |
| 14:55 | <karlcow> | http://www.w3.org/QA/2008/01/html5-is-html-and-xml |
| 14:55 | <karlcow> | http://wiki.whatwg.org/wiki/HTML_vs._XHTML |
| 14:55 | <Lachy> | maybe it's not. I'd need more than a cursory glance to know for sure |
| 14:57 | <Lachy> | karlcow, in that table, for some of the things like Unquoted Attributes and Attribute minimisation, the cells in the XHTML column say "not allowed" accompanied by an example. At first, it wasn't clear to me what about the example wasn't allowed. |
| 14:57 | <Lachy> | Then I realised that it was meant as an example of what is allowed in XHTML. It would be nice to find a way to make that clearer |
| 14:59 | <Philip`> | jgraham: Seems like division of labour - some people spend years learning all the subtleties of markup syntaxes, and then provide recommendations which can be followed by people who don't want to waste time repeating that whole learning exercise needlessly, and those recommendations necessarily have to summarise and condense and ignore some issues so they can be understood more readily |
| 14:59 | <Lachy> | and I object to you using <p class=sloppy> as an HTML example of unquoted attributes. Saying it's sloppy seems to convey that it is somehow bad, even though it's a perfectly valid technique |
| 15:00 | <Philip`> | (and hopefully the people providing the recommendations can minimise the harm caused by misinterpretation of the truth) |
| 15:00 | <Lachy> | I would prefer it if documents like this stayed more neutral on the debate about which syntax is better to use, and let authors make a more informed choice |
| 15:01 | <Philip`> | (and those experts would still be expected to justify their arguments in detail if someone asked them about it, so they're not expecting people to blindly rely on their expert status) |
| 15:01 | <jgraham> | Philip`: There is truth in that. In this case I dispute that "XHTML good, HTML bad" (which was roughly the message that many people got) is at the right level of simplicity |
| 15:03 | <Lachy> | the right message that should be conveyed to web developers is that HTML and XHTML are optimised for different requirements, and that authors should choose the one most appropriate for their needs |
| 15:03 | <jgraham> | If people think that quoating their attribute values and closing their tags helps them to write maintainable code then that's a decision that they should be free to make. We should have tools to help them enforce those coding standards. We shouldn't conflate that with pseudo-XHTML |
| 15:04 | <Lachy> | In most cases, given the current browser market, the right choice is generally HTML. But there are legitimate uses for XHTML when done right |
| 15:04 | <Lachy> | jgraham, indeed |
| 15:05 | <Philip`> | jgraham: So it sounds like you agree with the general idea of lying to simple-minded developers, and just have a different view (than people promoting only-XHTML-is-semantic) of whether the harmful consequences of one particular simple lie are unpleasantly (and hopefully avoidably) high |
| 15:07 | <jgraham> | Philip`: I agree that it shouldn't be necessary to know the full detail of the whole language to get started. I wouldn't haracterise that as "lying" |
| 15:07 | <Philip`> | jgraham: "Not telling the truth", then? :-) |
| 15:08 | <jgraham> | (I also think that the language should be designed in such a way that it is easy to learn the full language) |
| 15:09 | <jgraham> | Philip`: Not really. "Not trying to present the full complexity of everything all at once but making it avaliable when necessary" sould be more like it |
| 15:09 | <jgraham> | *would |
| 15:11 | <Philip`> | jgraham: (The full language is a simplification of reality devised by experts, who e.g. understand all the details of what <br /> means in SGML and XML and how it's implemented in browsers and why it was promoted instead of <br/> and how common all these things are, and choose to hide all those details and just recommend "you must do this or this") |
| 15:12 | <Philip`> | jgraham: (so "learn[ing] the full language" just seems like another instance of following experts' recommendations without really understanding them) |
| 15:13 | <Philip`> | jgraham: (and doesn't seem fundamentally different to other instances of the same thing, at higher levels than the markup language syntax) |
| 15:14 | <Philip`> | jgraham: (I hope you don't ask me to clarify my argument since I've now forgotten what my argument was or whether I even had one) |
| 15:17 | <jgraham> | Philip`: It seems like a big leap to go from "practitioners of the language should be knowledgeable enough to to make informed choices within the constraints imposed by the conformance criteria" to "there should be no conformance criteria since they represent arbitary limitations" |
| 15:17 | <Philip`> | jgraham: It does |
| 15:17 | <Philip`> | jgraham: Who is suggesting the latter? |
| 15:18 | <Philip`> | (...in this discussion, i.e. ignoring sayrer :-p ) |
| 15:22 | <jgraham> | Philip`: It seemed like the logical conclusion of your argument that "learn[ing] the full language" just seems like another instance of following experts' recommendations without really understanding them |
| 15:22 | <Lachy> | I can't believe how many of my twitter followers have joined in the "campaign" to get me on the Australian version of Beauty and the Geek, despite me not being there to participate :-/ |
| 15:22 | <Philip`> | jgraham: My argument is that people should be able to follow experts' recommendations without really understanding them, and the experts should provide recommendations to optimise that behaviour |
| 15:24 | <Philip`> | e.g. by providing conformance criteria for the language, and providing simple marketing terms like "XHTML" or "AJAX" or "HTML5" to encourage people to adopt a whole platformful of goodness (while trying to minimise the harm from the simplifications) |
| 15:26 | <jgraham> | I agree with the first part, less so with the second |
| 15:26 | <jgraham> | (and of course with the thrd) |
| 15:28 | <beowulf> | web developers don't need to understand the sticks with which they beat each other with, but it helps |
| 15:30 | <nikto> | and then comes the Black Swan, and your 'web developers' that just follow guidelines and expert counseling are faced to understand the 'display' property and it's values, and how this affects rendering of DIVs, SPANs, and everything |
| 15:30 | <nikto> | so.. a nice learning path for the very deep details should be offered to them, imho |
| 15:31 | <nikto> | not to mention the difference it makes when in quirks mode, or when in stric mode, etc. |
| 15:32 | <nikto> | the most efficient dev teams I've met had zero 'web developers'. They had developers, and graphic designers at the other side, but no glue in between |
| 15:37 | <hsivonen> | jgraham: saying that non-conforming parts aren't part of the "full language" for learning is a bit questionable |
| 15:51 | <hsivonen> | adactio: fwiw, I didn't mean to "belittle" anyone for their preferences. What I have an issue with is telling people to jump through hoops due to bogus rationale. |
| 15:51 | <hsivonen> | (HTML5 conformance has some of that, too, IMO) |
| 15:52 | <adactio> | I don't recall anyone ever saying that XHTML was better than HTML; just different. |
| 15:52 | <hsivonen> | adactio: oh, there have been people who've complained or ridiculed stuff for being valid HTML 4 but not XHTML |
| 15:52 | <adactio> | Citation needed. |
| 15:53 | <takkaria> | I'm no citation, but it was certainly a popular meme |
| 15:53 | <gsnedders> | +1 |
| 15:53 | <karlcow> | Lachy: "At first, it wasn't clear to me what about the example wasn't allowed." Yes I agree, I will improve that. |
| 15:54 | <jgraham> | http://www.webstandards.org/learn/articles/askw3c/oct2003/ |
| 15:54 | <jgraham> | (that article is quite balanced but still says things like "The syntax proposed by XHTML 1.0 has several important benefits") |
| 15:55 | <hsivonen> | adactio: I can't find the post that inspired me to write http://hsivonen.iki.fi/wannabe/ anymore, but it was someone saying that OpenOffice.org was wrong to emit uppercase tags |
| 15:55 | <hsivonen> | (and I do agree that lower case is nicer) |
| 15:55 | <hsivonen> | I guess these should be cataloged just like Namespace confusion in order to be able to respond to "citation needed" when prompted |
| 15:56 | <adactio> | hsivonen: But when you lump everyone who prefers XHTML syntax into a group labelled "HTML haters", you are guilty of equal FUD as the minority who claimed that XHTML was somehow better than HTML. |
| 15:58 | <jacobolus> | too bad mark pilgrim isn't around. his little poem is great :) |
| 15:58 | <adactio> | If I say "I like XHTML syntax", that is not the same thing as saying "XHTML is superior to HTML." Just because someone else who prefers XHTML syntax has made that erroneous claim doesn't mean we share that belief. |
| 15:58 | <hsivonen> | adactio: I don't think everyone who prefers XHTML syntax is an HTML hater. I just think that using XHTML-as-text/html is a marketing thing, and has no technical basis in how it gets consumed (as has been claimed; passive voice, no citation) |
| 15:59 | <jacobolus> | adactio: people say "I like XHTML because ..." and then their reasons are simply wrong |
| 15:59 | karlcow | still fixing http://www.la-grange.net/2009/06/html5-xhtml5/ |
| 15:59 | <karlcow> | I'll write more tonight when I have time |
| 15:59 | <hsivonen> | adactio: no, but "I like XHTML syntax" is different from going on to give utterly bogus technical reasons |
| 16:00 | <adactio> | hsivonen: in your Q & A you made it sound like anyone who uses XHTML syntax was doing so out of a desire to feel superior. |
| 16:00 | <adactio> | jacobolus: http://simonwillison.net/2008/Jun/5/cafes/#c39300 |
| 16:00 | <hsivonen> | adactio: I believe some are, but I don't think everyone is |
| 16:00 | <hsivonen> | adactio: I think many are just doing what gurus have told them to do |
| 16:00 | <jacobolus> | s/many/most/ |
| 16:01 | <adactio> | hsivonen: that distinction was not clear in your article. You tarred anyone who chooses to use XHTML with the same brush. |
| 16:01 | <hsivonen> | adactio: I concede that the "new and shiny" part was snarky. I think the "marketing" part was correct, though. |
| 16:02 | <jacobolus> | adactio: sure, and I can make my Python code conform to PEP 8, and I can write C code using some style guideline, K&R, say |
| 16:02 | <jacobolus> | adactio: there's no need to call that 'xhtml' though |
| 16:03 | <adactio> | Look, let me give you my own personal reasons why I have historically used XHTML syntax: |
| 16:03 | <adactio> | The extra bit of strictness makes the validator a more powerful tool... |
| 16:04 | <jacobolus> | really depends whether your validator is doing proper parsing |
| 16:04 | <adactio> | Just recently, a colleague who was using HTML 4, couldn't find a problem because it was being caused by a missing closing </p>. |
| 16:04 | <hsivonen> | adactio: it also makes the W3C validator not report some parsing problems |
| 16:04 | <adactio> | I'd quite happily use HTML 4 if there were some kind of "lint" options available that enforced the strictness of XHTML syntax. |
| 16:04 | <hsivonen> | adactio: a better solution is to make an HTML validator warn about implied tags |
| 16:05 | <hsivonen> | (yeah, I know, vaporware feature at this point) |
| 16:05 | <adactio> | In the absence of such a lint tool, XHTML 1 suits me quite nicely. |
| 16:05 | <adactio> | And I'll call it XHTML (even though I'm serving it as text/html) because there isn't a better term. |
| 16:05 | <hsivonen> | adactio: what about <script src='...'/>? |
| 16:05 | <jgraham> | If this whole thing comes down to "people want a HTML lint" then it seems like we should give up talking and just write the damn lint |
| 16:05 | <adactio> | Yes please. :-) |
| 16:05 | <hsivonen> | adactio: XHTML validation trades in a different set of silently passed problems |
| 16:06 | Philip` | has noticed several people in blog comments saying they like XHTML's strict syntax, so presumably they'd be happy with a lint-like tool too |
| 16:06 | <adactio> | Yeah, I know it's not ideal but I'm happy to live with those issues for the sake of a stricter validation check. |
| 16:06 | <hsivonen> | adactio: HTML5 validation overall is stricter than an XHTML 1.0 DTD check |
| 16:06 | <hsivonen> | adactio: although not for implied close tags |
| 16:06 | <adactio> | From my own anecdotal evidence, chatting with colleagues, an HTML lint tool would be *very* welcome. |
| 16:07 | <Philip`> | Require explicit end tags, lowercase element/attribute names, quoted attribute values, maybe still allow valueless attributes |
| 16:07 | <jgraham> | (the sensible place to write the lint is as part of the validator of course which really means that I'm volunteering hsivonen. But it wouldn't be impossible to make something that didn't validate but did waarn about unquoted attributes and so on) |
| 16:07 | <adactio> | Something like Douglas Crockford's JSLint: cruel but fair. :-) |
| 16:07 | <adactio> | I too nominate hsivonen to build this. ;-) |
| 16:07 | Philip` | thinks JSLint is a bit excessive when it complains about semicolons after function bodies, and therefore gave up attempting to use it |
| 16:08 | <adactio> | Seriously though, it would be sooooo useful! |
| 16:08 | <hsivonen> | adactio: I very much see the demand for some XHTML-like lint options in V.nu. I plead having been busy. |
| 16:08 | <adactio> | hsivonen: understandably busy! |
| 16:09 | <adactio> | hsivonen: and don't get me wrong; validator.nu kicks ass. |
| 16:09 | <hsivonen> | anyway, if I had had someone vet my post ahead of publishing, I would have toned down the snark |
| 16:09 | <Philip`> | http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Jul/0013.html |
| 16:09 | <hsivonen> | but by now it's been quoted all over, so there's no point hiding it |
| 16:09 | <hsivonen> | at best I could put a strike-over there and add a note that it distracts from the other points |
| 16:10 | <hsivonen> | although it seems controversial stuff gets more incoming links :-/ |
| 16:10 | <adactio> | hsivonen: ah, the Digg school of online publishing. ;-) |
| 16:10 | <hsivonen> | adactio: it was unintentional |
| 16:11 | <hsivonen> | but I observe that saying stuff that XHTML authors take offense with gets more links faster than just-the-facts like /doctype/ |
| 16:12 | <hsivonen> | Philip`: ah. "personal preferences" rides again. sigh. |
| 16:12 | <Philip`> | hsivonen: I guess you are unintentionally discovering the effectiveness of trolling :-) |
| 16:24 | <hsivonen> | :-( http://www.zeldman.com/2009/07/07/in-defense-of-web-developers/ spreads misinformation about the history of quirks mode and HTML 4 |
| 16:24 | <Lachy> | hsivonen, yeah, I noticed that too |
| 16:24 | <hsivonen> | "A few who became disenchanted with XHTML early retreated to HTML 4, and as browsers stopped going into Quirksmode in its presence, valid, structural HTML 4 became a reasonable option again." |
| 16:27 | <takkaria> | I don't see Zeldman or the comments pointing out that there is an XHTML5 too |
| 16:28 | <Lachy> | I'm not sure if Zeldman is aware of it. Though, I'd like to give him the benefit of the doubt and assume he is, and that he just didn't mention it. |
| 16:29 | gsnedders | groans |
| 16:29 | <Lachy> | gsnedders, what are you groaning about? |
| 16:30 | <gsnedders> | #svn can't even say whether I can rely upon all SVN checkouts of one revision being byte-for-byte equivalent |
| 16:30 | <gsnedders> | s/equivalent/identical/ |
| 16:30 | <Lachy> | wtf? |
| 16:30 | <gsnedders> | svn:eol-style makes it fun for text files. |
| 16:30 | <takkaria> | as in, given two SVN checkouts of the same revision, will they be exactly identical |
| 16:30 | <takkaria> | ? |
| 16:30 | <gsnedders> | Yeah. |
| 16:30 | <Lachy> | if they're not, then wouldn't that be a bug in SVN? |
| 16:31 | <gsnedders> | No, it's a feature. |
| 16:31 | <Lachy> | ??? |
| 16:31 | <takkaria> | except for svn:eol-style, I can't think of any reason why they wouldn't be |
| 16:31 | <gsnedders> | takkaria: svn:eol-style is alone a big enough problem for me. |
| 16:31 | <Lachy> | does svn:eol-style control whether new lines in text files are CR/CRLF/LF ? |
| 16:32 | <takkaria> | Lachy: yeah. and there's a 'native' setting that checks them out as whatever is platform-native |
| 16:32 | <Philip`> | If svn:eol-style=native then it automatically converts between the repository and working copies |
| 16:32 | <takkaria> | gsnedders: I can't see why, unless your two checkouts are on different machines with different OSes |
| 16:32 | <Lachy> | grrr. What stupid moron copied that feature from FTP ASCII transfers. |
| 16:32 | <matt_mastracci> | Hey all... I started working on a <video> compatibility shim for older browsers: http://code.google.com/p/video4all/ It works in virtually every browser I've tested so far, with the exception of Windows Safari w/o QuickTime installed. |
| 16:33 | <Philip`> | Lachy: It's very useful for projects that are developed partly with Windows tools and partly with Linux/OSX tools |
| 16:33 | <Lachy> | Philip`, there are developer tools available for Windows that support LF line endings |
| 16:33 | <Philip`> | e.g. you set all .cpp files to native and then everyone can use their favourite text editor and you won't get spurious diffs from line-ending changes |
| 16:33 | <Lachy> | Unless of course, you're using Notepad, in which case, you have bigger issues. |
| 16:33 | <Philip`> | Lachy: It's more of a pain to find and configure those tools than to just make SVN do the right thing automatically |
| 16:34 | <Lachy> | it's just a pain that Microsoft continues to perpetuate that CRLF as default nonsense |
| 16:34 | <Philip`> | It'd be even more of a pain if they changed it |
| 16:35 | <Lachy> | they would need a migration strategy, which would first involve ensure all Microsoft software support LF |
| 16:36 | <Philip`> | and secondly would involve ensuring a zillion pieces of third-party software, much of it written in the 90s with no source code available and being used as critical pieces of business infrastructure, all support LF? |
| 16:37 | <Lachy> | if they had done that with Notepad in XP (even if CRLF was still the default then), and encouraged other vendors to begin adopting LF as the default in more recent years, the migration could go smoothly |
| 16:37 | <takkaria> | this comment wins an award: http://www.zeldman.com/2009/07/07/in-defense-of-web-developers/#comment-44177 |
| 16:37 | <Philip`> | Lachy: That would be great until the first time someone edited a file in Notepad and then opened it in some other program and it stopped working, and they'd blame XP for being stupid and buggy and would go back to Win98 |
| 16:38 | <Philip`> | Oh, I suppose you might mean supporting LF only on input, not output |
| 16:38 | <Lachy> | yes |
| 16:38 | <Lachy> | at least for XP |
| 16:39 | <Philip`> | Since Notepad is just a standard text control, it would involve either changing the standard text control (which would break a zillion other applications) or rewriting Notepad to not use that control (in which case it would be Wordpad) |
| 16:39 | <Lachy> | If they'd done that then, now with Windows 7 given that pre-XP users are negligable in most places (especially work environments), they could start switching to LF as default for output |
| 16:40 | <Lachy> | there's no reason the standard text control shouldn't support LF on input |
| 16:40 | <Philip`> | Lachy: That would be great until the first time someone edited a file in Notepad in Windows 7 and then opened it in some other program and it stopped working, and they'd blame XP for being stupid and buggy and would go back to Vista |
| 16:40 | <Philip`> | s/XP/Windows 7/ |
| 16:40 | <gsnedders> | takkaria: On the same machine they'll be identical, sure. |
| 16:42 | <gsnedders> | Lachy: Win 2000 is not non-negilable |
| 16:43 | <Philip`> | gsnedders: On different machines you might have different versions of SVN with different working copy formats |
| 16:44 | <Lachy> | gsnedders, citation? |
| 16:44 | <Lachy> | (I'm aware of the fact that I should provide a citation for my claim too) |
| 16:44 | <gsnedders> | Lachy: See my old school. |
| 16:45 | <Lachy> | wow, still running Win2k? Why? |
| 16:45 | gsnedders | shrugs |
| 16:45 | <gsnedders> | Why pay for XP? |
| 16:46 | <Lachy> | they will need to upgrade by mid next year, since Win2k support ends, meaning no more security fixes |
| 16:47 | Philip` | prefers to grind skeleton warriors for XP, so he doesn't have to pay anything |
| 16:47 | <Lachy> | at least, no more free fixes, if I understand how Microsoft support works |
| 16:47 | <Lachy> | Philip`, what? |
| 16:48 | <Philip`> | Lachy: Just a pointless irrelevant MMORPG reference |
| 16:59 | <hsivonen> | http://www.zeldman.com/2009/07/07/in-defense-of-web-developers/#comment-44166 |
| 16:59 | <hsivonen> | "To be fair to Hixie, unlike Henri and Mark Pilgrim, he hasn’t been sowing deliberate seeds of confusion since the XHTML 2 announcement" |
| 16:59 | <hsivonen> | umm. what? |
| 17:01 | <adactio> | hsivonen: a reference to "new shiny" and "the house that Jeffrey built" respectively. |
| 17:02 | Philip` | wonders why it takes forever for zeldman.com pages to finish loading and scroll down to the right comment |
| 17:02 | beowulf | wishes hsivonen's font's would load quicker... |
| 17:02 | <beowulf> | grrr, fonts |
| 17:03 | <hsivonen> | beowulf: Firefox and Opera do the right thing. |
| 17:04 | <beowulf> | hsivonen: yeah, i noticed |
| 17:04 | <Rik|work> | oh, geolocation api last call |
| 17:04 | <takkaria> | that was fast |
| 17:04 | <hsivonen> | adactio: if I've sown confusion, it hasn't been deliberate. on the contrary, I've deliberately tried to unconfuse people who've been confused technically as a side effect of using "XHTML" as a social marketing platform |
| 17:05 | <Rik|work> | anybody proposed a <input type=coordinates> for HTML5 ? |
| 17:07 | <adactio> | hsivonen: understood. And for the most part, your post was just the facts ...but the bit where you (mis)characterised everyone who ever chose to author in XHTML, while clear, was an opinion, snarkily delivered. It sits awkwardly amongst the rest of the Q & A. |
| 17:07 | <Lachy> | "The W3C should stop writing new specifications, not only do they take so damn long, but they take years longer before they’re fully supported to a level where web professionals can use them. |
| 17:07 | <Lachy> | XHTML 1 works, leave it be. Go improve SVG, X3D, SMIL, or anything you want – but if it ain’t broke, don’t fix it – because fixing it WILL break it." |
| 17:07 | <Lachy> | from the comment directly after that from Jeremy Keith linked above |
| 17:08 | <Philip`> | Rik|work: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-November/017479.html |
| 17:08 | <hsivonen> | adactio: I added a note to the post. I don't know if it makes it better or worse, but editing it away at this point wouldn't be good |
| 17:09 | <eighty4> | gsnedders: still at work? |
| 17:09 | <Rik|work> | Philip`: I believe there is now a need since applications have to rely on javascript |
| 17:10 | <Philip`> | Rik|work: I guess it should be raised on the list again if the situation has changed since then |
| 17:10 | <adactio> | hsivonen: I appreciate the clarification. |
| 17:13 | <Rik|work> | wow, a french news site just titled "w3c abandon xhtml and audio and video elements |
| 17:14 | <hsivonen> | Rik|work: yay for reporting |
| 17:15 | <Rik|work> | can't they read emails ? |
| 17:17 | <gsnedders> | eighty4: yeah, going home now though |
| 17:19 | <gsnedders> | eighty4: I'll probably drop in to ICA on the way home, but I'll be back by 7 — could you bring a screwdriver like you said you could? |
| 17:19 | <eighty4> | gsnedders: will do |
| 17:19 | <eighty4> | ok if I'm outside you 19? |
| 17:20 | <takkaria> | http://twitter.com/diveintomark/status/2515353673 |
| 17:20 | <gsnedders> | eighty4: yeah |
| 17:22 | <jacobolus> | :D |
| 18:02 | gsnedders | guesses eighty4 was wrong at working out how long it would take to get here |
| 18:48 | <gsnedders> | jgraham: You will be pleased to know, that with eighty4's help, I now have washing on. |
| 18:52 | <jgraham> | gsnedders: Just in time |
| 19:06 | <eighty4> | gsnedders: this town is way to hard to find in |
| 19:06 | <gsnedders> | eighty4: :) |
| 19:26 | <gavin_> | Hixie: is it possible for video.currentSrc to be empty while video.src isn't, per spec? |
| 19:36 | <Hixie> | yes |
| 19:37 | <Hixie> | at least, i think the answer is yes not having checked to see if it actually is yes |
| 19:40 | <gavin_> | good enough for me, thanks :) |
| 22:09 | <Hixie> | aw, hsivonen didn't reply to mjs' idea |
| 22:29 | <othermaciej> | were you hoping he would? |
| 22:30 | <othermaciej> | I guess his opinion on it is pretty relevant |
| 22:33 | <Hixie> | i asked him to :-) |
| 22:33 | <Hixie> | i'm waiting for his opinion before replying myself |
| 22:42 | <takkaria> | mjs's idea? |
| 22:44 | <Hixie> | introducing a level of conformance between "conforming" and "down-played errors" that shows a validator warning but is still conforming |
| 22:44 | <takkaria> | I guess that makes sense |
| 22:45 | <gavin_> | I don't see any downsides |
| 22:46 | <Hixie> | personally i'm not a big fan of new conformance levels |
| 22:46 | <Hixie> | we've already got three |
| 22:47 | <Hixie> | and generally i don't see the point |
| 22:47 | <Hixie> | Transitional showed that people don't hesitate to use the low-level ones |
| 22:48 | <takkaria> | what stuff is downplayed errors atm? presentational guff? |
| 22:48 | <Hixie> | and XHTML has shown that other people try to be as strict as possible regardless of whether it's ok or not |
| 22:48 | <Hixie> | so it doesn't really have the desired effect |
| 22:49 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/#conformance-checkers-0 |
| 22:49 | <Hixie> | takkaria: ^ |
| 22:52 | <takkaria> | mm |
| 22:52 | <takkaria> | I don't know what I'd do in this situation |
| 22:52 | <takkaria> | four levels of conformance seems a little silly |
| 22:53 | <gavin_> | the point is that peope don't like being told that they are non-conformant |
| 22:53 | <gavin_> | yet they might like being told that they're wrong |
| 22:53 | <gavin_> | and since the distinction doesn't really matter in practice, you might as well accomodate them |
| 22:54 | <Hixie> | i think i might be more happy to making the others downplayed errors into this "warning" class than introducing a new class altogether |
| 22:54 | <takkaria> | I'm just worried about falling into the HTML4 Transitional trap |
| 22:54 | <Hixie> | but anyway, i'd like hsivonen's input on the topic |
| 22:54 | <Hixie> | since right now, he's the main one this affects |
| 22:54 | <takkaria> | Hixie: yeah, I was thinking that might be an idea |
| 22:55 | <takkaria> | probably make people happier about profile="" too |
| 22:55 | <Hixie> | yeah i'm really worried about doing another transitional |
| 22:55 | <takkaria> | (though I'm not sure that's necessarily a good thing. :)) |
| 22:55 | <Hixie> | well that's the thing |
| 22:55 | <Hixie> | profile="" and summary="" are very similar in this respect. They make people think they're doing something useful while really they are only wasting their time, even though in theory it could have been useful |
| 22:55 | <Hixie> | s/it/they/ |
| 23:09 | <othermaciej> | Hixie: I think making downplayed errors into warnings (that must be reported but don't break conformance) would be a good change and one that hsivonen could probably go along with |
| 23:10 | <Hixie> | othermaciej: i'm worried about making teh HTML4 Transitional mistake. |
| 23:10 | <Lachy> | Hixie, profile is worse than summary in that respect. The usefulness of summary is dependent upon the competence of the author writing it |
| 23:10 | <gavin_> | what were the consequences of "the HTML4 transitional mistake"? |
| 23:10 | <hober> | right, whereas the usefulness of profile is zero, regardless of the competence of the author using it. :) |
| 23:10 | <Hixie> | Lachy: yeah, arguably profile is worse, but then arguably summary is worse because when the author inevitably screws up, he hurts more users. |
| 23:10 | gavin_ | doesn't really know the history |
| 23:11 | <Hixie> | gavin_: people never transitioned. |
| 23:11 | <Lachy> | whereas the usefulness of profile is dependent upon a majority of authors getting it right, not just the one who chooses to use it |
| 23:11 | <Hixie> | gavin_: basically, people went to the lowest common denominator |
| 23:11 | <gavin_> | how is that relevant to this case? you're afraid people won't transition to the better markup unless it's a conformance error? |
| 23:12 | <Hixie> | gavin_: i guess we might avoid that if hsivonen makes these downplayed errors into warnings that are always present, and doesn't just say "yup you're valid" |
| 23:12 | <Hixie> | gavin_: yeah, more or less |
| 23:12 | <Lachy> | I know. summary isn't particularly conducive to people getting it right |
| 23:12 | <Hixie> | Lachy: arguably summary="" is useless because too many people have gotten it wrong already. :-) |
| 23:12 | <gavin_> | seems like we're mostly debating semantics :) |
| 23:13 | <Hixie> | worse than semantics |
| 23:13 | <Hixie> | psychology :-/ |
| 23:13 | <gavin_> | maybe the validator can call them "conformant errors" |
| 23:13 | <Hixie> | hah |
| 23:14 | <Hixie> | othermaciej: so you'd be ok with the proposal i described, if we just made the downplayed errors into frowned-upon-but-ok? i.e. changed from "not serious errors" to "discouraged non-errors"? |
| 23:14 | <Lachy> | Hixie, that depends on the ability of ATs being able to reliably filter out useless values, which we've been told they do, but been given no details about how |
| 23:14 | <Hixie> | Lachy: yeah, that information would definitely help |
| 23:14 | <Lachy> | nor how reliable it is |
| 23:14 | <Hixie> | we really should do usability testing of this |
| 23:14 | <Hixie> | it's hard to do usability testing for AT users, though |
| 23:14 | <Philip`> | Maybe the validator should do like what the Yahoo and Google page speed tools do, giving each message a red/yellow/green severity and showing them in decreasing order of severity, so you can see which issues are the most worth fixing |
| 23:15 | <Hixie> | because you have to go to their setups, you can't just have them come in |
| 23:15 | <Lachy> | if they have laptops, then you could have them come in with them |
| 23:15 | <othermaciej> | Hixie: I think that would be a good improvement beyond even the summary="" issue |
| 23:15 | <Hixie> | from what i've seen of usability studies of AT users here at google, they don't. at least not the ones we found for our studies. |
| 23:16 | <othermaciej> | Hixie: and yes, I would be in favor of that proposal |
| 23:17 | <Hixie> | othermaciej: ok. i'll see what hsivonen thinks, and will see if i can find a way to phrase it that satisfies both the desire to keep these things as second-class citizens, and the desire to allow them to some extent. |
| 23:17 | <Lachy> | how is that different from the downplayed errors we already have in the spec? |
| 23:18 | <othermaciej> | a document containing a downplayed error is nonconforming |
| 23:19 | <Lachy> | so maybe we should have conformance levels, conceptually like the WCAG accessibility levels? |
| 23:19 | <Philip`> | Sounds like it's going to be horrendously confusing to someone who just wants to be told whether their page is okay or not |
| 23:19 | <Lachy> | but then, that's getting closer to the idea of Strict/Transitional conformance |
| 23:22 | <gavin_> | Philip`: "good", "ok, with some problems" and "bad" doesn't seem that confusing |
| 23:23 | <gavin_> | but even with those three buckets there are going to be arguments as to what goes where |
| 23:25 | <Lachy> | yeah, my idea sucks. Don't listen to me. |
| 23:25 | <Dashiva> | "strict", "transitional", "invalid" |
| 23:26 | <othermaciej> | I think "good with no warnings", "good with warnings" and "bad, there are some errors" are sensible states |
| 23:26 | <othermaciej> | that is what programmers get from compilers and such |
| 23:29 | <Lachy> | that seems reasonable. Though it still sucks having to settle for compromises like this just to stop the arguments. |
| 23:29 | <othermaciej> | my personal point of view is that it's the right thing to do, not just a compromise |
| 23:30 | <othermaciej> | because sometimes, there are things that are often but not universally bad, and the right treatment is a mandatory but nonfatal diagnostic |
| 23:30 | <othermaciej> | some people will only care about error-free and take the warnings on a case-by-case basis, some will want their pages warning-free too |
| 23:31 | <Lachy> | I disagree, but fair enough |
| 23:31 | <Dashiva> | That's fair enough, I think the problem are the people who just see "Okay" and ignore the "but" |
| 23:31 | <Philip`> | The validator could let you specify -Werror and Wno-foo |
| 23:31 | <othermaciej> | yes, the validator should be able to toggle warning classes and also optionally treat warnings as fatal |
| 23:31 | <Philip`> | (for all values of foo) |
| 23:31 | <othermaciej> | I personally would want to use it in that mode, I think |
| 23:32 | <Lachy> | othermaciej, are you suggesting this be applied to all attributes that are currently considered downplayed errors, or just to summary? |
| 23:34 | <othermaciej> | Lachy: Hixie proposed converting the whole "downplayed error" class into something warning-like (i.e. does not break conformance) |
| 23:35 | <othermaciej> | I think having "error" and "warning" instead of "error" and "downplayed error" as the two classes of diagnostics makes sense |
| 23:35 | <Hixie> | yeah, i could reluctantly be convinced of that, i think |
| 23:35 | <othermaciej> | I haven't reviewed every current use of downplayed error, so I don't have an opinion on which should become nonfatal and which should be fatal |
| 23:36 | <othermaciej> | it would probably merit case-by-case review, since for some it may make sense to just convert them back to errors |
| 23:36 | <Hixie> | (i was reading a scifi story recently where people could run simulations of their mental processes on a computer, and at one point one of the characters is in a debate |
| 23:36 | <Hixie> | and he just ran 15,000 simulations of how the argument might play out with his simulated mental processes in "vm"s |
| 23:37 | <Hixie> | in 95% of cases he was convinced. so he just decided to assume he'd be convinced, and skip the 15 minutes discussion that would otherwise ensue. |
| 23:37 | <Hixie> | i find this a fascinating concept.) |
| 23:38 | <othermaciej> | I suppose that is meta-rational |
| 23:38 | <othermaciej> | (though sometimes it's important to know why you have been convinced) |
| 23:39 | Lachy | begins simulating tomorrow's public-html discussions |
| 23:39 | <Philip`> | Isn't that kind of missing the part of the debate that involves exchanging information with the other participant and becoming better informed? |
| 23:39 | <Hixie> | the simulation included a simulation of the other participant |
| 23:40 | <othermaciej> | hmm, did annevk ever put the client rect and caret from point stuff in the CSSOM Editor's Draft? |
| 23:40 | <othermaciej> | or is it somewhere else? |
| 23:41 | <Philip`> | A simulation based on the first person's mental representation of the second person? (in which case the first person still wouldn't be able to learn anything new) |
| 23:41 | <Hixie> | i don't recall the specifics |
| 23:43 | <Philip`> | "authors should be aware that additional media resources might be necessary" - but they are allowed to be unaware if they have valid reasons to be, though presumably they will have to somehow wipe their minds after carefully weighing the full implications and deciding to become unaware of it |
| 23:43 | <Hixie> | is that in html5? |
| 23:44 | Philip` | notes that he is aware he is making a trivial point and ignoring all the important issues presented in the email |
| 23:44 | <Philip`> | Hixie: No, in sayrer's recent post to public-html |
| 23:45 | <Hixie> | ah ok |
| 23:50 | Philip` | supposes a more constructive way to phrase the feedback would be to say that the proposed text uses the word "should" with no intended normative meaning, which is inconsistent with how that word is used throughout the rest of HTML 5 |
| 23:50 | <Philip`> | but that's a boring way to say it |
| 23:59 | <Lachy> | "New standards shouldn't carry the burden of backwards compatibility, that's the browser's (implementor's) job!" -- http://twitter.com/minusfive/statuses/2522138950 |