| 00:21 | <Rik`> | dandaman: you can use media queries in CSS |
| 00:21 | <dandaman> | see |
| 00:21 | <dandaman> | i did that |
| 00:21 | <dandaman> | <style> |
| 00:21 | <dandaman> | @media handheld { |
| 00:21 | <dandaman> | div.larger{font-size:72px;} |
| 00:21 | <dandaman> | } |
| 00:21 | <dandaman> | </style> |
| 00:21 | <dandaman> | but nothing has changed on my android emulator |
| 00:22 | <Rik`> | dandaman: handheld is not really supported on mobile devices |
| 00:22 | <dandaman> | so which should i be using? |
| 00:22 | <Rik`> | @media screen and (max-device-width: 400px) should work better |
| 00:22 | <Rik`> | (don't know if 400px is a good value) |
| 00:23 | <dandaman> | do you know if there is a way to make it so it automatically fits the resolution of the device it's on? |
| 00:24 | <Rik`> | the meta viewport should work |
| 00:24 | <dandaman> | i tried that |
| 00:24 | <dandaman> | didnt change anything on the emulator |
| 00:25 | <dandaman> | i did width=device-width |
| 00:29 | <dandaman> | actually I'm mistaken with my question |
| 00:29 | <dandaman> | everytime i try to see the webpage on the emulator |
| 00:29 | <dandaman> | the font is really really small |
| 00:29 | <dandaman> | and i need to zoom in a lot |
| 00:30 | <dandaman> | i want to make it so i start zoomed in on the content |
| 00:31 | <Rik`> | the meta viewport is here for that normally |
| 00:32 | <dandaman> | <meta name="viewport" content="width=device-width, height=device-height,minimum-scale=20, user-scalable=yes" /> |
| 00:32 | <dandaman> | is that right? |
| 00:32 | <Rik`> | just use <meta name="viewport" content="width=device-width"> |
| 00:33 | <Rik`> | there's no "height" parameter, user-scalable defaults to yes |
| 00:33 | <Rik`> | try that first |
| 00:34 | <dandaman> | I think the android emulator might just be broken |
| 00:41 | <dandaman> | Rik`: do you know if there is a way to extract the resolution from a device? |
| 00:41 | <dandaman> | maybe through javascript or something? |
| 00:42 | <Rik`> | screen.width and screen.height I think |
| 00:42 | <Rik`> | see http://quirksmode.org/mobile/tableViewport.html and http://quirksmode.org/mobile/ in general |
| 00:42 | <dandaman> | oh cool, thanks |
| 00:43 | <dandaman> | sweey |
| 00:43 | <dandaman> | sweet |
| 01:15 | <Hixie> | wow, only the rendering rules to go |
| 02:41 | <arosenberg> | HTML5 finally guarantees that an injected script tag will receive an 'error' Event if the src can't be fetched. |
| 02:41 | <arosenberg> | However, XMLHTTPRequest has provisions to learn the status code, etc. from the fetch. |
| 02:42 | <arosenberg> | It seems like a big gap to not have status code and headers from the response made available somewhere. |
| 02:45 | <arosenberg> | Case in point: Twitter can return rate limiting information as headers along with a 400 response to a JSONP query, but there's no standard way to retrieve that info. |
| 02:46 | <arosenberg> | One might call that a failure of their API design, but it still seems like functionality that should be in the standard. |
| 04:25 | <MikeSmith> | boblet: naoko and I are on a nozomi on the way to Kyoto |
| 04:26 | <boblet> | MikeSmith: nice — what are you up to down ther? |
| 04:26 | <boblet> | there? |
| 04:27 | <MikeSmith> | we'll be around this afternoon sightseeing about it, have some plans tomorrow, then free I think on Sunday in the daytime, and leaving around 7pm to head back to Tokyo |
| 04:27 | <MikeSmith> | I don't know what I meant by "about it".. |
| 04:27 | <MikeSmith> | oh |
| 04:27 | <MikeSmith> | I meant "sightseeing a bit" |
| 04:30 | <MikeSmith> | boblet: btw, I made some minor formatting tweaks to the generated "HTML5 (Edition for Web Authors)" subset of the spec - http://dev.w3.org/html5/spec-author-view/ |
| 04:30 | <MikeSmith> | please take a look at it when you have time and let me know if there's anything else that might be worthwhile to tweak in it to make it more useful |
| 04:30 | <MikeSmith> | I can't add substantive content |
| 04:31 | <MikeSmith> | because it needs to remain a strict subset of the spec |
| 04:31 | <boblet> | MikeSmith: will do |
| 04:31 | <boblet> | MikeSmith: when are you arriving? |
| 04:31 | <MikeSmith> | 2:30 |
| 04:32 | <boblet> | hrm not much time for sightseeing before dark. I guess you’ll have eaten? |
| 04:32 | <MikeSmith> | eating bento on the train |
| 04:32 | <MikeSmith> | but probably hungry again and eat once we get there |
| 04:32 | <MikeSmith> | we are meeting up with Naoko's mom there |
| 04:32 | <MikeSmith> | and going to Nanzenji |
| 04:33 | <MikeSmith> | (about the doc I can add, e.g., some CSS generated text where it might be useful) |
| 04:33 | <boblet> | i have to pick my daughter up tonight so I prolly couldn’t leave the station if I came today. want to eat dessert like last time, or would Sunday be better? |
| 04:35 | <MikeSmith> | either way that works for you, if you have time |
| 04:35 | <MikeSmith> | otherwise, you're visiting Tokyo again soon, right? |
| 04:35 | <MikeSmith> | we could definitely meet for tea or whatever at the station today if you can make time |
| 04:35 | <boblet> | I’ll be up for next WDE-Ex, 24-26th |
| 04:36 | <boblet> | yeah let’s do that |
| 04:36 | <MikeSmith> | OK |
| 04:36 | <boblet> | I’ll pop up and aim to be there by 2:30 |
| 04:36 | <MikeSmith> | OK |
| 04:37 | <MikeSmith> | will give you a call when we arrive |
| 04:38 | <boblet> | rgr |
| 04:38 | <MikeSmith> | boblet: btw, I did add a paragraph to the Abstract of the doc - the part with the green outline at http://dev.w3.org/html5/spec-author-view/ |
| 04:38 | <MikeSmith> | if you have suggestions for the wording of that, lemme know |
| 04:39 | <boblet> | ok will check it out |
| 04:41 | <MikeSmith> | Hixie: when you have some time, please peruse http://dev.w3.org/html5/spec-author-view/ a bit and let me know if you have any suggestions for refining formatting or whatever to make it clear that it's the author view and not the full spec |
| 04:42 | <MikeSmith> | one thing I did was to append the text "(details for Web authors)" to all the h3 and below headings |
| 04:43 | <boblet> | MikeSmith: would you also be interested in asking Ben Schwarz for feedback? ref: http://www.germanforblack.com/articles/moving-towards-readable-w3c-specs |
| 04:43 | <MikeSmith> | yeah sure |
| 04:43 | <MikeSmith> | if you know him, please feel free to give him a heads-up about it yourself |
| 04:43 | <boblet> | ok will ping him (although his viewpoint might be more to start from scratch :) ) |
| 04:43 | <boblet> | (he’s coming over for RubyConf in Aug btw) |
| 04:44 | <MikeSmith> | maybe the author view would be a good place to experiment with some of his formatting ideas |
| 04:44 | <MikeSmith> | ah great |
| 04:44 | <MikeSmith> | will be good to meet him |
| 04:44 | <boblet> | agreed |
| 04:46 | <MikeSmith> | boblet: thanks for thinking about that -- it seems like it could work out really well if he has the time |
| 04:47 | <MikeSmith> | what I'd be inclined to do is just give him commit access to the source in that subdir and just leave it up to him to experiment with as he'd like, and I'd just step out of the way |
| 04:47 | <MikeSmith> | as far as the CSS/styling goes, I mean |
| 04:48 | <MikeSmith> | it occurs to me that it might be a very nice thing indeed for the author view to have a very different look and feel from the full spec |
| 04:50 | <MikeSmith> | and a very CSS Zen Gardeny way of showing how formatting can be used to tune views of the same spec source for different audience needs |
| 05:05 | <MikeSmith> | boblet: the more I think about the idea of having Ben come and do work some design magic, the more I like it |
| 05:05 | <MikeSmith> | feel free to let him know I'm really enthusiastic about it |
| 05:06 | <MikeSmith> | or I guess, maybe just Cc me on the message you send him |
| 05:06 | <MikeSmith> | (if you've not already sent it) |
| 05:09 | <boblet> | writing it now |
| 05:10 | <boblet> | think he’s really busy (presenting on HTML5 at WDS) but he’s also really keen. maybe I can tag-team with him |
| 05:12 | <MikeSmith> | yeah, definitely |
| 05:13 | <MikeSmith> | you guys would have free rein |
| 05:13 | <MikeSmith> | unless/until you do something that I don't like |
| 05:13 | <MikeSmith> | then I pull the plug |
| 05:13 | <MikeSmith> | and the fun will be over |
| 05:14 | <boblet> | heh |
| 05:14 | <MikeSmith> | anyway, seriously, I think this could be a big win on getting HTML5 info out to authors in a more friendly and usable form |
| 05:28 | <cardona507> | i'm late to the party |
| 05:30 | <MikeSmith> | cardona507: hey |
| 05:30 | <MikeSmith> | what's the 507 for? |
| 05:31 | <MikeSmith> | cardona507: if this is your first time here, I have to give you the obligatory warning that if you spend time on this channel, you are guilty by association |
| 05:31 | <MikeSmith> | and anything you say here can and will be held against you |
| 05:31 | <cardona507> | my name is carlos and my fam/friends call me los - 507 is los upside down - it's a tag that remains from when I was younger |
| 05:32 | <cardona507> | MikeSmith: you got me feelin nervous |
| 05:32 | <MikeSmith> | :) |
| 05:32 | <MikeSmith> | Carlos, we met before, right? |
| 05:32 | <MikeSmith> | at TPAC? |
| 05:32 | <cardona507> | i heard what happens when you hang out with that WHATWG crowd |
| 05:32 | <cardona507> | yep we sure did |
| 05:32 | <cardona507> | how have you been? |
| 05:33 | <MikeSmith> | cardona507: yeah, the whatwg disease is contagious |
| 05:33 | <MikeSmith> | cardona507: I am doing great |
| 05:33 | <cardona507> | busy? |
| 05:33 | <MikeSmith> | getting some good stuff done |
| 05:33 | <MikeSmith> | "busy" is relative, I guess |
| 05:33 | <MikeSmith> | I have recently been spending a lot of time on telephone conferences |
| 05:34 | <cardona507> | working groups? |
| 05:34 | <MikeSmith> | groups and other W3C calls |
| 05:35 | <MikeSmith> | it is really useful to be able to talk with other people on the phone in real time |
| 05:35 | <cardona507> | i agree - i'm not sure if I like working with a group or virtual better |
| 05:35 | <MikeSmith> | but it can also be pretty wearying sometimes |
| 05:35 | <cardona507> | I really like working from home - but with my son running around it can be hectic |
| 05:36 | <MikeSmith> | yeah |
| 05:36 | <MikeSmith> | but that gives you a good excuse to take a break now and then |
| 05:36 | <MikeSmith> | I sit at my PC for hours sometimes without getting up |
| 05:36 | <MikeSmith> | reading through e-mail, or on calls |
| 05:36 | <cardona507> | thats for sure |
| 05:37 | <MikeSmith> | so it takes up a lot of time, but seems odd a bit to call it "busy" |
| 05:37 | <MikeSmith> | I don't know how much e-mail you get, but I am finding it more and more difficult to keep up with the volume of mail I need to read every day |
| 05:37 | <MikeSmith> | I spend several hours a day just reading e-mail |
| 05:38 | <MikeSmith> | some days, I'd guess 4 hours or so |
| 05:38 | <MikeSmith> | of just reading |
| 05:39 | <MikeSmith> | for me at least, I really find IRC and other real-time chat/IM mechanisms a better way of dealing with a lot of daily communication tasks |
| 05:39 | <MikeSmith> | anyway, I digress |
| 05:39 | <MikeSmith> | as usual |
| 05:39 | <MikeSmith> | cardona507: it's great to have you here |
| 05:40 | <MikeSmith> | I hope you end up hanging around more |
| 05:40 | <MikeSmith> | because there's a lot of really useful discussion that takes place here |
| 05:40 | <MikeSmith> | amidst my nonsense |
| 05:40 | <MikeSmith> | and a lot you can learn here |
| 05:43 | <estellevw> | Question about microdata versus microformats..... |
| 05:43 | <estellevw> | microdata is the idea of using the itemprop attribute, not class attribute like microformats, so microformats really aren't part of html5, or are they somehow? |
| 05:43 | <estellevw> | Do i have that right? |
| 05:44 | <cardona507> | MikeSmith: thanks - I really appreciate it |
| 05:44 | <cardona507> | and truthfully I learned a whole lot from a localStorage tutorial of yours that I found on the web - just for the record :) |
| 05:44 | <MikeSmith> | cardona507: oh man, that thing is so old.. I should take it down |
| 05:44 | <MikeSmith> | if you mean that little demo I made |
| 05:45 | <cardona507> | yep |
| 05:45 | <cardona507> | i've found my way back to it more than once |
| 05:45 | <cardona507> | it's the #1 search for 'localstorage demo' on google |
| 05:45 | <MikeSmith> | estellevw: Microdata is integrated in the whatwg version of the spec, but is published as a separate document from the W3C version of HTML5 |
| 05:46 | <MikeSmith> | cardona507: wow |
| 05:46 | <MikeSmith> | I'm #1! |
| 05:46 | <MikeSmith> | that's scary, really |
| 05:47 | <estellevw> | http://people.w3.org/mike/localstorage.html |
| 05:47 | <estellevw> | very nice! |
| 05:47 | <MikeSmith> | there are much better local storage demos on the Web than that one |
| 05:47 | <estellevw> | yours is nice and simple though, so easy to learn from |
| 05:47 | <MikeSmith> | and anyway I just stole that one from a demo that Hixie made! |
| 05:48 | <cardona507> | haha |
| 05:48 | <cardona507> | then I don't feel so bad that I reused that code :) |
| 05:48 | <MikeSmith> | estellevw: oh, I see you were asking about microformats |
| 05:48 | <MikeSmith> | not microdata |
| 05:48 | <MikeSmith> | I guess the microformats mechanism could be considered "part of" HTML |
| 05:49 | <MikeSmith> | but not the specific microformats vocabularies |
| 05:49 | <MikeSmith> | the actual semantics of the microformats class values are not defined in HTML5 |
| 05:50 | <estellevw> | mikesmith: yes, wondering about the relation. I always thought they were the same, but now that i am looking at the spec for the first time for microdata, it seems that microdata liked the microformats idea, but may never take off on its own, though i guess we can use both microdata classes and itemprop style attributes at the same time |
| 05:50 | <estellevw> | ok, thanks for clarifying. |
| 05:52 | <MikeSmith> | estellevw: I think the way that Web developers/authors should consider the various linked-data/structured-data markup mechanisms is, use whichever one like and that solves whatever problem or need it is that you're aiming for |
| 05:53 | <MikeSmith> | I think one specific need that people have now is to add data so that it will get indexed in the right way for Google's Rich Snippets thing |
| 05:53 | <estellevw> | Right now microformats are fairly well supported by humans and machines alike. |
| 05:53 | <MikeSmith> | yeah |
| 05:53 | <MikeSmith> | exactly |
| 05:53 | <MikeSmith> | and by things like Google Rich Snippets |
| 05:54 | <estellevw> | so i don't see microdata taking off and replacing it |
| 05:54 | <estellevw> | they likely can co-exist nicely in code |
| 05:54 | <MikeSmith> | microformats was never intended to replace anything |
| 05:54 | <estellevw> | but wondering if anyone major will start scraping for itemprop/itemscope in the next 2 years |
| 05:54 | <estellevw> | no, i understand that |
| 05:55 | <kennyluck> | estellevw, their are microdata parsers alreday. |
| 05:55 | <estellevw> | reading the spec i feel like the spec is offering a standardiszed replacement |
| 05:55 | <kennyluck> | s/their/there/ |
| 05:55 | <MikeSmith> | OK -- yeah, it was intended to provide another option for those people who want one -- a mechanism that has a very clearly and precisely defined processing model, and an API |
| 05:55 | <estellevw> | who is parsing microdata |
| 05:56 | <estellevw> | yeah, i like it. i think having an attribute just for that is helpful. multipurposing class was a good idea too |
| 05:57 | <kennyluck> | Well, the Goolge Rich Snippets mechanism reads Microdata as well as microformat and RDFa. |
| 05:58 | <estellevw> | thanks kennyluck |
| 05:58 | <estellevw> | an you too, mikesmith |
| 05:58 | <MikeSmith> | cheers |
| 05:59 | <MikeSmith> | there's a statement I came across in an odd place yesterday that I think is nonetheless really well put: "Developers should favor the lightest-weight technology that will meet their requirements." |
| 06:01 | <MikeSmith> | so I guess that if, e.g., somebody finds that using microformats markup is all they need in order to get Google Rich Snippets to index their docs the way they want, then it makes good sense for them to use that |
| 06:01 | <estellevw> | is anyone else having an issue with microformats.org ... it reloads to a blank page on page load for me. |
| 06:02 | <MikeSmith> | estellevw: same here |
| 06:02 | <MikeSmith> | it seems to be borked |
| 06:02 | <estellevw> | wondering if there is a need to duplicate efforts: <span class="fn" itemprop="fn" |
| 06:02 | <estellevw> | some script |
| 06:02 | <estellevw> | should I let @t know, or is he no longer involved with it? |
| 06:03 | <MikeSmith> | dunno if he is responsible for the site or not |
| 06:03 | <MikeSmith> | maybe ask on #microformats channel here? |
| 06:06 | <estellevw> | ah, thanks. good idea |
| 06:09 | <estellevw> | no one has said anything there in days, but i'll try pinging the actives there, like jonathan neal, boblet and adactio. maybe one of them knows |
| 06:10 | <MikeSmith> | kennyluck: the concept of a "property" is something that is part of RDF, right? |
| 06:11 | <MikeSmith> | basically meaning the same thing as "predicate" |
| 06:11 | <MikeSmith> | or not? |
| 06:11 | <kennyluck> | MikeSmith, yeah it means predicate, and is defined in RDF. |
| 06:11 | <MikeSmith> | OK |
| 06:12 | <kennyluck> | But a property usually connects two things, so it is also similar to "property" in programming language's sense. |
| 06:12 | <kennyluck> | s/usually/always/ |
| 06:15 | <MikeSmith> | kennyluck: so let me ask another question |
| 06:15 | <MikeSmith> | a very high level one |
| 06:16 | <MikeSmith> | HTML has a number of elements which have attributes whose values are URIs |
| 06:16 | <MikeSmith> | It would seem to me that those elements are particularly useful for marking up linked data |
| 06:16 | <MikeSmith> | regardless of what markup mechanism you use |
| 06:17 | <kennyluck> | Yes. |
| 06:17 | <kennyluck> | RDFa uses this mechanism |
| 06:17 | <kennyluck> | So href and src could be the target of a property. |
| 06:17 | <MikeSmith> | but I am not sure how to describe in general terms why they are useful for linked-data purposes |
| 06:17 | <kennyluck> | But it's a bit hacky, I have to say. |
| 06:18 | <MikeSmith> | kennyluck: understood |
| 06:18 | <kennyluck> | Uhh, in some sense, HTML is linked data. |
| 06:18 | <kennyluck> | So what you are saying makes a lot of sense. |
| 06:18 | <kennyluck> | s/HTML/HTML-HTML linking/ |
| 06:19 | <MikeSmith> | OK, so what I am asking is, how do I describe how, at a conceptual level, an element that has a URI attribute is especially useful for linked-data purposes? |
| 06:19 | <kennyluck> | or HTML-IMG linking |
| 06:21 | <kennyluck> | Because, in the linked-data world, we actually prefer cross domain/cross machine links. And HTML has this ability <img> <a> <script>. Becuase they use URLs. |
| 06:21 | <kennyluck> | (I am not sure I am answering the question...) |
| 06:21 | <kennyluck> | MikeSmith |
| 06:22 | <kennyluck> | The Web is interesting because it's decentralized. I guess. |
| 06:22 | <kennyluck> | The links are cool data. Not just the contents. |
| 06:22 | <kennyluck> | So I really like blog posts. :) |
| 06:47 | mhausenblas | now realises where kennyluck has been in the past 30min :P |
| 06:49 | <Hixie> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=10061 http://www.w3.org/Bugs/Public/show_bug.cgi?id=10113 |
| 06:49 | <Hixie> | i can't win |
| 06:59 | <hsivonen> | Hixie: awesome |
| 07:18 | <estellevw> | question about itemtype attribute |
| 07:18 | <estellevw> | if the value is an URL |
| 07:19 | <estellevw> | what is supposed to be at that URL? |
| 07:19 | <estellevw> | a DTD for the microdata type? |
| 07:19 | <hsivonen> | estellevw: nothing |
| 07:19 | <kennyluck> | I am interested in this question as well. |
| 07:20 | <annevk> | http://www.whatwg.org/specs/web-apps/current-work/complete/microdata.html#item-type |
| 07:20 | <kennyluck> | If nothing, then I guess using reverse DNS would be more appropriate. |
| 07:20 | <annevk> | "Item types are opaque identifiers, and user agents must not dereference unknown item types, or otherwise deconstruct them, in order to determine how to process items that use them." |
| 07:20 | <annevk> | kennyluck, we had that, but people did not like that |
| 07:21 | <estellevw> | so, each itemscope with the same microdata structure should point to the same URI, but nothing needs to be at that URI - a 404 is ok? |
| 07:21 | <hsivonen> | estellevw: 404 is OK |
| 07:21 | <kennyluck> | This is kind of a stupid result, I suppose. |
| 07:21 | <annevk> | yeah, it's just an identifier that looks like a URL |
| 07:21 | <annevk> | kennyluck, RDF has pretty much the same |
| 07:21 | <kennyluck> | Linked Data asks every URL to be dereferenceable. |
| 07:21 | <annevk> | kennyluck, in fact, we got this because of RDF afaik |
| 07:21 | <kennyluck> | No, this is some old thinking. |
| 07:21 | <estellevw> | or it could be any random string, other than the official microdata uri |
| 07:22 | <kennyluck> | You shouldn't listen to those who don't understand HTTP. |
| 07:22 | <annevk> | estellevw, well, it has to be a URL |
| 07:22 | <estellevw> | "Item types are opaque identifiers, and user agents must not dereference unknown item types, or otherwise deconstruct them, in order to determine how to process items that use them." is confusing to me, and i am a native english speaker |
| 07:22 | <annevk> | kennyluck, see e.g. XML Namespaces |
| 07:23 | <kennyluck> | Yes, I know I know. But FYI, W3C has the policy that XML Namespaces should ALL be dereferenceable. |
| 07:23 | <kennyluck> | I mean XML namesapces starting with http://www.w3.org/ |
| 07:23 | <annevk> | W3C also does that wrongly in several cases |
| 07:24 | <hsivonen> | estellevw: I think it's more confusing the the RDF folks thought it was a great idea to use Web addresses for identifying things that aren't the documents pointed to by the URLs |
| 07:24 | <kennyluck> | That's a mistake, then. |
| 07:24 | <annevk> | ever tried to dereference http://www.w3.org/1999/xhtml or http://www.w3.org/1999/xlink ? |
| 07:24 | <annevk> | oh, maybe it was not the XLink namespace |
| 07:24 | <annevk> | or maybe it got fixed |
| 07:24 | <kennyluck> | Maybe it's a new policy.. let me find a reference. |
| 07:25 | <hsivonen> | annevk: it's super-annoying that the XHTML namespace redirects, so you can't do ctrl-L, ctrl-C |
| 07:25 | <kennyluck> | http://www.w3.org/1999/10/nsuri |
| 07:26 | <kennyluck> | "Namespace URIs MUST be dereferenceable" |
| 07:26 | <hsivonen> | (well, you can, but you get the wrong thing on the clipboard) |
| 07:26 | <annevk> | hsivonen, yeah |
| 07:27 | <annevk> | kennyluck, sure, but that's just for w3.org namespaces |
| 07:27 | <kennyluck> | agreed |
| 07:27 | <annevk> | kennyluck, it's not something you can enforce anyway and really in retrospect using URLs as identifiers is a bad idea |
| 07:28 | <nessy> | cause there's no resource behind it? |
| 07:28 | <kennyluck> | If we couldn't enforce this practice, I have to say using URLs as identifiers is indeed a bad idea. |
| 07:29 | <annevk> | nessy, cause clients start fetching them while they shouldn't |
| 07:30 | <mhausenblas> | whatz the harm in this, annevk? |
| 07:30 | <mhausenblas> | that is, why shouldn't clients fetch them? |
| 07:31 | <kennyluck> | This is indeed a difficult question. Ted of the W3C systems team indeed treats those dereferenceable namespaces as a big problem. |
| 07:32 | <kennyluck> | But I think the sentence "item types are opaque identifiers, and user agents must not dereference unknown item types" is not very nice. |
| 07:33 | <mhausenblas> | kennyluck, it's confusing |
| 07:33 | <mhausenblas> | (but maybe this is the intention ;) |
| 07:33 | <hsivonen> | mhausenblas: http://www.w3.org/blog/systeam/2008/02/08/w3c_s_excessive_dtd_traffic |
| 07:33 | <mhausenblas> | hsivonen, yeah I know this |
| 07:34 | <annevk> | mhausenblas, you get distributed denial of service attacks and there's no real benefit |
| 07:34 | <mhausenblas> | so, your argument is: cause some people fcked it up in the past, we won't allow it, right? |
| 07:34 | <mhausenblas> | let me ask you one thing, please |
| 07:35 | <hsivonen> | mhausenblas: don't give people rope to shoot others in the foot with |
| 07:35 | <mhausenblas> | with your mind set - would the WWW in 1994-1996 have taken off? |
| 07:35 | <mhausenblas> | hsivonen please don't pretend to be sooo protective ;) |
| 07:35 | <mhausenblas> | I doubt it |
| 07:36 | <mhausenblas> | yes, there will be people fcking it up again |
| 07:36 | <annevk> | me too, I was like eight back then |
| 07:36 | <mhausenblas> | so whatz next? disallowing knifes |
| 07:36 | <mhausenblas> | cause someone could be killed? |
| 07:36 | <hsivonen> | mhausenblas: sure. there were no namespaces back then |
| 07:36 | <annevk> | but I don't see how that matters |
| 07:36 | <mhausenblas> | annevk, me too? |
| 07:36 | <annevk> | I doubt it too |
| 07:36 | <mhausenblas> | doubt what? :) |
| 07:36 | <annevk> | there's clear harm and benefit has not been demonstrated |
| 07:37 | <mhausenblas> | annevk, please back up your arguments |
| 07:37 | <mhausenblas> | where is the clear harm |
| 07:37 | <annevk> | mhausenblas, see the link above |
| 07:37 | <mhausenblas> | come on |
| 07:37 | <hsivonen> | mhausenblas: non-stab knives are actually a pretty good idea, but they aren't gonna take off until 20 have elapsed |
| 07:37 | <hsivonen> | (hooray for patents) |
| 07:37 | <mhausenblas> | that is not a clear proof this is a random link you pull out of the thin air |
| 07:37 | <mhausenblas> | hsivonen, are you trying to be funny or having a real convo? |
| 07:38 | <annevk> | mhausenblas, same thing happened with the Netscape DTD for RSS |
| 07:38 | <hsivonen> | mhausenblas: I'm serious |
| 07:38 | <mhausenblas> | hsivonen, aha |
| 07:38 | <mhausenblas> | annevk, what happened? |
| 07:38 | <annevk> | mhausenblas, it's clear anecdotical evidence that it doesn't work |
| 07:38 | <annevk> | mhausenblas, and again, there's no benefit |
| 07:38 | <mhausenblas> | annevk, cause you say so? |
| 07:38 | <mhausenblas> | annevk, I see a lot of benefit |
| 07:38 | <annevk> | such as? |
| 07:38 | <mhausenblas> | some 20billion RDF triples provided by institutions such as BBC, NYT, etc. |
| 07:39 | <mhausenblas> | proof, on a daily basis that there is |
| 07:39 | <annevk> | that has nothing to do with the benefit of opaque identifiers being URLs |
| 07:39 | <kennyluck> | annevk, if we have a stable URI for jQuery. Caching it would reduce a lot of service load. |
| 07:39 | <mhausenblas> | but prolly these people there don't have the great insights you chaps have ;) |
| 07:39 | <annevk> | that just means RDF is useful to some instutitions |
| 07:39 | <mhausenblas> | annevk: |
| 07:39 | <annevk> | but it does not mean all properties of RDF are |
| 07:39 | <mhausenblas> | sorry ;) |
| 07:40 | <mhausenblas> | RDF itself is not much worth |
| 07:40 | mhausenblas | started with TopicMaps in 2000 |
| 07:40 | <mhausenblas> | then switched over to RDF |
| 07:40 | <mhausenblas> | however, the real potential was only revealed in 2006 |
| 07:40 | <mhausenblas> | when Tim wrote up the Linked Data note |
| 07:40 | <annevk> | look, all I'm interested in is the benefit of identifiers being URLs that can be fetched |
| 07:40 | <annevk> | you saw a lot of benefit |
| 07:40 | <mhausenblas> | yes |
| 07:40 | <annevk> | so what is it? |
| 07:41 | <mhausenblas> | ok, annevk, one step at a time, ok? |
| 07:41 | <mhausenblas> | so can we agree on that the Web is a decentralised, system? |
| 07:41 | <mhausenblas> | a *global*, decentralised system |
| 07:41 | <hsivonen> | kennyluck: single point of failure with a single URL for jquery |
| 07:41 | <mhausenblas> | yes, annevk? |
| 07:42 | <kennyluck> | hsivonen, true. Let's have two stable URLs for jQuery, shall we? |
| 07:42 | <mhausenblas> | hm - do you need to think that long about the answer or is the question to confusing, annevk? :) |
| 07:42 | <mhausenblas> | s/to/too |
| 07:43 | <annevk> | mhausenblas, I'm still waiting for your argument |
| 07:43 | <mhausenblas> | I asked you a question, annevk ... :) |
| 07:43 | <kennyluck> | hsivonen, I would <script src="<the URI of jQuery on Google>"> because I think it's the most efficient. |
| 07:43 | hsivonen | mumbles about the dissonance of single url and decentralization |
| 07:43 | <annevk> | mhausenblas, also, I'm not fulltime on this channel |
| 07:43 | <kennyluck> | I think Google recommend that as well. |
| 07:43 | <annevk> | mhausenblas, but yes, the Web is somewhat decentralized |
| 07:44 | <mhausenblas> | somewhat? |
| 07:44 | <mhausenblas> | hu? |
| 07:44 | <mhausenblas> | anyway |
| 07:44 | <annevk> | if you cut a few lines it all falls apart |
| 07:44 | <mhausenblas> | so, assuming that the Web is *somewhat* decentralised |
| 07:44 | <hsivonen> | kennyluck: it's a pure address. if the address no longer works, you can change it without breaking apps that want a pure identifier |
| 07:44 | <mhausenblas> | and, there is no single point of registration, right? |
| 07:44 | <hsivonen> | (this was already discussed at TPAC 2008) |
| 07:45 | <mhausenblas> | I can just put my documents (or my data) on the Web without registering it |
| 07:45 | <mhausenblas> | however, from an application perspective, one needs single points of access |
| 07:45 | <mhausenblas> | such as Google is for documents (mainly ;) |
| 07:45 | <kennyluck> | hsivonen, RDF has the mechanism to identify the same thing with different URIs. We want serveal URIs for a thing we often use, but not too many. |
| 07:45 | <hsivonen> | mhausenblas: you need to register for DNS |
| 07:46 | <mhausenblas> | hsivonen, correct |
| 07:46 | <mhausenblas> | but not for the *content* hsivonen |
| 07:46 | <mhausenblas> | DNS is a necessary thing for decentralised management |
| 07:46 | <mhausenblas> | back to annevk |
| 07:46 | <mhausenblas> | so, as a human user, what do you do, really? |
| 07:47 | <mhausenblas> | you're a manual dereferencing processor |
| 07:47 | <mhausenblas> | you click on links |
| 07:47 | <mhausenblas> | HATEOAS and the like, you remember annevk? |
| 07:47 | <mhausenblas> | now, why can you do this? |
| 07:48 | <mhausenblas> | cause the when you (or better say, your User Agent) tries to GET stuff |
| 07:48 | <mhausenblas> | you actually receive something (that can be rendered in a browser) |
| 07:48 | <mhausenblas> | so much for the human users |
| 07:49 | <mhausenblas> | IFF I want to enable the Web as a single global database |
| 07:49 | <mhausenblas> | the machines need to be able to do exactly the same |
| 07:49 | <mhausenblas> | IMO |
| 07:49 | <mhausenblas> | even MS seems to get it in OData |
| 07:49 | <mhausenblas> | see http://www.odata.org/blog/2010/4/22/queryable-odata-metadata |
| 07:50 | <mhausenblas> | (and also my initial post on comparing OData with Linked Data at http://webofdata.wordpress.com/2010/04/14/oh-it-is-data-on-the-web/) |
| 07:50 | <mhausenblas> | so, annevk, I guess you still don't see the benefit, right? |
| 07:51 | <annevk> | right |
| 07:51 | <mhausenblas> | aha |
| 07:51 | <mhausenblas> | thanks for the convo |
| 07:51 | <mhausenblas> | bye |
| 07:53 | <annevk> | no worries; I just don't see how that follows from what we have now and I have not seen concrete benefit in the case of RDF being deployed today |
| 07:57 | <kennyluck> | annevk, we are working hard. And we need more communication as well, not just playing locally with academic people. |
| 07:58 | <kennyluck> | Web is not built up by those who publish paper. I think Semantic Web will be the same. |
| 08:38 | <estellevw> | http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#microdata |
| 08:39 | <estellevw> | the example uses <address>, which is not the correct usage of the <address> element |
| 09:09 | <Smylers> | estellevw: Why not? That address could be the contact details for the page on which they appear. |
| 09:11 | <estellevw> | The address element represents the contact information for its nearest article or body element ancestor. If that is the body element, then the contact information applies to the document as a whole. |
| 09:11 | <estellevw> | The address element must not be used to represent arbitrary addresses (e.g. postal addresses), unless those addresses are in fact the relevant contact information. (The p element is the appropriate element for marking up postal addresses in general.) |
| 09:12 | <estellevw> | from that example, it seems like an arbitrary address, not the contact details for the page |
| 09:12 | <Smylers> | Clearly it must be the contact details for the page — cos it's using the <address> element! |
| 09:12 | <estellevw> | lol |
| 09:14 | <Smylers> | That's a really common scenario: many sites have their own contact details, whereas relatively few have other people's addresses. So I think the example is fine. |
| 09:14 | <Smylers> | But ... it may be worth clarifying, so people don't see it and think it's the thing to do for any address. |
| 09:15 | <estellevw> | exactly |
| 09:15 | <Smylers> | Raise a bug if you're bothered by it. |
| 09:15 | <estellevw> | i didn't know how on that page |
| 09:15 | <estellevw> | just a general whatwg bug? |
| 09:16 | <annevk> | at the bottom of the viewport there's an arrow |
| 09:16 | <Smylers> | Do you see a bar along the bottom of the viewport with a textbox? |
| 09:16 | <annevk> | if you hit the arrow you get a textbox |
| 09:16 | <annevk> | if you fill in the textbox and hit enter and hit confirm or something you filed a bug |
| 09:16 | <estellevw> | oh, doh, i had turned off js to see the micorformats page (which currently has a bad script in it). |
| 09:17 | <estellevw> | forgot to turn it back on |
| 09:17 | <annevk> | ah |
| 09:17 | <annevk> | that might be problematic indeed |
| 09:17 | <micheil> | annevk: howdy' |
| 09:18 | <annevk> | hey |
| 09:18 | <annevk> | thought I'd recognized your name :) |
| 09:21 | <micheil> | annevk: one of the few people in the world with the same name. |
| 09:21 | <micheil> | (that I know of.) |
| 09:21 | <micheil> | the only other person who I met on the net with the same name created a fake twitter account and wanted to kill me for "stealing their name" |
| 09:22 | <annevk> | oh god |
| 09:24 | <micheil> | heh heh, it's a rare name. |
| 09:25 | <annevk> | for my gender mine is also pretty rare ;p |
| 09:25 | <micheil> | heh heh |
| 10:14 | <annevk> | lol -- http://twitter.com/opera/status/18105017768 |
| 10:19 | <micheil> | anyone know if opera or firefox or ie9 do websockets? |
| 10:20 | <annevk> | I think Firefox 4 does |
| 10:21 | <annevk> | We still have some work to do and the first sign Microsoft gave was giving feedback on the requirements document so I guess not yet, but maybe... |
| 10:31 | <micheil> | annevk: okay, just checking |
| 10:31 | <micheil> | thinking I might start maintaining a matrix of browsers and websocket versions supported |
| 10:32 | <annevk> | you could modify http://wiki.whatwg.org/wiki/Implementations_in_Web_browsers#Web_Sockets |
| 10:32 | <annevk> | modify / fix |
| 10:32 | <micheil> | oh? |
| 10:33 | <micheil> | I think I'll just start fresh.. |
| 10:33 | <micheil> | although, I do need to know what version of websockets chrome 4 supported |
| 10:33 | <micheil> | (I can only get chrome 5 or 6 on mac) |
| 10:33 | <annevk> | if you make a page of your own please link to it from there at least |
| 10:35 | <micheil> | sure, will do |
| 10:35 | <micheil> | it'll probably be up at: http://wiki.github.com/miksago/node-websocket-server/browser-support or something |
| 10:36 | <micheil> | in fact, it will be. |
| 10:36 | <micheil> | I'll write up what I know already later on tonight, have to do work now. |
| 10:37 | <Philip`> | I thought Chrome updated automatically so nobody should be using/caring about Chrome 4? |
| 10:39 | <micheil> | Philip`: on the off chance. |
| 10:39 | <Rik`> | micheil: Firefox 4 has WebSockets support (-76) |
| 10:39 | <micheil> | I thought firefox automatically updated, but people still use firefox 1.5 |
| 10:39 | <Philip`> | Maybe they only added automatic updating in version 2.0? :-) |
| 10:39 | <Rik`> | micheil: I used your nodejs websocket server to code some demos for Firefox 4 (thanks btw) |
| 10:40 | <micheil> | Rik`: oh, cool |
| 10:40 | <micheil> | Rik`: any feedback? |
| 10:41 | <Rik`> | micheil: works well generally |
| 10:41 | <micheil> | Rik`: was this which version? v1.3.00 or? (released yesterday) |
| 10:41 | <Rik`> | I had an error message with Firefox about a big endian stuff which was not implemented yet |
| 10:41 | <Rik`> | old version, I should update when I publish a demo |
| 10:42 | <micheil> | Rik`: hmm.. that would be the draft 76 protocol. |
| 10:42 | <micheil> | draft 75 doesn't do big endian |
| 10:42 | <Rik`> | and I also changed the broadcast method to broadcast to everyone but the sender |
| 10:43 | <Rik`> | (I should have added a method though but quick hack during testing) |
| 10:43 | <micheil> | Rik`: I'd appreciate any feedback & / or publicity on you using node-websocket-server, purely there are many dead websocket-servers for node, which only do partial draft 75 and don't work well |
| 10:43 | <micheil> | Rik`: that's changed in v1.3.00 slightly |
| 10:43 | <Rik`> | I know, I've had a hard time finding a cool one :) |
| 10:43 | <Rik`> | well, we should publish something on hacks.mozilla.org soon |
| 10:44 | <micheil> | there's a new method on the connection objects called broadcast which will send to all connections but that connection |
| 10:44 | <micheil> | and the server broadcast remains the same |
| 10:44 | <micheil> | there's also an improvement to http server doubling |
| 10:45 | <micheil> | Rik`: check out: http://groups.google.com/group/nodejs/browse_thread/thread/a9a49056b65ea5ae/c63c9584e973c099 |
| 10:46 | <Rik`> | micheil: I've demoed your server in front of Ryan Dahl last saturday ;) |
| 10:46 | <micheil> | Rik`: yeah, I worked closely with ryan when developing it |
| 10:46 | <micheil> | there was a hint at one stage that ryan may add websocket support to node core. |
| 10:46 | <micheil> | anyway, bbl, dinner. |
| 10:47 | <Rik`> | - Modified echo-server example to also serve up the client.html by default. |
| 10:47 | <Rik`> | micheil: that's what I did myself :) |
| 11:42 | <karlcow> | http://jaredforsyth.com/blog/2010/jul/8/announcing-codetalker/ |
| 11:43 | <karlcow> | using Python to define a grammar instead of say BNF |
| 11:52 | <karlcow> | annevk: Opera starts a H1N1 campaign? |
| 11:54 | <annevk> | a mutated version that gives everyone direct access to a browser |
| 11:54 | <karlcow> | ;) |
| 11:54 | <micheil> | Rik`: heh heh, I've been working on it when I get a chance |
| 11:55 | <micheil> | Rik`: there's been a request from one user to be able to store data with a connection, so I'm trying to implement a way to do that. |
| 11:55 | <Rik`> | micheil: store as in "store while the server is running" |
| 11:55 | <micheil> | karlcow: That'd be funnier if it was the Mozilla Messaging team, you know, bird flu. |
| 11:55 | <micheil> | Rik`: yea |
| 11:56 | <micheil> | Rik`: feel free to join me in #node.js if you wish to talk more in-depth about it |
| 11:57 | <Rik`> | thanks |
| 11:57 | <Rik`> | I'm not working on it at the moment though |
| 11:58 | <micheil> | well, I mean, to continue discussion about it, not necessarily if you needed any help ;) |
| 14:18 | <hallvors> | what does a "host-specific component" mean in the text "an absolute URL with a <host-specific> component that is either empty or a single U+002F SOLIDUS character (/)" ? |
| 14:18 | <hallvors> | I tried following the links to RFC9387 but I couldn't find host-specific defined there |
| 14:23 | <Philip`> | hallvors: http://www.w3.org/html/wg/href/draft |
| 14:24 | <Philip`> | hallvors: I think it used to be defined in HTML5, then got split into a separate document, then got abandoned |
| 15:01 | <hallvors> | Philip`: thanks. Actually that document isn't complete/clear either. But I think "an absolute URL with a <host-specific> component that is either empty or a single U+002F SOLIDUS character (/)" means a URL with protocol, hostname but no path. |
| 15:02 | <zcorpan_> | hallvors: the path might be / |
| 15:02 | <Philip`> | hallvors: Why is "The decoded substring that follows the decoded substring matched by the <authority> production, or the whole string if the <authority> production wasn't matched." not clear? |
| 15:02 | <hallvors> | (This is from text regarding window.postMessage's targetOrigin argument. I think the spec says that a complete URL with path more than just a / is not a valid targetOrigin) |
| 15:03 | <hallvors> | Philip`: because I don't find <authority> defined in that spec..? |
| 15:03 | <hallvors> | ..with a quick search.. |
| 15:03 | <Philip`> | hallvors: <authority> is from RFC3986 |
| 15:03 | <zcorpan_> | hallvors: you're reading an outdated draft |
| 15:04 | <zcorpan_> | oh maybe not |
| 15:04 | hallvors | hates drafts and the fact that whatever I'm reading is always outdated when he needs to understand something |
| 15:04 | hallvors | isn't sure if he is first or third person grammatically. oh well. |
| 15:05 | zcorpan_ | is reading http://www.whatwg.org/specs/web-apps/current-work/complete/web-messaging.html#web-messaging |
| 15:05 | <jre> | zcorpan, http://www.w3.org/html/wg/href/draft is outdated, and abandoned, and known to be buggy |
| 15:06 | <Philip`> | Is there a better reference to use instead? |
| 15:06 | <hallvors> | I was reading http://www.whatwg.org/specs/web-apps/current-work/multipage/comms.html#security-postmsg |
| 15:06 | <jre> | the problem is that there's currently nothing replacing it |
| 15:06 | <zcorpan_> | hallvors: yeah that should be up to date |
| 15:06 | <hallvors> | but - window.postMessage('foo', 'http://www.example.com/this/is/a/path';) is invalid, right? |
| 15:06 | <jre> | optimally, the HTML5 spec would simply use RFC3986/RFC3987 terminology, and define everything additionally it needs on top of it |
| 15:07 | <jre> | see http://dev.w3.org/html5/status/issue-status.html#ISSUE-056 for some related information |
| 15:08 | <boblet> | foolip: in microdata, if a <meta> itemprop is moved into <head> by a browser, it should still work if referred to via @itemid right? (does it work like that in microdatajs?) |
| 15:13 | <foolip> | bobchao, if you mean itemref, then yes, it doesn't matter that the parser moves it |
| 15:13 | <foolip> | and itemref works in microdatajs |
| 15:15 | <boblet> | hrm. ages ago when I asked about this someone suggested using CSS to hide a span with the itemprop over meta due to this moving. I can’t think of any situation where that’d be better than using @itemref |
| 15:17 | <foolip> | I don't think that's a great idea really |
| 15:19 | <boblet> | well ugly kludges are good when they work, but using itemref is much nicer |
| 15:19 | <foolip> | alternatively, don't use hidden metadata |
| 15:21 | <boblet> | yeah I’m strongly suggesting not to |
| 15:26 | <Workshiva> | boblet: It could be the advice predates itemref |
| 15:26 | <boblet> | Workshiva: this was back in Feb, so entirely possible |
| 15:26 | <Workshiva> | It used to be a different mechanism, and before that, no mechanism at all |
| 15:37 | <boblet> | no browsers support microdata natively yet, right? also, is the same still true for RDFa and microformats? |
| 15:38 | <boblet> | (should anyone know :) ) |
| 15:38 | <foolip> | right, no browser supports anything |
| 15:39 | <boblet> | foolip: ta |
| 15:40 | <foolip> | It's amusing to see Facebook's flavor of RDFa allowing the xmlns:og declaration to be left out completely |
| 15:43 | <foolip> | boblet, what javascript errors are you getting with the vcard download example you couldn't get working? |
| 15:44 | <boblet> | no visible errors, didn’t check console |
| 15:44 | <boblet> | will look |
| 15:44 | <boblet> | “getVcard is not defined” |
| 15:45 | <foolip> | ok, did you really include all the needed scripts? |
| 15:45 | <boblet> | as an uncacught ReferenceError |
| 15:45 | <boblet> | this is on http://foolip.org/cv/ |
| 15:45 | <foolip> | right, that's broken |
| 15:45 | <foolip> | and I forgot the password so I can't fix it right away :/ |
| 15:45 | <boblet> | (I think your repo naming has changed) |
| 15:46 | <boblet> | har! |
| 15:47 | <boblet> | foolip: in a locally saved version of your page, with local links to microdata.js and microdata.vcard.js I get no errors but no download either |
| 15:49 | <foolip> | will have a look |
| 15:49 | <boblet> | what was the logic behind not having a XML literals equivalent in microdata? no/not compelling enough use cases? (datatypes seems like a vocabulary issue but including tags seems useful) |
| 15:50 | <boblet> | foolip: thanks — much appreciated as always |
| 15:53 | <boblet> | another microdata q; do we need @itemtype when using RDFa vocabularies, or are full URL names enough? seems strange that using globally unique names without itemtype is ok for microdata vocabs but not RDFa ones |
| 15:53 | <foolip> | boblet, if I just fix the paths in my cv example the download link works |
| 15:53 | <foolip> | are you using chrome by any chance? |
| 15:53 | <boblet> | foolip: why yes :/ |
| 15:53 | <boblet> | KNO |
| 15:53 | <boblet> | |
| 15:53 | <foolip> | boblet, I think chrome has a bug with that |
| 15:53 | <foolip> | and I think you're the one who told me |
| 15:54 | <boblet> | foolip: lol. i should blog more so Google can tell me what I know |
| 15:54 | <boblet> | (I use “know” loosely here) |
| 15:54 | <foolip> | boblet, I can't think of any good way to fix it really, other than bouncing it off the server |
| 15:55 | <boblet> | foolip: what do you mean by that? generating a static file? |
| 15:56 | <foolip> | boblet, or having a microdata parser and vcard writer on the server side |
| 15:56 | <foolip> | I think that's the only sane option in real life, as with the data: approach you get a crappy file name |
| 15:56 | <brucel> | Who would like to tell me I'm wrong about input type=range? |
| 15:56 | <boblet> | foolip: so nothing of the level of “add these scripts and make a download link like this”-level easy |
| 15:57 | <foolip> | boblet, no, not that easy |
| 15:57 | <boblet> | brucel: you’re wrong |
| 15:57 | <brucel> | thought so, thanks boblet |
| 15:57 | <foolip> | brucel, you're right, file a bug |
| 15:57 | <foolip> | besides, we all know Hixie likes to fix bugs |
| 15:57 | <boblet> | foolip: sheesh. sucks |
| 15:57 | <brucel> | OK: input type=range min=100 max=130 step=10 |
| 15:58 | <boblet> | brucel: so what exactly are you wrong about? |
| 15:58 | <foolip> | about XML literals, I don't know how widely used that is in RDF, but am guessing it only exists because it was relatively easy to make with RDF/XML |
| 15:58 | <brucel> | no value defined, so spec says ""The default value is the minimum plus half the difference between the minimum and the maximum, unless the maximum is less than the minimum, in which case the default value is the minimum." |
| 15:58 | <brucel> | " |
| 15:58 | <foolip> | boblet, and there's no such thing as RF |
| 15:58 | <foolip> | RDFa vocabularies, only RDF |
| 15:59 | <boblet> | foolip: aah, thanks. still mighty confused by RDF(a)’s acronym soup |
| 15:59 | <brucel> | but that would make default value = 115, which is invalid cos of the step attribute. Chrome and Saf set it to 120, the nearest valid value to the mid point. Which makes sense to me. |
| 16:00 | <brucel> | so is the spec more sensible than the chome/ saf behaviour? |
| 16:01 | <boblet> | brucel: seems like the spec should start “the devault value is *the closest value (defined by @step) to* the minimum plus…” |
| 16:02 | <foolip> | boblet, itemtype is involved in the RDF conversion algorithm, but I don't remember the exact details now |
| 16:02 | <boblet> | foolip: aah, I’ll check that then. thanks for the nudge |
| 16:03 | <brucel> | boblet, yes. Wondering why it doesn't. |
| 16:03 | <foolip> | I also don't really think anyone will ever use that algorithm, since it makes "ugly" RDF, and people who like RDF probably do it because they think it's elegant and such |
| 16:03 | <brucel> | Ugly RDF? tautology surely? |
| 16:03 | <foolip> | brucel, 110 is just as close as 120 |
| 16:03 | boblet | ducks |
| 16:04 | <boblet> | foolip: I’m thinking more of being able to use all the pre-defined RDF vocabularies as-is, without needing to make a microdata version |
| 16:05 | <boblet> | rather than with the intention of generating RDFa |
| 16:05 | <brucel> | foolip, yup. but both are (at least) valid; and rounding-up is the norm. |
| 16:05 | <foolip> | it'd be extremely messy to try to mix and match different vocabularies like you usually do with RDF |
| 16:05 | <foolip> | brucel, then the spec should say to round up (just saying) |
| 16:06 | <boblet> | brucel: you filing the bug, or are you gonna make me do it? |
| 16:06 | <foolip> | I will be leaving now |
| 16:07 | <foolip> | boblet, mail me if you need |
| 16:07 | <boblet> | thanks for your help, foolip! |
| 16:07 | <brucel> | boblet, well I ain't sure that it's a bug. It's weird but in HTML5 land that's not necessarily an indicator of incorrect |
| 16:08 | <boblet> | brucel: definitely worth filing. spec having a interpretation loophole = bug |
| 16:12 | <Matjas> | brucel: if you get me a free copy of your book I'll use chapter 5 to recreate that Flash thingy |
| 16:13 | <Matjas> | >] |
| 16:16 | <brucel> | Matjas you can't get the prize *before* the competiton |
| 16:17 | <Matjas> | YES I CAN I JUST ORDERED IT |
| 16:17 | <boblet> | lol |
| 16:18 | <boblet> | is it one of those “if you know this song call us to win a free CD of it now!” type competitions? |
| 16:19 | <nimbupani> | brucel: Matjas is winning it unfairly *mummmmmmy* |
| 16:20 | <Matjas> | boblet: sounds like it :) |
| 16:20 | <brucel> | ah, i just idly said that I'd give a free copy to 1st person to duplicate http://lab.andre-michelle.com/pulsate with canvas and html5 audio |
| 16:21 | <boblet> | nimbupani: don’t worry dear, I’ll have a chat to him… BRUCEL! |
| 16:21 | <boblet> | ;-) |
| 16:21 | <brucel> | Pat Lauke was gonna have a go using JS to generate random tones by making on-the-fly data URIs as src for <audio> when the circles collide |
| 16:21 | <paul_irish> | that's hot. |
| 16:21 | <boblet> | brucel: re: input type="range", I think the relevant bit is a bit below that beginning “When the element is suffering from a step mismatch…” |
| 16:22 | <boblet> | I don’t get it, it’s just that blue lego I see everywhere |
| 16:24 | <brucel> | I favoured an option to have radomised swear-words like Buffy's swearing keyboard (Google it, NSFW) |
| 16:24 | <brucel> | boblet looking |
| 16:28 | <brucel> | boblet you're right. I got confused by the change from "default value" to "When the element is suffering from a step mismatch". But it still doesn't define what you do in mny example above; as Chrome and Safari round up, suggest spec says that |
| 16:29 | <boblet> | good point |
| 16:36 | <boblet> | brucel: http://www.w3.org/Bugs/Public/show_bug.cgi?id=10122 |
| 16:40 | <brucel> | boblet by golly, that's groovier than Lou Reed duetting with Lady Gaga |
| 16:40 | gsnedders | dances |
| 16:40 | <gsnedders> | (with brucel) |
| 16:40 | <TabAtkins> | boblet: My post is turning pretty long. That ok? |
| 16:40 | <boblet> | brucel: crap — forgot to add raspberries to you for making me file it |
| 16:40 | <brucel> | TabAtkins long is good |
| 16:41 | <TabAtkins> | Excellent. |
| 16:41 | <brucel> | gsnedders you're a lovely mover |
| 16:41 | <boblet> | TabAtkins: sure. doubt it’s as long as the microdata one ;-) |
| 16:41 | <boblet> | brucel: consider yourself raspberried |
| 16:42 | <brucel> | bizarrely, was just eating a bowl of raspberries |
| 16:43 | <boblet> | hope they were making farty sounds |
| 16:44 | <boblet> | brucel: added raspberries to the bug. everything’s right in the world |
| 16:46 | <boblet> | jeez, back to formula |
| 16:53 | <brucel> | boblet thanks |
| 16:54 | <boblet> | np yo. gotta have your raspberries |
| 16:54 | <boblet> | partial to a nice boysenberry myself |
| 16:54 | <akamike> | i've always pictured you to be a man of the forbidden fruits boblet |
| 16:55 | <boblet> | akamike: just to check, are boysenberries forbidden atm? |
| 16:55 | <akamike> | I'm too young to know :( |
| 16:56 | <brucel> | boblet: remember the old adage: boysen is poison |
| 16:56 | <boblet> | akamike: lucky brucel was here! ;-) |
| 16:56 | <akamike> | ;) |
| 17:08 | <boblet> | nn all |
| 17:54 | <jgraham> | Philip`: No I hadn't, thanks |
| 17:59 | <micheil> | just a quick table on the state of browser support for websockets: http://wiki.github.com/miksago/node-websocket-server/browser-support |
| 18:12 | <Hixie> | there's nothing that quite illustrates why i have a problem with the a11y task force proposals than the problem they're having with the alt issue |
| 18:13 | <Hixie> | they literally have been at a loss to come up with rationale for some of their proposals for weeks now |
| 18:13 | <Hixie> | but they are still adamant that that's what the spec should say! because, you know, consensus! |
| 18:13 | <Hixie> | sigh |
| 18:34 | jgraham | forsees dull email in his future |
| 19:19 | <llrcombs> | is it possible to make clicks on an element pass through to the element beneath? |
| 19:21 | <TabAtkins> | You want the pointer-events property in CSS. |
| 20:59 | <llrcombs> | should HTML5 video's controls be shown onmouseover, or should they always be shown? |
| 20:59 | <TabAtkins> | UA choice. |
| 20:59 | <TabAtkins> | My personal preference is to always show them when the video isn't playing, and only on mouseover when it is playing |
| 21:02 | <llrcombs> | that's WebKit's behavior |
| 21:02 | <TabAtkins> | Precisely. ^_^ |
| 21:05 | <annevk> | jgraham, your future is a bit like my present, but maybe I'll make it future too by postponing some of it |
| 21:09 | <jgraham> | annevk: Yeah, may strategy so far is to put the future on hold |
| 21:09 | <jgraham> | *my |
| 21:11 | <gsnedders> | jgraham: You're on holiday. Go away. |
| 21:11 | <gsnedders> | annevk: I'm stealing your bike. |
| 21:12 | <jgraham> | gsnedders: I appear to be at home |
| 21:12 | <jgraham> | gsnedders: Also theft is wrong and you will probably go to hell (according to several major world religions) |
| 21:13 | <jgraham> | (but not according to me, so that's OK) |
| 21:13 | <gsnedders> | I'll put it back, and he'll never know it moved! |
| 21:13 | <jgraham> | gsnedders: Also, I would use %timeit in ipython to measure the performance of the two approaches |
| 21:13 | <jgraham> | Or realise that the one I preferred was fast enough |
| 21:13 | <gsnedders> | jgraham: Mine is quicker, as well as giving the right answer |
| 21:14 | <gsnedders> | (If you mean what you're talking about) |
| 21:14 | <gsnedders> | s/what/what I think/ |
| 21:14 | <jgraham> | gsnedders: Yes, but "one I preferred" wasn't a statement of a particular preference |
| 21:15 | <jgraham> | I just meant if you have a particular preference it makes sense to use it if it is fast enough |
| 21:15 | <gsnedders> | Yeah, true. I don't know which you prefer. :) |
| 21:15 | <jgraham> | (I guess the itertools approach is faster because it is lazier so you don't need to allocate huge arrays) |
| 21:15 | <TabAtkins> | ?_? |
| 21:16 | <gsnedders> | TabAtkins: Given an iterator in Python, compute the run lengths within it |
| 21:16 | <jgraham> | TabAtkins: The problem was "In python, calculate the average run length of the same value consecutively in an array" |
| 21:17 | <jgraham> | i.e. given [1,1,1,2,2,3] you would have one run of thength 3, one of length 2 and one of length 1 |
| 21:17 | <jgraham> | So the average would be 2 |
| 21:17 | <gsnedders> | My approach became: import itertools; runs = [len(tuple(x)) for x in itertools.groupby(foo)] |
| 21:17 | <gsnedders> | (then average runs) |
| 21:18 | <TabAtkins> | So just iterate, tracking each run, and when a run finishes, add it to the running average. |
| 21:18 | <jgraham> | gsnedders: Why construct an explicit lift? |
| 21:18 | <jgraham> | *list |
| 21:18 | <gsnedders> | jgraham: for len? you can't get the len of an iterator. |
| 21:18 | <jgraham> | gsnedders: You don't need to |
| 21:19 | <annevk> | gsnedders, are you now? beware of the booby traps |
| 21:19 | <jgraham> | TabAtkins: There was a theory that explicit loops are slow |
| 21:19 | <jgraham> | TabAtkins: That might still be the fastest way though |
| 21:19 | <TabAtkins> | How long is the list? |
| 21:20 | <gsnedders> | TabAtkins: There are around 200k lists each with 100 items in them |
| 21:20 | <TabAtkins> | I somehow doubt that an explicit loop is slower than an iterator that then rips the subiterators out into a tuple before measuring their lengths. |
| 21:21 | <gsnedders> | (and they aren't lists, they're iterators, from itertools._grouper)) |
| 21:21 | <TabAtkins> | Bah, lists and iterators are nearly identical. |
| 21:21 | <gsnedders> | Really not. |
| 21:21 | <TabAtkins> | You're not thinking abstractly enough. |
| 21:21 | <jgraham> | Similar at least |
| 21:22 | <Slaanesh> | They're different where it matters |
| 21:22 | TabAtkins | is constantly confused when he runs into something that doesn't work the same for the two. |
| 21:22 | <jgraham> | Conceptually the main difference is that you can only use an iterator once |
| 21:22 | <jgraham> | I think |
| 21:22 | TabAtkins | is kinda confused by Python iterators anyway, because they're impure. |
| 21:22 | <Slaanesh> | Conceptually, a list exists in full |
| 21:22 | <jgraham> | Like your Mum |
| 21:22 | <gsnedders> | jgraham: And iterators are lazily |
| 21:22 | <gsnedders> | evaluated |
| 21:22 | jgraham | couldn't resist, sorry |
| 21:23 | <Slaanesh> | To be fair, python seems to mix iterator and generator some |
| 21:23 | <gsnedders> | And as we concluded at lunch today, your mum jokes mean it's time to go |
| 21:23 | <jgraham> | Oh yes, indeed |
| 21:23 | <Slaanesh> | gsnedders: That's what she said |
| 21:23 | <gsnedders> | Yeah, that is confusing |
| 21:23 | <TabAtkins> | Slaanesh: Aren't the two identical in Python in terms of interface? |
| 21:23 | <jgraham> | Yes |
| 21:23 | <gsnedders> | TabAtkins: In terms of interface, yes |
| 21:24 | <jgraham> | It is quite beautiful really |
| 21:24 | <Slaanesh> | So the important part of iteritems() isn't the iterator (since lists are iterators too), but the generator aspect |
| 21:24 | <gsnedders> | yeah, indeed |
| 21:24 | <TabAtkins> | jgraham: I agree. |
| 21:25 | TabAtkins | wishes there was an easier way to get yield-style generator expressions in Lisp, so he didn't have to write his generators with explicit state. |
| 21:25 | <gsnedders> | I was talking to people about rewriting what I have in Haskell today, as so much of it is all relying upon lazy evaluation |
| 21:25 | <gsnedders> | (And Haskell is cool) |
| 21:25 | <gsnedders> | (And I can compile it to JS with YHC) |
| 21:25 | <TabAtkins> | You swedes and your haskell. |
| 21:25 | <TabAtkins> | (Haskell is pretty cool.) |
| 21:26 | <gsnedders> | TabAtkins: I'm going to Glasgow University, what do you expect? :P |
| 21:26 | <TabAtkins> | What's the term for people from the nordic peninsula? Nords? |
| 21:27 | <gsnedders> | Nordic |
| 21:27 | <gsnedders> | Nordics in plural |
| 21:27 | <TabAtkins> | kk |
| 21:29 | <Slaanesh> | Oh man, TGIF needs more filling food |
| 21:29 | <Slaanesh> | I don't have anything at home and now I'm hungry |
| 21:30 | <jgraham> | TabAtkins: Vikings |
| 21:30 | <TabAtkins> | jgraham: Makes sense. |
| 21:30 | <Slaanesh> | Superhumans |
| 21:36 | <TabAtkins> | Anyway, gsnedders' approach with itertools.groupby seems fine conceptually. It's simple and easy, and as long as the performance is acceptable, go crazy. |
| 21:37 | <gsnedders> | nested generators over an iterator for a database is all a bit of an overhead, though |
| 21:37 | <gsnedders> | Like, minutes of overhead. |
| 21:40 | <TabAtkins> | So then, if performance isn't acceptable, do it manually. Doing run-length computation is trivial. |
| 21:46 | <jgraham> | Oh no |
| 21:47 | <jgraham> | The TC39 people want to make ES6 versioned opt-in |
| 21:52 | <gsnedders> | TabAtkins: Won't help. Running it in a profiler shows 468s runtime, 305s of that being just iterating over the MySQL result set |
| 21:52 | <gsnedders> | MySQLdb is slow :\ |
| 21:53 | <AryehGregor> | So the WHATWG should make ES5 then, clearly. (I mean, TC39 jumped straight from ES4 to ES6, right? So the name is open.) |
| 21:53 | <AryehGregor> | (Failing that, call it JavaScript5, which is more reasonable anyway.) |
| 21:53 | <gsnedders> | AryehGregor: No, ES4 was abandoned, the latest spec is ES5 |
| 21:53 | <TabAtkins> | gsnedders: Ah, then you're screwed no matter what. That's liberating. |
| 21:54 | <gsnedders> | TabAtkins: Write my own MySQL Python module that's actually maintained? |
| 21:54 | <gsnedders> | Oh, wait, I won't maintain it :P |
| 21:54 | <jgraham> | gsnedders: (maintain the existing one would be the sane community spirited option) |
| 21:54 | <TabAtkins> | Can you optimize the query itself, or is it literally just iterating over the already-computed result that's the problem? |
| 21:55 | <gsnedders> | TabAtkins: The latter. |
| 21:55 | <gsnedders> | TabAtkins: I'm excluding the actual time to execute the query from that |
| 21:55 | <gsnedders> | (that's another 109s) |
| 21:56 | <gsnedders> | (which is only 41 seconds of runtime left for the loop contents) |
| 21:56 | <gsnedders> | (which is mostly just run length computation) |
| 21:56 | <TabAtkins> | Dang. |
| 21:57 | <gsnedders> | jgraham: The problem is you'd be better off totally rewriting it |
| 21:57 | <gsnedders> | Using Cython, moving way more stuff down into C |
| 21:58 | <AryehGregor> | gsnedders, oh, drat. Then I guess we'll have to go with JS5. |
| 21:58 | <AryehGregor> | Now, what's this problem you're talking about, on a high level? |
| 21:59 | <gsnedders> | No comment ;P |
| 22:00 | <gsnedders> | No, more seriously, computing stability of results from regression tracking system |
| 22:01 | <AryehGregor> | That sounds too much like applied math to be interesting to me. |
| 22:01 | <gsnedders> | Heh. |
| 22:01 | <AryehGregor> | (kind of funny how I like pure math and related programming, and also non-math programming, but applied math I find totally uninteresting) |
| 22:01 | <jgraham> | AryehGregor: WHATWG workign on Javascript likely wouldn't gain any traction |
| 22:02 | <AryehGregor> | Why not? Are there major non-browser consumers of JS that browsers need to be compatible with? |
| 22:02 | <jgraham> | unless TC39 disappear even more off into the weeds |
| 22:02 | <jgraham> | AryehGregor: Because TC39 has support from the major players |
| 22:03 | <jgraham> | Mozilla, Microsoft, etc. |
| 22:03 | <AryehGregor> | Not if they do opt-in versioning for EC6, right? :) |
| 22:03 | <jgraham> | what they are doing might fail, but we won't get buy-in to an alternative until it fails |
| 22:04 | <jgraham> | Unlike XHTML2 where we got a 4 year headstart |
| 22:04 | <jgraham> | because everyone looked at it and went "uh, no" |
| 22:04 | <jgraham> | Everyone important anyway |
| 22:12 | <llrcombs> | what's the difference between the "playing" event and the "play" event of a <video> tag? |
| 22:13 | <gsnedders> | AryehGregor: Also, MS aren't involved in WHATWG |
| 22:13 | <AryehGregor> | Well, no. |
| 22:13 | <AryehGregor> | But they've wound up implementing HTML anyway, and probably would have in the end even if it weren't at the W3C. |
| 22:14 | <AryehGregor> | llrcombs, I think I complained about that. The naming is confusing. |
| 22:14 | <AryehGregor> | Too late to change it, I guess. |
| 22:15 | <Slaanesh> | AryehGregor: Mozilla supports opt-in (to whatever degree Mozilla can support something) |
| 22:15 | <AryehGregor> | Really? |
| 22:15 | <AryehGregor> | That's surprising to me. |
| 22:15 | <gsnedders> | AryehGregor: Mozilla is still a major driving force within ES |
| 22:15 | <Slaanesh> | It was pretty much necessary for the original es4 |
| 22:15 | <Slaanesh> | considering the core changes |
| 22:15 | <gsnedders> | They probably have more control over it than the other browser vendors combined |
| 22:16 | <zcorpan_> | llrcombs: playing is fired when the video really starts playing. play is fired when play is requested |
| 22:16 | AryehGregor | doesn't follow ES. |
| 22:17 | <Slaanesh> | Mozilla was basically running javascript until es4 got started |
| 22:19 | <gsnedders> | In a lot of ways, it was not really much better than OOXML, ECMA pretty much just rubber-stamping it |
| 22:19 | <Slaanesh> | es5? |
| 22:19 | <gsnedders> | Before ES4 |
| 22:19 | <Slaanesh> | They didn't even stamp the stuff from js 1.5 to 2.0 |
| 22:19 | <llrcombs> | CS5? |
| 22:19 | <llrcombs> | wait lolwat? |
| 22:20 | <gsnedders> | Slaanesh: Well, yeah. :P |
| 22:22 | <Slaanesh> | I agree the original es4 went quite far into unbroken wilderness, but es5 was somewhat underwhelming |
| 22:22 | <gsnedders> | ES5 didn't go far enough at making the spec match what any browser will ever implement |
| 22:29 | <TabAtkins> | Heh, linking my demos more than double the total lifetime bandwidth expenditure for my site. ^_^ |
| 22:30 | <TabAtkins> | (And happened right when my account dropped below a dollar, so the two bucks or so that the swell cost knocked my site down temporarily.) |
| 22:32 | <Rik`> | TabAtkins: have you also measured the time people are loosing by downloading all those episodes for an almost forgotten cartoon ? |
| 22:33 | <TabAtkins> | NEVAR FORGET |
| 22:33 | <gsnedders> | NEVER! |
| 22:33 | <gsnedders> | NEVER! |
| 22:33 | <Slaanesh> | Which one? Gem? |
| 22:33 | <TabAtkins> | Samurai Pizza Cats |
| 22:33 | <TabAtkins> | Slaanesh: http://www.xanthir.com/video/demo3.html |
| 22:34 | <Slaanesh> | More like Samurai Making-stuff-up Cats |