| 00:30 | <roc_> | hmm |
| 00:30 | <roc_> | I am the stupidest person in the world |
| 00:31 | <roc> | spellcheck="true" does not turn off spellcheck in Firefox trunk |
| 00:31 | <roc> | spellcheck="false" does |
| 00:31 | <roc> | (and it works with contenteditable) |
| 00:35 | Hixie | receives spam with the subject line "Make a living on Google" |
| 00:35 | <Hixie> | i think i have that covered |
| 00:36 | <Philip`> | roc: data:text/html,<p contenteditable spellcheck="false">Slartibartfast still gives me a wavey red line |
| 00:37 | <Philip`> | ...at least in Gecko/20090119 Shiretoko/3.1b3pre |
| 00:40 | <Hixie> | Lachy: i think it's reasonable to assume that copy and paste is part of the problem statement, because that's a user-facing interface. |
| 00:41 | <Hixie> | Lachy: "i want to expose a way to have the user copy content from one document, paste it into my text editor, and have my text editor automatically generate the bibliographic entry" seems like a reasonable request. |
| 00:42 | <Lachy> | Hixie, that one is a copy and paste issue. But the other two about getting contact information from a web page into an address book, and getting an email from webmail into a native client, aren't copy and paste issues |
| 00:42 | <Lachy> | at least, they don't have to be |
| 00:43 | <Hixie> | sure |
| 00:48 | <roc> | Philip`: OK, it looks like spellcheck only works on "body" to control contenteditable spellchecking for the entire document :-( |
| 01:00 | <Philip`> | http://fonts.philip.html5.org/ |
| 01:00 | <Philip`> | Seems to work in at least FF3.1 and Opera 10 |
| 01:00 | <Philip`> | (It's pretty limited and buggy, but at least it can handle simple ASCII characters for a few fonts) |
| 01:17 | <Hixie> | rubys: i'm confused. what text about 'origin' are we discussing removing anyway? |
| 01:18 | <Hixie> | i thought the whole problem was that there was no definition and we needed an ID for one |
| 01:29 | <Lachy> | Hixie, based on the Origin thread, I assumed there was a specific section about Origin that was being moved out of the spec and into an ID. Is that not the case? |
| 01:29 | <Lachy> | perhaps I should look at the spec... |
| 01:29 | <Hixie> | i don't know |
| 01:29 | <Hixie> | nobody seems to have said what they are discussing |
| 01:29 | <Lachy> | wouldn't it be section 5.3 or 5.3.1? |
| 01:30 | <Lachy> | hmm, maybe not |
| 01:31 | <Hixie> | that's about scripting |
| 01:31 | <Hixie> | not the http header |
| 01:32 | <Lachy> | hsivonen said here that he'd discussed it with you http://lists.w3.org/Archives/Public/public-html/2009Jan/0210.html |
| 01:34 | <Lachy> | I see. The spec mentions XXX-Origin header in several places throughout the spec. But there doesn't seem to be anywhere that actually defines it. Now I'm confused too |
| 01:35 | <abarth> | politics makes my head hurt :( |
| 05:12 | <heycam> | annevk, http://twitter.com/waka/status/1132138169 |
| 07:28 | <Hixie> | i wish webkit had an upload progress bar |
| 08:11 | <annevk> | roc, also note that the spec says the values should be "on" and "off"... |
| 08:12 | <annevk> | roc, so maybe Chrome and Gecko do different things here? |
| 08:12 | <roc> | which spec? Ian's old spec doesn't |
| 08:12 | <roc> | http://www.damowmow.com/playground/spellcheck.txt |
| 08:14 | <annevk> | ah, but an even older one does: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2006-June/006762.html |
| 08:14 | <annevk> | sorry for the confusion |
| 08:19 | <annevk> | http://twitter.com/nlothian/statuses/1131828648 |
| 08:19 | <annevk> | (worth a read) |
| 08:21 | annevk | wonders if @color-profile is implemented |
| 08:24 | <Philip`> | http://blog.mozilla.com/rob-sayre/2009/01/19/where-memes-go-to-die/ (#7) |
| 08:25 | <deane> | annevk: you may already have seen this, but: http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Jan/0075.html http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Jan/0077.html |
| 08:28 | <annevk> | I hadn't |
| 08:28 | annevk | is not following public-rdf-in-xhtml-tf |
| 08:29 | annevk | thinks the voting Ian referred to would be in the HTMLWG, not the WHATWG |
| 08:48 | <deane> | yeah, Manu has it wrong, in fact the whole RDFa issue is a result of misinformation and misunderstandings, we wouldn't have this issue of RDFa deployment if the W3C admited publicly that only a handfull of people are using XHTML |
| 08:56 | <deane> | ...That way we wouldn't end up with specs that can't be implemented. |
| 09:01 | <annevk> | no big deal |
| 09:03 | <deane> | It's no big deal as long as everyone knows that HTML5 doesn't have an obligation to suport all these unimplementable specs |
| 09:03 | <annevk> | well, there is disagreement over that |
| 09:04 | <annevk> | and I think we've yet to establish who is right |
| 09:08 | <danbri> | deane, which spec re rdfa are you thinking is unimplementable? |
| 09:10 | <deane> | danbri: http://www.w3.org/TR/rdfa-syntax/ |
| 09:12 | <danbri> | that's 'rdfa in xhtml'; nobody is asking HTML5 to be XML-only. So I think this leaves the door open for figuring out which bits of it could be re-built on top of html5 without too much trauma |
| 09:12 | <danbri> | eg. a profile without using curies |
| 09:13 | <deane> | how can you implement a spec that is XHTML only, when there's only about one hundred people in the world using XHTML |
| 09:14 | <danbri> | this conversation isn't feeling very collaborative. i'm trying to talk to you about finding a version of that design which *is* implementable in html5... |
| 09:15 | <deane> | if you have to alter the spec to be usable in text/html, then the whole should be altered since the web is text/html, therefore the spec has no value |
| 09:16 | <danbri> | "if ... then .... therefore ...." |
| 09:16 | danbri | thinks |
| 09:16 | <danbri> | nope, you've lost me. |
| 09:16 | <danbri> | one more time? |
| 09:16 | <deane> | start again from scratch |
| 09:16 | <danbri> | i spent some time yesterday talking with henri, and looking at existing rdfa parser behaviour |
| 09:17 | <deane> | specs need to be written from scratch to suit text/html |
| 09:17 | <danbri> | idea was to find something that was close to what those parsers expect (i tested 6); and close to what could be done in html5 with no ugliness |
| 09:18 | <deane> | sure, so we are really starting from scratch then, right? |
| 09:18 | <danbri> | the rdfa spec tries to be a design that can be bound to specific concrete carrier languages (svg, atom, xhtml, ...) |
| 09:19 | <danbri> | binding to a non-xml language is more of a stretch; since you need to invent or avoid having a ns abbreviation mechanism |
| 09:19 | <danbri> | most of the dicussion to date has been around this re-inventing a ns prefixing mechanism |
| 09:19 | <danbri> | the idea we explored yesterday was about avoiding one by always using full URIs |
| 09:20 | <danbri> | ie. a compromise that might not be what either 'side' wants but which can still be useful |
| 09:21 | <deane> | I can understand that, but, not many people are using xhtml, most people using svg in the future will be using it in text/html which has no support for namespaces |
| 09:23 | <deane> | ...and people using xhtml in the future will be using xhtml5, not xhtml1, so unfortunately that spec we mentioned has no real value |
| 09:23 | <danbri> | so would you agree it was useful for me to spend my yesterday trying out rdfa tool support for a namespaces-free profile of rdfa? |
| 09:23 | <deane> | I think that's a step forward |
| 09:24 | <danbri> | notes are in http://svn.foaf-project.org/foaftown/2009/rdfa/tests/readme.txt if you're interested |
| 09:24 | <danbri> | unfortunately it seems the parsers mostly work on the html5 no-ns version, but so long as they find a magic hack, xmlns:http="http:" |
| 09:25 | <danbri> | henri said it's easier to make the validator tolerate this than to have it check real use of xml namespaces; but i hope we can approach from the other side too, and have rdf parsers not look for it |
| 09:26 | <danbri> | this would give a copy-and-paste-friendly profile of rdfa: subsets which, with some care, could be copy/pasted between html5 and xhtml5/atom/svg docs |
| 09:26 | <danbri> | (obviously you'd need to stick to pretty bland markup to avoid tripping up on other things, but that's life) |
| 09:30 | <deane> | has the RDFa crowd got any demo pages showing what they think RDFa could look like in a html5 page? I don't understand what the need is for namespaces or pseudo non xml namespaces |
| 09:33 | <deane> | actually, I don't think I've seen a concrete proposal put forward from the RDFa guys, I saw something about adding six attributes |
| 09:34 | <danbri> | see t5 and t6 examples in http://svn.foaf-project.org/foaftown/2009/rdfa/tests/ |
| 09:34 | <danbri> | those are what i was testing parsers with; one has xhtml boilerplate, the other htm5 |
| 09:35 | <danbri> | and http://svn.foaf-project.org/foaftown/2009/rdfa/tests/g2.html is one that breaks henri's current validation demo for html5+rdfa-minus-curies, cos it uses rev= |
| 09:36 | <danbri> | hi MikeSmith |
| 09:36 | <MikeSmith> | danbri: hej |
| 09:41 | Philip` | encounters a bug in the Font::TTF library |
| 09:42 | <Philip`> | which is surprising because I thought it'd take much less than a day to find one |
| 09:50 | Philip` | encounters a second bug in the Font::TTF library |
| 09:50 | <Philip`> | It's not going so well now :-( |
| 09:52 | <annevk> | have you tried error handling yet? |
| 09:54 | <Philip`> | What kind of error handling? |
| 09:54 | <annevk> | dunno, wrong tables, wrong set of bytes, etc. |
| 09:55 | <Philip`> | No - I'm just assuming the input is an approximately correct font |
| 09:57 | <annevk> | maybe I don't understand what you're doing |
| 10:00 | <Philip`> | I'm doing http://fonts.philip.html5.org/ |
| 10:01 | <annevk> | ah, cool |
| 10:01 | <annevk> | I thought you were testing font interpreters |
| 10:02 | <Philip`> | That's just a side-effect of trying to view fonts in browsers |
| 10:03 | <gsnedders> | Philip`: Trying a multi-byte character doesn't work well. |
| 10:05 | <Philip`> | gsnedders: I just want ASCII to work first :-) |
| 10:10 | <Philip`> | Currently I just want to be able to load a particular font and save it again, without it mysteriously stopping working in Firefox even though I can't see any non-trivial differences in the files |
| 10:22 | <Philip`> | Hmm, turns out that Firefox (or whatever font library it uses on Linux) is sensitive to the ordering of the tables, which is probably violating the TTF spec |
| 10:24 | <Philip`> | ...or maybe it isn't? |
| 10:24 | <Philip`> | I'm confused :-( |
| 10:25 | <Philip`> | Oh, yes, I was confused |
| 10:25 | Philip` | curses checksums that he forgot to update |
| 10:37 | <roc> | we use Freetype and Pango on Linux |
| 10:39 | <Philip`> | They seem to be working fine, and it's just my code that's being stupid |
| 11:13 | Philip` | decides to adopt a policy of not supporting fonts with CFF outlines, because he's lazy and doesn't care enough |
| 11:25 | <annevk> | rubys, oops, duh |
| 11:26 | annevk | should do something simple, like breaking his blog in all but beta browsers, while sick |
| 11:27 | <Dashiva> | You could make your blog all red and add lots of "comrade". |
| 11:28 | <annevk> | http://barslecht.nl/weblog/ |
| 11:28 | <annevk> | en ook de homepage http://barslecht.nl/ |
| 11:31 | <hsivonen> | Philip`: bummer. Wouldn't it be nice to support the Computer Modern Unicode family, which makes it possible to get the beloved LaTeX look on the Web |
| 11:34 | <jcranmer> | "[Put "almost" in front of most words in the following.] |
| 11:34 | <jcranmer> | " |
| 11:34 | <jcranmer> | okay... |
| 11:34 | <jcranmer> | The almost consistent almost DOM almost criteria is almost necessary but almost not almost not almost sufficient. |
| 11:36 | <hsivonen> | almost things like xml:lang take more than almost effort to support |
| 11:39 | <annevk> | but that would not have been the case if we could have put it in the XML namespace, right? |
| 11:41 | <annevk> | so many things are slightly broken because of design decision made in the past; makes me wonder what our mistakes are |
| 11:42 | <annevk> | according to olliej <canvas> should have had a Path object from the beginning, anything else? |
| 11:49 | <hsivonen> | annevk: when XML was specced, there were markup languages with attribute lang |
| 11:49 | <hsivonen> | and id |
| 11:49 | <hsivonen> | so the right way would have been to reserve the names id and lang in XML if cross-vocabulary ids and langs were considered important |
| 11:49 | <annevk> | oh yeah, no doubt they made lots of mistakes when speccing XML |
| 11:50 | <annevk> | but I'm sort of past the idea of getting that fixed (apart from maybe xml:id) |
| 11:50 | <hsivonen> | one of the mistakes being adding Namespaces to the XML layer instead of making it an RDF layer problem |
| 11:51 | <hsivonen> | if the damage done to XML is any indication, we should avoid adding CURIEs to HTML |
| 11:51 | <annevk> | (reasonable people would say here: I disagree) |
| 11:51 | <hsivonen> | SVG and MathML should get rid of xml:id, in my opinion |
| 11:51 | <hsivonen> | both have pre-existing id that works well enough |
| 11:52 | <annevk> | that would be a start, yes |
| 11:52 | <annevk> | as for CURIEs, they seem to hard to author to me |
| 11:52 | <annevk> | but I can also see that having to register short names to be used as prefix is a non-starter for many RDF folks |
| 11:53 | <annevk> | though maybe there is some middle way |
| 11:53 | <annevk> | where you either use a predefined prefix or just use the full URI |
| 11:53 | <hsivonen> | one of the fundamental problem that RDF-the-model has is that it uses URIs as identifiers and URIs are too long |
| 11:53 | <annevk> | that would at least survive under copy & paste |
| 11:54 | <hsivonen> | so RDF has had various syntaxes created for it |
| 11:54 | <hsivonen> | and the syntaxes (except N-Triples, yay for N-Triples) try to somehow make the length of the URIs disappear |
| 11:55 | <hsivonen> | but instead of making things simpler and shorter, the RDF serializations always create more cruft and complication when they try to make URIs shorter |
| 11:55 | <hsivonen> | so in time each serialization is declared as sucky, a new one is created and around we go again |
| 11:56 | <annevk> | has it been established that the URL length is the main issue with the serializations? |
| 11:56 | <annevk> | or the abbreviation mechanism |
| 11:57 | <annevk> | or are there other reasons why e.g. one might not like RDF/XML? |
| 11:58 | <hsivonen> | annevk: RDF serializations in practice if they abbreviate URIs want to have more than one URI prefix in scope, so just declaring a base won't work |
| 11:59 | <annevk> | yeah, but they allow for multiple abbreviations |
| 11:59 | <hsivonen> | annevk: and, yes, if you consider how other serializations of the same family differ from N-Triples, one of the main things they tend to address is URI length |
| 11:59 | <annevk> | ok |
| 11:59 | <hsivonen> | annevk: I mean, if you have an abbreviation mechanism, you can't use the one that works for HTML (i.e. base plus relative) |
| 12:00 | <hsivonen> | without being able point to one of many active bases |
| 12:01 | <annevk> | "Furthermore, experience in the wild (notably with SVG) shows that as soon as you have two versions a non-negligible subset of all documents start being labelled with the wrong version, meaning you now have a lot of useless metadata on your hands." |
| 12:01 | annevk | wonders when berjon will let go of the angle brackets :) |
| 12:02 | <annevk> | hsivonen, yeah, we can address that though, e.g. give <html> a prefix attribute or some such |
| 12:02 | <annevk> | but the main problem then is copy & paste |
| 12:04 | <hsivonen> | annevk: nope, there are more problems left even in that case: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-January/018283.html |
| 12:06 | <annevk> | I don't think "Negative savings in syntax length when I given prefix is only used a couple of times in a file." is correct |
| 12:06 | <annevk> | as typically those URLs are long |
| 12:06 | <hsivonen> | ok, s/couple of times/once/ |
| 12:07 | <annevk> | prefix="x:{url}" is 10 characters plus 2 characters each time you use it |
| 12:07 | <annevk> | well ok :) |
| 12:08 | <annevk> | but yeah, those concerns seem valid, although using registered prefixes instead requires maintaining a massive table of RDF vocabularies |
| 12:08 | <annevk> | but I'm not sure if you suggested something like that |
| 12:19 | <Philip`> | hsivonen: Computer Modern would indeed be nice; if I don't decide this project is a waste of time that nobody cares about and give up on it, I suppose CFF support would be a worthwhile thing to add in the future |
| 12:20 | Philip` | has no idea if "CFF" is the right term to use, but that's the name of the OpenType table that the offending fonts store glyph data in |
| 12:43 | <zcorpan> | annevk: is it a problem to have a massive table of rdf vocabularies? |
| 12:43 | <zcorpan> | we have a massive table of entities |
| 12:43 | <annevk> | shepazu, "<shepazu> definitely wrt URLs" HTML5 does not ignore IETF standards for URIs or IRIs, it just defines pre-processing when they are encountered within HTML attributes. According to e.g. Larry that is acceptable |
| 12:44 | <annevk> | zcorpan, hsivonen e-mail points it out as something that would be annoying, I don't feel strongly either way |
| 12:45 | <hsivonen> | zcorpan: Entities aren't open-ended. there's a new RDF vocabulary born every minute |
| 12:45 | <zcorpan> | hsivonen: ok |
| 12:48 | <annevk> | http://lists.w3.org/Archives/Public/public-ietf-w3c/2009Jan/0003.html has a long discussion on HTML5 by danc, shepazu, some IETF folks, etc. |
| 13:08 | <takkaria> | and people want RDFa because of decentralised extensibility |
| 13:09 | <takkaria> | make a table of URLs and no-one will want it anymore. :) |
| 13:11 | <Philip`> | Oh no, I forgot about ligatures :-( |
| 13:27 | <jgraham> | takkaria: Indeed. It seems like you could have a canonical table and allow people to use short prefixes from the table if they want. ALthough I guess there are problems if you want to append to the table and clients need to update |
| 13:28 | <hsivonen> | like I said, all attempts to hide the length of URIs lead to more complexity |
| 13:30 | <rubys> | hsivonen: are you up for helping me think through a thought experiment on this subject? I'm not sure what the outcome is, but it might be worth exploring... |
| 13:31 | <hsivonen> | rubys: perhaps later this week, not today or tomorrow |
| 13:32 | <rubys> | ok |
| 13:47 | <Dashiva> | Another "we should force implementors to implement what we want" post |
| 13:48 | <takkaria> | my brother made an amusing comment on identifying everything with URIs |
| 13:49 | <takkaria> | "makes me understand why we don't identify ourselves with genealogies anymore" |
| 13:50 | <Philip`> | Did people ever identify themselves with genealogies, other than in Tolkien? |
| 13:51 | <Dashiva> | Vikings |
| 13:51 | <takkaria> | the bible has a fair bit of it too |
| 13:51 | <takkaria> | anyway, I thought it was worth sharing, regarldess of factual accuracy. :) |
| 13:54 | <jgraham> | If only you had shared in in the form of an RDF triple we could have merged it with other humerous remarks |
| 13:55 | <takkaria> | jokes are not well representable in RDF, sadly |
| 13:55 | <Dashiva> | I propose using http://irc.whatwg.org/2009/01/20/13/55/52/0 to denote the statement |
| 13:56 | <Philip`> | I once heard a humerous remark from my fibula |
| 13:57 | <Dashiva> | This? http://dashiva.net/img/humerus.jpg |
| 14:07 | <zcorpan> | Hixie: s/with and height/width and height/ (in the img section) |
| 14:08 | <jgraham> | If jokes can't be represented in RDF, how will the film "Short Circuit" ever become reality? Clearly we need a solution for the use case of making friends with robots |
| 14:09 | <Dashiva> | Surely robots will find RDF itself funny enough :P |
| 14:16 | <zcorpan> | Hixie: is http://www.libpr0n.com/tests/frames/004/004.html invalid? (i get the same result in opera, firefox, ie8 and safari) |
| 14:48 | <Lachy> | oh no. The doctype thread is degrading into a versioning debate again :-( |
| 14:48 | Lachy | adds that to my list of threads to not respond |
| 15:02 | <annevk> | you can probably use the http://www.w3.org/2005/Incubator/emotion/XGR-emotionml/ vocabulary to express humor in RDF |
| 15:05 | <annevk> | brucel suggested Mr Last Week might be Hixie gone schizophrenic, as in Fight Club :p |
| 15:06 | jgraham | already suggested My Last Week was in fact Hixie |
| 15:07 | Philip` | fixes a bug that probably caused gsnedders to see multi-byte characters not working |
| 15:07 | <Philip`> | (Characters like â are typically stored as composite glyphs, so I need to recurse into those to work out which component glyphs need to be included too, and then renumber them all) |
| 15:08 | <Philip`> | (This all seems like a bit of a pain really) |
| 15:10 | <annevk> | maybe that's why nobody has done it yet ;) |
| 15:12 | <annevk> | meanwhile Last Week has a "follow your leader" post http://lastweekinhtml5.blogspot.com/2009/01/stark-choice-for-html5-and-future-of.html :p |
| 15:12 | <Philip`> | But it's not a lot of pain, so someone else should have already spent a few days doing the same kind of thing |
| 15:12 | <Philip`> | (but not that I can find anywhere) |
| 15:13 | <Philip`> | (and anyway it's more fun to rewrite it) |
| 15:22 | Philip` | finds that Opera does really crazy things, like using entirely the wrong font for some paragraphs |
| 15:29 | <Philip`> | It seems to pretty much ignore the font-family <-> src mapping in the CSS, and just make up its own mapping |
| 15:29 | <Lachy> | Philip`, file some bugs |
| 15:32 | <Philip`> | Oh, how annoying, it works correctly if I upload the fonts somewhere |
| 15:32 | <Philip`> | Oh, maybe it's just a stale cache |
| 15:33 | <Philip`> | Hmm, yes, that was it |
| 15:39 | <takkaria> | oh, heh, I'm on the list for who could be the biggest smeghead |
| 15:39 | <takkaria> | win |
| 16:58 | gsnedders | realizes it's actually impossible to do footnotes, pretty much |
| 17:00 | <jgraham> | gsnedders: Try taking off your shoes |
| 17:25 | <danbri> | humm http://www.whitehouse.gov/ is xhtml served as text/html |
| 17:26 | <gsnedders> | Because serving XHTML as text/html is really helpful when you want to move to real XHTML! |
| 17:28 | <Philip`> | All you need is to wait until IE supports XHTML, and then flip your web server's big red switch labelled "application/xhtml+xml", and the world will become many times zazzier than before! |
| 17:45 | Philip` | marvels at the ligatures now visible in the word "fluffiest" |
| 17:45 | Philip` | doesn't marvel so much at the completely broken spacing in "التلفون" |
| 18:25 | <Philip`> | Hmm |
| 18:26 | <Philip`> | My multipage splitter script has an instance of curl which has been attempting to downloading the spec for 21 days and 19 hours |
| 18:28 | <rubys> | sweet, not only is whitehouse.gov well-formed, it also has a number of links to feeds. Atom feeds. :-) |
| 18:28 | <gsnedders> | Philip`: Hmm… Has it started getting a response yet? |
| 18:29 | <Philip`> | rubys: It'd be more impressive if the feeds weren't 0 bytes in length |
| 18:29 | <rubys> | http://feedvalidator.org/check.cgi?url=http%3A%2F%2Fwww.whitehouse.gov%2Ffeed%2Fblog.aspx |
| 18:30 | rubys | is not thrilled with '.aspx' |
| 18:30 | <gsnedders> | 'Your feed appears to be encoded as "utf-8", but your server is reporting "US-ASCII"' — ergh. |
| 18:30 | <gsnedders> | And duplicate IDs. |
| 18:31 | <rubys> | Content-Type: text/xml |
| 18:31 | <gsnedders> | Yeah, I was guessing that was the cause of the first error |
| 18:31 | <rubys> | ewwww.... *all* of the ids are the same. |
| 18:31 | <Philip`> | gsnedders: No, but it's dumped a hundred megabytes of progress bar into my log file |
| 18:32 | <gsnedders> | rubys: At least it means it's easier to work around, having duplicate IDs in the feed at the same time |
| 18:32 | <gsnedders> | Philip`: Yay :\ |
| 18:33 | <rubys> | time for a blog entry. :-) |
| 18:33 | gsnedders | should blog more :) |
| 18:36 | <karlcow> | http://www.legacy.com/SeattleTimes/DeathNotices.asp?Page=LifeStory&PersonId=122901048 |
| 18:36 | <karlcow> | Norbert Hannes Mikula |
| 18:37 | <karlcow> | http://lists.w3.org/Archives/Public/w3c-sgml-wg/1997Feb/0054.html |
| 19:26 | <Philip`> | Ooh, excellent |
| 19:27 | <Philip`> | If I use a combining diacritic in a line of text that uses @font-face, then Safari 3.1 scrunches up all the characters from the whole line on top of each other |
| 19:27 | <Philip`> | but only when I run from a local web server, not when I upload it to somewhere else :-/ |
| 19:27 | <Philip`> | (even after I empty my cache) |
| 19:27 | <Dashiva> | charset thing? |
| 19:29 | <Philip`> | Oh |
| 19:29 | <Philip`> | Oops |
| 19:29 | <Philip`> | Probably because it's being served as text/plain on the remote server |
| 19:31 | <Philip`> | Oh, that's not it, the font is now application/octet-stream on both |
| 19:34 | <Philip`> | Oh, right, I'm still just being an idiot |
| 19:34 | <Philip`> | since the version on the remote web server pointed to a non-existent URL for the font |
| 19:40 | <Philip`> | Could someone say what http://philip.html5.org/tests/font/combining-chars.html looks like in a recent Safari on OS X? |
| 19:42 | <takkaria> | ooks fine to me |
| 19:42 | <takkaria> | in Safari 3.2.1 |
| 19:42 | <takkaria> | same text on different lines in different fonts |
| 19:43 | <rubys> | second line looks like a smaller font |
| 19:45 | <rubys> | http://intertwingly.net/tmp/combining-chars.png |
| 19:46 | <Philip`> | Okay, thanks |
| 19:46 | <Philip`> | I get http://philip.html5.org/tests/font/combining-chars-safari-win.png |
| 19:46 | <Philip`> | which is suboptimal from a readability perspective |
| 19:47 | <Dashiva> | But can you live with it? |
| 19:47 | rubys | chuckles |
| 19:48 | <Philip`> | I tend not to use many combining diacritics in my daily writing, so I suppose it's not *that* much of a problem |
| 19:49 | takkaria | is happy that he started blackholing html5-related mail at teh beginning of the month |
| 19:52 | <Philip`> | Hmph, I tried replacing the offending characters in my test page with the text "(Safari sucks)", and it turns out that one of my test fonts implements ')' as an upside-down '(', which reveals that my font subsetter has another bug and breaks the '(' :-( |
| 19:53 | <Philip`> | Uh |
| 19:53 | <Philip`> | ...breaks the ')' :-( |
| 19:54 | <Philip`> | ...except in Opera, which doesn't use the font's glyphs for either '(' or ')' |
| 19:55 | <Philip`> | ...Oh, sorry, that's just it caching too much again |
| 19:58 | <gsnedders> | Is there any way to hide the first and last characters of a string using CSS? |
| 19:59 | Philip` | discovers http://www.fontembedding.com/eot/ |
| 20:07 | <roc> | gsnedders: for the first, you might be able to use first-letter if you're really lucky |
| 20:08 | <gsnedders> | roc: first char. is punctuation, so I'd get that matching the first two chars. fun. |
| 20:08 | <roc> | Then, I guess, no |
| 20:09 | <gsnedders> | The only vaguely possible suggestion I've got involves XSLT. Yay. |
| 20:11 | <roc> | that's not CSS |
| 20:12 | <gsnedders> | No, it isn't. |
| 20:12 | <gsnedders> | I also don't want to learn XSLT for this. |
| 20:45 | <Philip`> | Now I can convert my subsetted fonts into EOT so they work in IE |
| 20:46 | <Philip`> | which I suppose is a good thing, even if EOT is evil, because otherwise nobody will use these fonts when they don't work in IE |
| 22:04 | <hsivonen> | there's something wrong with a favicon when its weirdness makes it to the BBC News front page: http://news.bbc.co.uk/2/hi/uk_news/magazine/7839744.stm |
| 22:20 | <Philip`> | hsivonen: Seeing the last entry on that page ("I wanted to show someone using their hands to open the BBC and see inside."), I guess they didn't learn from http://www.flickr.com/photos/qwghlm/529967993/ |
| 22:28 | <jruderman> | heh, BBC itself has a terrible favicon |
| 23:14 | <doodlewarrior> | what's the correct markup to include a phone number? |
| 23:14 | <doodlewarrior> | i imagine it goes into an address element |
| 23:14 | <doodlewarrior> | should i wrap the number itself in a tag? |
| 23:16 | <webben> | doodlewarrior: is the phone number contact information for the author of the page (or at least that distinctive section of the page?) |
| 23:16 | <webben> | http://dev.w3.org/html5/spec/Overview.html#the-address-element |
| 23:16 | <doodlewarrior> | yes |
| 23:16 | <doodlewarrior> | its a feedback viewer |
| 23:17 | <doodlewarrior> | theres a header element containing info about the user who submitted the feedback |
| 23:17 | <doodlewarrior> | followed by his content |
| 23:17 | <doodlewarrior> | i think i found what i need |
| 23:17 | <doodlewarrior> | href='tel:5551212' |
| 23:17 | <webben> | there's no special markup for phone numbers specifically afaik although I would note http://en.wikipedia.org/wiki/E.123 and http://microformats.org/wiki/hcard |
| 23:18 | <webben> | hmm, does anything implement tel: ? |
| 23:18 | <webben> | Skype? |
| 23:18 | <tantek> | in practice phone numbers are nearly always present as part of contact info, and thus should be marked up with hCard as noted. |
| 23:19 | <doodlewarrior> | mobile phones, for one |
| 23:20 | <webben> | point |
| 23:20 | <webben> | (though I think mobile phones probably try to make phone-number-like strings phoneable) |
| 23:20 | <doodlewarrior> | generally yes |
| 23:20 | <doodlewarrior> | but as long as im marking it up i might as well do it properly |
| 23:21 | <webben> | true |
| 23:21 | <doodlewarrior> | http://developer.apple.com/webapps/designingcontent.php |
| 23:21 | <doodlewarrior> | Integrate with Phone, Mail, Maps, and YouTube |
| 23:21 | <doodlewarrior> | symbian and winmo also support it |
| 23:22 | <webben> | interesting |
| 23:22 | <doodlewarrior> | thx |
| 23:39 | Hixie | puts his gloves back on and begins working on html5 again |
| 23:39 | <Hixie> | right |
| 23:40 | <Hixie> | how does navigating to a javascript: URI that returns a string affect the baseURI of the frame? |
| 23:42 | Hixie | wishes he'd get as many replies to technical questions like this as he does to questions about the introduction section |
| 23:42 | <Hixie> | maybe we should have a system where you have to contibute technical feedback before you're allowed to have an opinion on simpler stuff |
| 23:42 | <Hixie> | that would reduce the bikeshedding... |
| 23:47 | <doodlewarrior> | Hixie: i'm not the most advanced js user, so take this all with salt |
| 23:48 | <doodlewarrior> | but i was under the impression that a javascript:my_function() link would be a prettier way to do onclick='my_function' |
| 23:49 | <Hixie> | i mean something like http://www.hixie.ch/tests/adhoc/uri/javascript/003.html |
| 23:49 | <Hixie> | where an iframe has its location set to a javascript: URL that returns a string |
| 23:50 | <Hixie> | opera and firefox seem to agree that the base uri is the base uri of the script that was evaluated |
| 23:50 | <doodlewarrior> | you are officially deeper into JS than i |
| 23:50 | <Hixie> | hehe |
| 23:50 | <Hixie> | no worries |
| 23:51 | <doodlewarrior> | this is a nit |
| 23:51 | <doodlewarrior> | but i was just looking at the address record |
| 23:51 | <doodlewarrior> | in the example <ADDRESS> is capped |
| 23:52 | <doodlewarrior> | isn't the current preferred styling for tags lowercase? |
| 23:52 | <Hixie> | preferred by whom? |
| 23:52 | <doodlewarrior> | fair point |
| 23:53 | <doodlewarrior> | but a general styling consensus |
| 23:53 | <doodlewarrior> | the same way that in python this_var is prefered to thisVar |
| 23:53 | <Hixie> | as far as the spec cares, it makes no difference |
| 23:53 | <Hixie> | lowercase is the case used in xhtml |
| 23:53 | <doodlewarrior> | thanks for clarifying :) |
| 23:53 | <doodlewarrior> | and for kicking ass on the new spec |
| 23:53 | <Hixie> | but other than that it doesn't really matter |
| 23:53 | <Hixie> | (uppercase is the case used by the DOM) |
| 23:53 | <doodlewarrior> | i've been writing a new app in html 5 |
| 23:53 | <Hixie> | yeah? |
| 23:54 | <doodlewarrior> | yeah |
| 23:54 | <doodlewarrior> | very much in dev |
| 23:54 | <doodlewarrior> | but imaregular.com |
| 23:55 | <doodlewarrior> | the data- property is very much my friend |
| 23:55 | <Hixie> | cool |
| 23:56 | <doodlewarrior> | :) |
| 23:56 | <doodlewarrior> | not really relevant, just figured id share |
| 23:56 | <Hixie> | always good to hear about people liking html5 :-) |
| 23:57 | <Hixie> | especially these days, where a lot of hte feedback is from groups who have been spurned by html5 :-) |
| 23:57 | <doodlewarrior> | as a matter of fact. . . |
| 23:58 | <doodlewarrior> | regarding boolean attributes |
| 23:58 | <doodlewarrior> | it would be nice if their was a caveat in the spec that value='true' and value='false' are not valid |
| 23:58 | <doodlewarrior> | since the term boolean is defined to be true/false in so many other web languages |
