From ian at hixie.ch Wed Jun 10 00:05:08 2009 From: ian at hixie.ch (Ian Hickson) Date: Wed, 10 Jun 2009 07:05:08 +0000 (UTC) Subject: [html5] a question about canvas element's shadow effect In-Reply-To: <2E9A38A0754E4E22B85E659E028883ED@FNSTcyp> References: <2E9A38A0754E4E22B85E659E028883ED@FNSTcyp> Message-ID: On Wed, 20 May 2009, caoyipeng wrote: > > canvas element's shadow effect use Gaussian blur to fulfill. > > In Gaussian blur, when do convolution operation, Gaussian template's size(or > Gaussian kernel width and height) affect the blur effect > > ,is there default value about Gaussian template's size(such as the size is > (6?+1)*(6?+1),? is the standard deviation). The bigger the better. This is just a quality of implementation issue. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From ian at hixie.ch Thu Jun 11 12:16:46 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:16:46 +0000 (UTC) Subject: [html5] Microdata and Meta In-Reply-To: References: Message-ID: On Mon, 25 May 2009, Sean B. Palmer wrote: > > I know the Microdata stuff isn't stable yet, but I was just wondering > if the following is a valid and sensible thing to do: > > > > Microdata Example > > It's valid, but I would urge people to just use normal elements instead, as in: Microdata Example -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From ian at hixie.ch Thu Jun 11 12:17:50 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:17:50 +0000 (UTC) Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: References: Message-ID: On Mon, 25 May 2009, Sean B. Palmer wrote: > > HTML 5 currently says that img/@width and so forth use CSS 2.1 pixel > values. Pixel values in CSS 2.1 are of course a relative rather than an > absolute unit. In other words, CSS px is a non-linear alias for radians; > 675.522px, for example, is a quarter of a radian. > > But non-vector image formats like JPG and PNG have discrete pixels, and > screens have pixel displays with certain ppi values. So has anybody > looked at whether browsers actually implement px as a relative unit? And > how do they render images? I'm presuming that they don't resample to fit > the relative measurement. It's hard to say today because monitors are all assumed to be 96dpi. The idea of using CSS pixels everywhere is to make the spec be ready for when screens have higher resolutions. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From mrmazda at ij.net Thu Jun 11 12:56:58 2009 From: mrmazda at ij.net (Felix Miata) Date: Thu, 11 Jun 2009 15:56:58 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: References: Message-ID: <4A31618A.8030809@ij.net> On 2009/06/11 19:17 (GMT) Ian Hickson composed: > monitors are all assumed to be 96dpi. OS X makes this assumption. Windows makes this assumption by default, but enables application of alternative assumptions. One particular alternative, 120, is commonly applied prior to sale by vendors of laptops having materially above average DPI. Other operating systems vary whether and to what extent assumptions about display DPI are made, particularly when display DPI is in fact greater than 96. -- "Cast but a glance at riches, and they are gone, for they will surely sprout wings and fly off to the sky like an eagle." Proverbs 23:5 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://fm.no-ip.com/ From ian at hixie.ch Thu Jun 11 12:58:53 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:58:53 +0000 (UTC) Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <4A31618A.8030809@ij.net> References: <4A31618A.8030809@ij.net> Message-ID: On Thu, 11 Jun 2009, Felix Miata wrote: > On 2009/06/11 19:17 (GMT) Ian Hickson composed: > > > monitors are all assumed to be 96dpi. > > OS X makes this assumption. Windows makes this assumption by default, > but enables application of alternative assumptions. One particular > alternative, 120, is commonly applied prior to sale by vendors of > laptops having materially above average DPI. Other operating systems > vary whether and to what extent assumptions about display DPI are made, > particularly when display DPI is in fact greater than 96. Even on Windows with a non-96dpi, browsers today still assume 96dpi, as far as I can tell. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From mrmazda at ij.net Thu Jun 11 13:23:17 2009 From: mrmazda at ij.net (Felix Miata) Date: Thu, 11 Jun 2009 16:23:17 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: References: <4A31618A.8030809@ij.net> Message-ID: <4A3167B5.8030406@ij.net> On 2009/06/11 19:58 (GMT) Ian Hickson composed: > On Thu, 11 Jun 2009, Felix Miata wrote: >> On 2009/06/11 19:17 (GMT) Ian Hickson composed: >> > monitors are all assumed to be 96dpi. >> OS X makes this assumption. Windows makes this assumption by default, >> but enables application of alternative assumptions. One particular >> alternative, 120, is commonly applied prior to sale by vendors of >> laptops having materially above average DPI. Other operating systems >> vary whether and to what extent assumptions about display DPI are made, >> particularly when display DPI is in fact greater than 96. > Even on Windows with a non-96dpi, browsers today still assume 96dpi, as > far as I can tell. On Windows, some assume 96 (e.g. Safari, which I reported as a Safari bug "UI text is too small for high PPI display environment" https://bugreport.apple.com/cgi-bin/WebObjects/RadarWeb.woa/61/wo/APqcRdz7DDWD1jtFE8yeWw/3.79.28.0.9 and related Webkit bug http://bugs.webkit.org/show_bug.cgi?id=18001 ), while for others 96 is only a floor (e.g. Gecko). IE uses whatever DPI has been specified for the desktop generally. -- "Cast but a glance at riches, and they are gone, for they will surely sprout wings and fly off to the sky like an eagle." Proverbs 23:5 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://fm.no-ip.com/ From sirokai at gmail.com Fri Jun 12 08:31:10 2009 From: sirokai at gmail.com (Christian Montoya) Date: Fri, 12 Jun 2009 11:31:10 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <4A31618A.8030809@ij.net> References: <4A31618A.8030809@ij.net> Message-ID: <1a19d7d10906120831x36f24a31o5d795cdfd8c98492@mail.gmail.com> On Thu, Jun 11, 2009 at 3:56 PM, Felix Miata wrote: > On 2009/06/11 19:17 (GMT) Ian Hickson composed: > >> monitors are all assumed to be 96dpi. > > OS X makes this assumption. Windows makes this assumption by default, but > enables application of alternative assumptions. One particular alternative, > 120, is commonly applied prior to sale by vendors of laptops having > materially above average DPI. Other operating systems vary whether and to > what extent assumptions about display DPI are made, particularly when display > DPI is in fact greater than 96. 8 years ago I bought a Dell widescreen laptop, Inspiron 6000, and it was pre-configured at 120 DPI. I've been telling people ever since, 96 DPI is not a standard. Dell does this for a lot of widescreen laptops, and that's just 1 vendor. -- -- Christian Montoya mappdev.com :: christianmontoya.net From chris at martinilab.com Thu Jun 18 10:53:07 2009 From: chris at martinilab.com (chris at martinilab.com) Date: Thu, 18 Jun 2009 12:53:07 -0500 (CDT) Subject: [html5] toDataURL calling itself causes security error Message-ID: <1245347587.134132250@192.168.1.71> Clever me thinking I could come up with a undo script. I ended up with an error I can't get past. var img = new Image(); img.src = canvas.toDataURL(); ctx.drawImage(img,0,0); This only works once though. The next time I use line 2. img.src = canvas.toDataURL(); I get a security error even though img is created from the same script on the server. The only way around this (and very ineffeciently I light add) is to post save the canvas to an image on the server. This is happening on FireFox 3, I haven't checked Safari 4 though. Is this working as it should? Is it a bug? Is there a better workaround than posting to the server? Thanks! Chris Williams www.martinilab.com From excors+whatwg at gmail.com Thu Jun 18 11:02:29 2009 From: excors+whatwg at gmail.com (Philip Taylor) Date: Thu, 18 Jun 2009 19:02:29 +0100 Subject: [html5] toDataURL calling itself causes security error In-Reply-To: <1245347587.134132250@192.168.1.71> References: <1245347587.134132250@192.168.1.71> Message-ID: On Thu, Jun 18, 2009 at 6:53 PM, wrote: > [...] > var img = new Image(); > img.src = canvas.toDataURL(); > ctx.drawImage(img,0,0); > > This only works once though. ?The next time I use line 2. > img.src = canvas.toDataURL(); > > I get a security error even though img is created from the same script on the server. > [...] > This is happening on FireFox 3, I haven't checked Safari 4 though. > > Is this working as it should? ?Is it a bug? It's a bug according to the security rules in HTML 5 (or at least it was when I last checked), and was broken in all implementations (when I last checked, ages ago). Sounds like https://bugzilla.mozilla.org/show_bug.cgi?id=417836 fixes it, presumably for Firefox 3.5, but I haven't tested that. >?Is there a better workaround than posting to the server? Do you have to use toDataURL? e.g. could you construct a new temporary canvas element, then draw from the current canvas onto it, then draw back, instead of using an Image? i.e. something like var tmp = document.createElement('canvas'); tmp.width = canvas.width; tmp.height = canvas.height; ctx.drawImage(tmp, 0, 0); tmp.getContext('2d').drawImage(canvas, 0, 0); -- Philip Taylor excors at gmail.com From chris at martinilab.com Thu Jun 18 11:34:45 2009 From: chris at martinilab.com (chris at martinilab.com) Date: Thu, 18 Jun 2009 13:34:45 -0500 (CDT) Subject: [html5] toDataURL calling itself causes security error In-Reply-To: References: <1245347587.134132250@192.168.1.71> Message-ID: <1245350085.461817920@192.168.1.70> On Thu, Jun 18, 2009 at 6:53 PM, wrote: > It's a bug according to the security rules in HTML 5 (or at least it > was when I last checked), and was broken in all implementations (when > I last checked, ages ago). Sounds like > https://bugzilla.mozilla.org/show_bug.cgi?id=417836 fixes it, > presumably for Firefox 3.5, but I haven't tested that. Good to know! > Do you have to use toDataURL? e.g. could you construct a new temporary > canvas element, then draw from the current canvas onto it, then draw > back, instead of using an Image? i.e. something like > > var tmp = document.createElement('canvas'); > tmp.width = canvas.width; > tmp.height = canvas.height; > ctx.drawImage(tmp, 0, 0); > tmp.getContext('2d').drawImage(canvas, 0, 0); That works great actually. Thanks so much! Chris Williams http://www.martinilab.com From webmaster at keryx.se Sat Jun 27 04:22:39 2009 From: webmaster at keryx.se (Keryx Web) Date: Sat, 27 Jun 2009 13:22:39 +0200 Subject: [html5] Trying to help, please help me Message-ID: <4A4600FF.1090104@keryx.se> Hello all! I have updated my Best Practice Sheet at http://keryx.se/resources/html-elements/ It now has an HTML 5 doctype and some experimental CSS transformations (Only for the Fox 3.5 so far). All feedback about the content is appreciated. I have also written the following on a blog of mine: http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html Title: Validation and doctype myths and (inconvenient) truths It's an effort to educate. Please let me know if anything can be improved. -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From zcorpan at gmail.com Sat Jun 27 11:47:59 2009 From: zcorpan at gmail.com (Simon Pieters) Date: Sat, 27 Jun 2009 20:47:59 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A4600FF.1090104@keryx.se> References: <4A4600FF.1090104@keryx.se> Message-ID: On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web wrote: > Hello all! > > I have updated my Best Practice Sheet at > > http://keryx.se/resources/html-elements/ is now allowed in XHTML5 too. Maybe it would make sense to have one column for HTML5 and one for XHTML5? is deprecated in HTML4. is allowed in HTML5 but only in SVG and MathML contexts. was dropped. (X)HTML5 doesn't have , , . The "?" label seems unused. I guess , , should be labeled "B". <noscript>, <ins>, <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, <embed>, <applet> are "I". <table> is marked "B" but is not display:block. I guess none of the frameset-related elements should have "B", since framesets can't be expressed with CSS. The <ruby>-related elements also have special values for 'display'. I'm not sure about <layer> and <nolayer> -- I think they're just unknown (and hence inline) in browsers these days. <figure> is an "N" element. <isindex> is a bit weird since the parser expands it into <form>...<input>. <bb> can be either normal or replaced, I think. <command> is "I". I think <details> is always "N". HTML 5 doesn't say that <meter> or <progress> are replaced elements. <img> and <object> can be either "R" or "N". <audio> can be either "R" or "I". The "I" for <source> is left-aligned while the others are centered. <map> is a normal inline element. I wouldn't say <frameset> is an "N" -- maybe leave the cell empty. Same for <frame>. I think <xml> is invisible in IE. > It now has an HTML 5 doctype and some experimental CSS transformations > (Only for the Fox 3.5 so far). Nice. Hixie: some tables in the spec could benefit from the same styling. > All feedback about the content is appreciated. > > I have also written the following on a blog of mine: > > http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html > > Title: Validation and doctype myths and (inconvenient) truths > > It's an effort to educate. Please let me know if anything can be > improved. > -- Simon Pieters From mgainty at hotmail.com Sat Jun 27 19:02:13 2009 From: mgainty at hotmail.com (Martin Gainty) Date: Sat, 27 Jun 2009 22:02:13 -0400 Subject: [html5] Trying to help, please help me In-Reply-To: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> Message-ID: <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> Goddag Simon for each HTML specificnation tag i see ? which converts to '<LF> using system font is this correct? this is good information laid out in a very intuitive matrix Sverige Altid! Martin ______________________________________________ Verzicht und Vertraulichkeitanmerkung/Note de d?ni et de confidentialit? Diese Nachricht ist vertraulich. Sollten Sie nicht der vorgesehene Empfaenger sein, so bitten wir hoeflich um eine Mitteilung. Jede unbefugte Weiterleitung oder Fertigung einer Kopie ist unzulaessig. Diese Nachricht dient lediglich dem Austausch von Informationen und entfaltet keine rechtliche Bindungswirkung. Aufgrund der leichten Manipulierbarkeit von E-Mails koennen wir keine Haftung fuer den Inhalt uebernehmen. Ce message est confidentiel et peut ?tre privil?gi?. Si vous n'?tes pas le destinataire pr?vu, nous te demandons avec bont? que pour satisfaire informez l'exp?diteur. N'importe quelle diffusion non autoris?e ou la copie de ceci est interdite. Ce message sert ? l'information seulement et n'aura pas n'importe quel effet l?galement obligatoire. ?tant donn? que les email peuvent facilement ?tre sujets ? la manipulation, nous ne pouvons accepter aucune responsabilit? pour le contenu fourni. > To: webmaster at keryx.se; help at whatwg.org > Date: Sat, 27 Jun 2009 20:47:59 +0200 > From: zcorpan at gmail.com > Subject: Re: [html5] Trying to help, please help me > > On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: > > > Hello all! > > > > I have updated my Best Practice Sheet at > > > > http://keryx.se/resources/html-elements/ > > <base> is now allowed in XHTML5 too. > > Maybe it would make sense to have one column for HTML5 and one for XHTML5? > > <basefont> is deprecated in HTML4. > > <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. > > <eventsource> was dropped. > > (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. > > The "?" label seems unused. > > I guess <html>, <listing>, <xmp> should be labeled "B". <noscript>, <ins>, > <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, <embed>, > <applet> are "I". <table> is marked "B" but is not display:block. I guess > none of the frameset-related elements should have "B", since framesets > can't be expressed with CSS. The <ruby>-related elements also have special > values for 'display'. I'm not sure about <layer> and <nolayer> -- I think > they're just unknown (and hence inline) in browsers these days. > > <figure> is an "N" element. <isindex> is a bit weird since the parser > expands it into <form>...<input>. <bb> can be either normal or replaced, I > think. <command> is "I". I think <details> is always "N". HTML 5 doesn't > say that <meter> or <progress> are replaced elements. <img> and <object> > can be either "R" or "N". <audio> can be either "R" or "I". The "I" for > <source> is left-aligned while the others are centered. <map> is a normal > inline element. I wouldn't say <frameset> is an "N" -- maybe leave the > cell empty. Same for <frame>. I think <xml> is invisible in IE. > > > > It now has an HTML 5 doctype and some experimental CSS transformations > > (Only for the Fox 3.5 so far). > > Nice. Hixie: some tables in the spec could benefit from the same styling. > > > > All feedback about the content is appreciated. > > > > I have also written the following on a blog of mine: > > > > http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html > > > > Title: Validation and doctype myths and (inconvenient) truths > > > > It's an effort to educate. Please let me know if anything can be > > improved. > > > > > -- > Simon Pieters > _______________________________________________ > Help mailing list > Help at lists.whatwg.org > http://lists.whatwg.org/listinfo.cgi/help-whatwg.org _________________________________________________________________ Windows Live? SkyDrive?: Get 25 GB of free online storage. http://windowslive.com/online/skydrive?ocid=TXT_TAGLM_WL_SD_25GB_062009 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.whatwg.org/pipermail/help-whatwg.org/attachments/20090627/1ed05aa7/attachment.htm> From webmaster at keryx.se Sun Jun 28 04:29:27 2009 From: webmaster at keryx.se (Keryx Web) Date: Sun, 28 Jun 2009 13:29:27 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> Message-ID: <4A475417.5070408@keryx.se> On 2009-06-28 04:02, Martin Gainty wrote: > Goddag Simon > > for each HTML specificnation tag i see ? which converts to '<LF> using > system font > is this correct? Actually it is the entity &#10003; in the source code. Already when viewing source it converts to a check-mark i FFox, a bug I suppose. Opera gets it right in its view-source! If it converts to line-feed that sounds like another bug to me. In what Browser? -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From webmaster at keryx.se Sun Jun 28 10:45:59 2009 From: webmaster at keryx.se (Keryx Web) Date: Sun, 28 Jun 2009 19:45:59 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> Message-ID: <4A47AC57.8070308@keryx.se> On 2009-06-27 20:47, Simon Pieters wrote: > On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: >> http://keryx.se/resources/html-elements/ > > <base> is now allowed in XHTML5 too. Missed that. Thank you. > Maybe it would make sense to have one column for HTML5 and one for XHTML5? Maybe for my next *major* update. > <basefont> is deprecated in HTML4. Ooops! > <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. Thats why it is marked "s" = "sometimes". > <eventsource> was dropped. I thought it just moved to a separate spec? Note to self: "Must pay better attention!" > (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. OK. I just took for granted that it would include all ruby markup from XHTML 1.1. And that the rest just wasn't spec'd yet. Has this been discussed? Where? (I have not got time to follow every conversation, so sometimes I miss stuff. That's why I am asking for help ;-) ) > The "?" label seems unused. Yes. Since my last revision it is. > I guess <html>, <listing>, <xmp> should be labeled "B". Will fix. > <noscript>, > <ins>, <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, > <embed>, <applet> are "I". Will fix. > <table> is marked "B" but is not > display:block. I guess none of the frameset-related elements should have > "B", since framesets can't be expressed with CSS. Perhaps I should be more clear about what I mean with "blockish"... This table is goes back more than a few years and is the result of talking to my students. It is based on my personal quick way of explaining this to complete newbies: 1. Block elements will cause line breaks and must be floated or positioned in order to have contents on their sides. 2. Inline elements may never contain block elements. Both rules apply to table. Anyway, I think it is probably wise to be more precise. I will change it. Today I also teach CSS from day one, making it more clear what I mean anyway. > The <ruby>-related > elements also have special values for 'display'. I'm not sure about > <layer> and <nolayer> -- I think they're just unknown (and hence inline) > in browsers these days. I thought they were ignored completely...? At least they should be, IMHO. > <figure> is an "N" element. Ooops! > <isindex> is a bit weird since the parser > expands it into <form>...<input>. Weird indeed, but an input control will be shown. Sounds replaced to me! > <bb> can be either normal or replaced, > I think. Anyone who can conform this? > I think <details> is always "N". Anyone who can conform this? > <command> is "I". I thought some control would be shown. in my rightmost column I have said tha "Button, radiobutton or checkbox" will be shown. > HTML 5 > doesn't say that <meter> or <progress> are replaced elements. That was deduction from my part. How else would they display if not through some widget? > <img> and > <object> can be either "R" or "N". You mean, usually "R", but "N" when providing text-fallback? > <audio> can be either "R" or "I". The > "I" for <source> is left-aligned while the others are centered. Sounds like it's too complex to describe with just one letter... > <map> is > a normal inline element. As indeed it is in my table. > I wouldn't say <frameset> is an "N" -- maybe > leave the cell empty. Same for <frame>. Sounds reasonable. Will change. > I think <xml> is invisible in IE. A quick test proves this to be the case. >> It now has an HTML 5 doctype and some experimental CSS transformations >> (Only for the Fox 3.5 so far). > > Nice. Hixie: some tables in the spec could benefit from the same styling. > It is a very fragile technique that requires fixed widths for table columns and is almost impossible to get pixel perfect right. I think it is a very clear use case for something that should be very easy to accomplish in CSS, and I have sent an e-mail to the CSS WG. So far I have received no answer from anyone, though. -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From major at minet.sk Sun Jun 28 10:55:45 2009 From: major at minet.sk (Juraj Major) Date: Sun, 28 Jun 2009 19:55:45 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A4600FF.1090104@keryx.se> References: <4A4600FF.1090104@keryx.se> Message-ID: <4A47AEA1.3090508@minet.sk> Hello! I think this issue about CDATA > Opera may ignore CSS if MIME is true XHTML is not right. <style><![CDATA[ #foo { css rules } ]]></style> These css rules are ignored if MIME is text/html. It's all right with */*+xml. From webmaster at keryx.se Tue Jun 30 02:56:23 2009 From: webmaster at keryx.se (Keryx Web) Date: Tue, 30 Jun 2009 11:56:23 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A47AC57.8070308@keryx.se> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <4A47AC57.8070308@keryx.se> Message-ID: <4A49E147.40209@keryx.se> On 2009-06-28 19:45, Keryx Web wrote: > On 2009-06-27 20:47, Simon Pieters wrote: > ... I think I have addressed all of your concerns. An updated version is now up at: http://keryx.se/resources/html-elements/ (I tried to thank you on twitter, but could not find you...) -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From webmaster at keryx.se Tue Jun 30 03:17:00 2009 From: webmaster at keryx.se (Keryx Web) Date: Tue, 30 Jun 2009 12:17:00 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A49E147.40209@keryx.se> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <4A47AC57.8070308@keryx.se> <4A49E147.40209@keryx.se> Message-ID: <4A49E61C.2080008@keryx.se> On 2009-06-30 11:56, Keryx Web wrote: > I think I have addressed all of your concerns. An updated version is now > up at: http://keryx.se/resources/html-elements/ FYI, here is the changelog: More detailed explanations. Print stylesheet, version is shown. <base> allowed in XHTML 5 (spec change) <basefont> is deprecated in HTML 4.01 (fix) <ventsource> has been dropped *as an element* (spec change) <html>, <listing>, <xmp> are B (block elements) (fix) <a> Added "May contain block elements in (X)HTML 5." <noscript>, <ins>, <del> <object>, <noembed> and <canvas> are now marked I* (display: inline, but may contain block elements) (pedagogic change) <embed>, <applet>, <meter>, <progress> and <keygen> are now marked I (inline) (fix) <keygen> is now marked R (replaced) (fix) Ruby elements are not marked inline anymore (blank) (pedagogic change) Ruby elements <rbc>, <rtc> and <rb> marked as ? for HTML 5 (Not in spec, maybe never will be?) <isindex> Is more described as they are treated by an HTML 5 parser. (pedagogic change) <layer>, <nolayer>, <ilayer>, <xml>, <multicol>, <spacer> have no information any longer about inline/block normal/invisible/replaced, since current behavior differs from the original intention (pedagogic change) Expanded info about <isindex> (Block, text appears) (pedagogic change) <details> is N (normal), not I/N/R (pedagogic change) <thead> repeats on each page when printing (fix) Table is now T (not B) (pedagogic change) -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From ian at hixie.ch Wed Jun 10 00:05:08 2009 From: ian at hixie.ch (Ian Hickson) Date: Wed, 10 Jun 2009 07:05:08 +0000 (UTC) Subject: [html5] a question about canvas element's shadow effect In-Reply-To: <2E9A38A0754E4E22B85E659E028883ED@FNSTcyp> References: <2E9A38A0754E4E22B85E659E028883ED@FNSTcyp> Message-ID: <Pine.LNX.4.62.0906100701551.1648@hixie.dreamhostps.com> On Wed, 20 May 2009, caoyipeng wrote: > > canvas element's shadow effect use Gaussian blur to fulfill. > > In Gaussian blur, when do convolution operation, Gaussian template's size(or > Gaussian kernel width and height) affect the blur effect > > ,is there default value about Gaussian template's size(such as the size is > (6?+1)*(6?+1),? is the standard deviation). The bigger the better. This is just a quality of implementation issue. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From ian at hixie.ch Thu Jun 11 12:16:46 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:16:46 +0000 (UTC) Subject: [html5] Microdata and Meta In-Reply-To: <b6bb4d890905250914x41f3c2a1q12a5d58cd89a22ad@mail.gmail.com> References: <b6bb4d890905250914x41f3c2a1q12a5d58cd89a22ad@mail.gmail.com> Message-ID: <Pine.LNX.4.62.0906111915370.1648@hixie.dreamhostps.com> On Mon, 25 May 2009, Sean B. Palmer wrote: > > I know the Microdata stuff isn't stable yet, but I was just wondering > if the following is a valid and sensible thing to do: > > <!DOCTYPE html> > <head item> > <title>Microdata Example</title> > <link itemprop="about" href=""> > <meta itemprop="example" content="value"> It's valid, but I would urge people to just use normal <meta> elements instead, as in: <!DOCTYPE html> <head> <title>Microdata Example</title> <meta name="example" content="value"> -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From ian at hixie.ch Thu Jun 11 12:17:50 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:17:50 +0000 (UTC) Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> Message-ID: <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> On Mon, 25 May 2009, Sean B. Palmer wrote: > > HTML 5 currently says that img/@width and so forth use CSS 2.1 pixel > values. Pixel values in CSS 2.1 are of course a relative rather than an > absolute unit. In other words, CSS px is a non-linear alias for radians; > 675.522px, for example, is a quarter of a radian. > > But non-vector image formats like JPG and PNG have discrete pixels, and > screens have pixel displays with certain ppi values. So has anybody > looked at whether browsers actually implement px as a relative unit? And > how do they render images? I'm presuming that they don't resample to fit > the relative measurement. It's hard to say today because monitors are all assumed to be 96dpi. The idea of using CSS pixels everywhere is to make the spec be ready for when screens have higher resolutions. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From mrmazda at ij.net Thu Jun 11 12:56:58 2009 From: mrmazda at ij.net (Felix Miata) Date: Thu, 11 Jun 2009 15:56:58 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> Message-ID: <4A31618A.8030809@ij.net> On 2009/06/11 19:17 (GMT) Ian Hickson composed: > monitors are all assumed to be 96dpi. OS X makes this assumption. Windows makes this assumption by default, but enables application of alternative assumptions. One particular alternative, 120, is commonly applied prior to sale by vendors of laptops having materially above average DPI. Other operating systems vary whether and to what extent assumptions about display DPI are made, particularly when display DPI is in fact greater than 96. -- "Cast but a glance at riches, and they are gone, for they will surely sprout wings and fly off to the sky like an eagle." Proverbs 23:5 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://fm.no-ip.com/ From ian at hixie.ch Thu Jun 11 12:58:53 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:58:53 +0000 (UTC) Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <4A31618A.8030809@ij.net> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> <4A31618A.8030809@ij.net> Message-ID: <Pine.LNX.4.62.0906111958330.16244@hixie.dreamhostps.com> On Thu, 11 Jun 2009, Felix Miata wrote: > On 2009/06/11 19:17 (GMT) Ian Hickson composed: > > > monitors are all assumed to be 96dpi. > > OS X makes this assumption. Windows makes this assumption by default, > but enables application of alternative assumptions. One particular > alternative, 120, is commonly applied prior to sale by vendors of > laptops having materially above average DPI. Other operating systems > vary whether and to what extent assumptions about display DPI are made, > particularly when display DPI is in fact greater than 96. Even on Windows with a non-96dpi, browsers today still assume 96dpi, as far as I can tell. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From mrmazda at ij.net Thu Jun 11 13:23:17 2009 From: mrmazda at ij.net (Felix Miata) Date: Thu, 11 Jun 2009 16:23:17 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <Pine.LNX.4.62.0906111958330.16244@hixie.dreamhostps.com> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> <4A31618A.8030809@ij.net> <Pine.LNX.4.62.0906111958330.16244@hixie.dreamhostps.com> Message-ID: <4A3167B5.8030406@ij.net> On 2009/06/11 19:58 (GMT) Ian Hickson composed: > On Thu, 11 Jun 2009, Felix Miata wrote: >> On 2009/06/11 19:17 (GMT) Ian Hickson composed: >> > monitors are all assumed to be 96dpi. >> OS X makes this assumption. Windows makes this assumption by default, >> but enables application of alternative assumptions. One particular >> alternative, 120, is commonly applied prior to sale by vendors of >> laptops having materially above average DPI. Other operating systems >> vary whether and to what extent assumptions about display DPI are made, >> particularly when display DPI is in fact greater than 96. > Even on Windows with a non-96dpi, browsers today still assume 96dpi, as > far as I can tell. On Windows, some assume 96 (e.g. Safari, which I reported as a Safari bug "UI text is too small for high PPI display environment" https://bugreport.apple.com/cgi-bin/WebObjects/RadarWeb.woa/61/wo/APqcRdz7DDWD1jtFE8yeWw/3.79.28.0.9 and related Webkit bug http://bugs.webkit.org/show_bug.cgi?id=18001 ), while for others 96 is only a floor (e.g. Gecko). IE uses whatever DPI has been specified for the desktop generally. -- "Cast but a glance at riches, and they are gone, for they will surely sprout wings and fly off to the sky like an eagle." Proverbs 23:5 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://fm.no-ip.com/ From sirokai at gmail.com Fri Jun 12 08:31:10 2009 From: sirokai at gmail.com (Christian Montoya) Date: Fri, 12 Jun 2009 11:31:10 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <4A31618A.8030809@ij.net> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> <4A31618A.8030809@ij.net> Message-ID: <1a19d7d10906120831x36f24a31o5d795cdfd8c98492@mail.gmail.com> On Thu, Jun 11, 2009 at 3:56 PM, Felix Miata<mrmazda at ij.net> wrote: > On 2009/06/11 19:17 (GMT) Ian Hickson composed: > >> monitors are all assumed to be 96dpi. > > OS X makes this assumption. Windows makes this assumption by default, but > enables application of alternative assumptions. One particular alternative, > 120, is commonly applied prior to sale by vendors of laptops having > materially above average DPI. Other operating systems vary whether and to > what extent assumptions about display DPI are made, particularly when display > DPI is in fact greater than 96. 8 years ago I bought a Dell widescreen laptop, Inspiron 6000, and it was pre-configured at 120 DPI. I've been telling people ever since, 96 DPI is not a standard. Dell does this for a lot of widescreen laptops, and that's just 1 vendor. -- -- Christian Montoya mappdev.com :: christianmontoya.net From chris at martinilab.com Thu Jun 18 10:53:07 2009 From: chris at martinilab.com (chris at martinilab.com) Date: Thu, 18 Jun 2009 12:53:07 -0500 (CDT) Subject: [html5] toDataURL calling itself causes security error Message-ID: <1245347587.134132250@192.168.1.71> Clever me thinking I could come up with a undo script. I ended up with an error I can't get past. var img = new Image(); img.src = canvas.toDataURL(); ctx.drawImage(img,0,0); This only works once though. The next time I use line 2. img.src = canvas.toDataURL(); I get a security error even though img is created from the same script on the server. The only way around this (and very ineffeciently I light add) is to post save the canvas to an image on the server. This is happening on FireFox 3, I haven't checked Safari 4 though. Is this working as it should? Is it a bug? Is there a better workaround than posting to the server? Thanks! Chris Williams www.martinilab.com From excors+whatwg at gmail.com Thu Jun 18 11:02:29 2009 From: excors+whatwg at gmail.com (Philip Taylor) Date: Thu, 18 Jun 2009 19:02:29 +0100 Subject: [html5] toDataURL calling itself causes security error In-Reply-To: <1245347587.134132250@192.168.1.71> References: <1245347587.134132250@192.168.1.71> Message-ID: <ea09c0d10906181102j6eee5633n7e9ca87ce4ed7050@mail.gmail.com> On Thu, Jun 18, 2009 at 6:53 PM, <chris at martinilab.com> wrote: > [...] > var img = new Image(); > img.src = canvas.toDataURL(); > ctx.drawImage(img,0,0); > > This only works once though. ?The next time I use line 2. > img.src = canvas.toDataURL(); > > I get a security error even though img is created from the same script on the server. > [...] > This is happening on FireFox 3, I haven't checked Safari 4 though. > > Is this working as it should? ?Is it a bug? It's a bug according to the security rules in HTML 5 (or at least it was when I last checked), and was broken in all implementations (when I last checked, ages ago). Sounds like https://bugzilla.mozilla.org/show_bug.cgi?id=417836 fixes it, presumably for Firefox 3.5, but I haven't tested that. >?Is there a better workaround than posting to the server? Do you have to use toDataURL? e.g. could you construct a new temporary canvas element, then draw from the current canvas onto it, then draw back, instead of using an Image? i.e. something like var tmp = document.createElement('canvas'); tmp.width = canvas.width; tmp.height = canvas.height; ctx.drawImage(tmp, 0, 0); tmp.getContext('2d').drawImage(canvas, 0, 0); -- Philip Taylor excors at gmail.com From chris at martinilab.com Thu Jun 18 11:34:45 2009 From: chris at martinilab.com (chris at martinilab.com) Date: Thu, 18 Jun 2009 13:34:45 -0500 (CDT) Subject: [html5] toDataURL calling itself causes security error In-Reply-To: <ea09c0d10906181102j6eee5633n7e9ca87ce4ed7050@mail.gmail.com> References: <1245347587.134132250@192.168.1.71> <ea09c0d10906181102j6eee5633n7e9ca87ce4ed7050@mail.gmail.com> Message-ID: <1245350085.461817920@192.168.1.70> On Thu, Jun 18, 2009 at 6:53 PM, <chris at martinilab.com> wrote: > It's a bug according to the security rules in HTML 5 (or at least it > was when I last checked), and was broken in all implementations (when > I last checked, ages ago). Sounds like > https://bugzilla.mozilla.org/show_bug.cgi?id=417836 fixes it, > presumably for Firefox 3.5, but I haven't tested that. Good to know! > Do you have to use toDataURL? e.g. could you construct a new temporary > canvas element, then draw from the current canvas onto it, then draw > back, instead of using an Image? i.e. something like > > var tmp = document.createElement('canvas'); > tmp.width = canvas.width; > tmp.height = canvas.height; > ctx.drawImage(tmp, 0, 0); > tmp.getContext('2d').drawImage(canvas, 0, 0); That works great actually. Thanks so much! Chris Williams http://www.martinilab.com From webmaster at keryx.se Sat Jun 27 04:22:39 2009 From: webmaster at keryx.se (Keryx Web) Date: Sat, 27 Jun 2009 13:22:39 +0200 Subject: [html5] Trying to help, please help me Message-ID: <4A4600FF.1090104@keryx.se> Hello all! I have updated my Best Practice Sheet at http://keryx.se/resources/html-elements/ It now has an HTML 5 doctype and some experimental CSS transformations (Only for the Fox 3.5 so far). All feedback about the content is appreciated. I have also written the following on a blog of mine: http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html Title: Validation and doctype myths and (inconvenient) truths It's an effort to educate. Please let me know if anything can be improved. -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From zcorpan at gmail.com Sat Jun 27 11:47:59 2009 From: zcorpan at gmail.com (Simon Pieters) Date: Sat, 27 Jun 2009 20:47:59 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A4600FF.1090104@keryx.se> References: <4A4600FF.1090104@keryx.se> Message-ID: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: > Hello all! > > I have updated my Best Practice Sheet at > > http://keryx.se/resources/html-elements/ <base> is now allowed in XHTML5 too. Maybe it would make sense to have one column for HTML5 and one for XHTML5? <basefont> is deprecated in HTML4. <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. <eventsource> was dropped. (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. The "?" label seems unused. I guess <html>, <listing>, <xmp> should be labeled "B". <noscript>, <ins>, <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, <embed>, <applet> are "I". <table> is marked "B" but is not display:block. I guess none of the frameset-related elements should have "B", since framesets can't be expressed with CSS. The <ruby>-related elements also have special values for 'display'. I'm not sure about <layer> and <nolayer> -- I think they're just unknown (and hence inline) in browsers these days. <figure> is an "N" element. <isindex> is a bit weird since the parser expands it into <form>...<input>. <bb> can be either normal or replaced, I think. <command> is "I". I think <details> is always "N". HTML 5 doesn't say that <meter> or <progress> are replaced elements. <img> and <object> can be either "R" or "N". <audio> can be either "R" or "I". The "I" for <source> is left-aligned while the others are centered. <map> is a normal inline element. I wouldn't say <frameset> is an "N" -- maybe leave the cell empty. Same for <frame>. I think <xml> is invisible in IE. > It now has an HTML 5 doctype and some experimental CSS transformations > (Only for the Fox 3.5 so far). Nice. Hixie: some tables in the spec could benefit from the same styling. > All feedback about the content is appreciated. > > I have also written the following on a blog of mine: > > http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html > > Title: Validation and doctype myths and (inconvenient) truths > > It's an effort to educate. Please let me know if anything can be > improved. > -- Simon Pieters From mgainty at hotmail.com Sat Jun 27 19:02:13 2009 From: mgainty at hotmail.com (Martin Gainty) Date: Sat, 27 Jun 2009 22:02:13 -0400 Subject: [html5] Trying to help, please help me In-Reply-To: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> Message-ID: <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> Goddag Simon for each HTML specificnation tag i see ? which converts to '<LF> using system font is this correct? this is good information laid out in a very intuitive matrix Sverige Altid! Martin ______________________________________________ Verzicht und Vertraulichkeitanmerkung/Note de d?ni et de confidentialit? Diese Nachricht ist vertraulich. Sollten Sie nicht der vorgesehene Empfaenger sein, so bitten wir hoeflich um eine Mitteilung. Jede unbefugte Weiterleitung oder Fertigung einer Kopie ist unzulaessig. Diese Nachricht dient lediglich dem Austausch von Informationen und entfaltet keine rechtliche Bindungswirkung. Aufgrund der leichten Manipulierbarkeit von E-Mails koennen wir keine Haftung fuer den Inhalt uebernehmen. Ce message est confidentiel et peut ?tre privil?gi?. Si vous n'?tes pas le destinataire pr?vu, nous te demandons avec bont? que pour satisfaire informez l'exp?diteur. N'importe quelle diffusion non autoris?e ou la copie de ceci est interdite. Ce message sert ? l'information seulement et n'aura pas n'importe quel effet l?galement obligatoire. ?tant donn? que les email peuvent facilement ?tre sujets ? la manipulation, nous ne pouvons accepter aucune responsabilit? pour le contenu fourni. > To: webmaster at keryx.se; help at whatwg.org > Date: Sat, 27 Jun 2009 20:47:59 +0200 > From: zcorpan at gmail.com > Subject: Re: [html5] Trying to help, please help me > > On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: > > > Hello all! > > > > I have updated my Best Practice Sheet at > > > > http://keryx.se/resources/html-elements/ > > <base> is now allowed in XHTML5 too. > > Maybe it would make sense to have one column for HTML5 and one for XHTML5? > > <basefont> is deprecated in HTML4. > > <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. > > <eventsource> was dropped. > > (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. > > The "?" label seems unused. > > I guess <html>, <listing>, <xmp> should be labeled "B". <noscript>, <ins>, > <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, <embed>, > <applet> are "I". <table> is marked "B" but is not display:block. I guess > none of the frameset-related elements should have "B", since framesets > can't be expressed with CSS. The <ruby>-related elements also have special > values for 'display'. I'm not sure about <layer> and <nolayer> -- I think > they're just unknown (and hence inline) in browsers these days. > > <figure> is an "N" element. <isindex> is a bit weird since the parser > expands it into <form>...<input>. <bb> can be either normal or replaced, I > think. <command> is "I". I think <details> is always "N". HTML 5 doesn't > say that <meter> or <progress> are replaced elements. <img> and <object> > can be either "R" or "N". <audio> can be either "R" or "I". The "I" for > <source> is left-aligned while the others are centered. <map> is a normal > inline element. I wouldn't say <frameset> is an "N" -- maybe leave the > cell empty. Same for <frame>. I think <xml> is invisible in IE. > > > > It now has an HTML 5 doctype and some experimental CSS transformations > > (Only for the Fox 3.5 so far). > > Nice. Hixie: some tables in the spec could benefit from the same styling. > > > > All feedback about the content is appreciated. > > > > I have also written the following on a blog of mine: > > > > http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html > > > > Title: Validation and doctype myths and (inconvenient) truths > > > > It's an effort to educate. Please let me know if anything can be > > improved. > > > > > -- > Simon Pieters > _______________________________________________ > Help mailing list > Help at lists.whatwg.org > http://lists.whatwg.org/listinfo.cgi/help-whatwg.org _________________________________________________________________ Windows Live? SkyDrive?: Get 25 GB of free online storage. http://windowslive.com/online/skydrive?ocid=TXT_TAGLM_WL_SD_25GB_062009 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.whatwg.org/pipermail/help-whatwg.org/attachments/20090627/1ed05aa7/attachment-0001.htm> From webmaster at keryx.se Sun Jun 28 04:29:27 2009 From: webmaster at keryx.se (Keryx Web) Date: Sun, 28 Jun 2009 13:29:27 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> Message-ID: <4A475417.5070408@keryx.se> On 2009-06-28 04:02, Martin Gainty wrote: > Goddag Simon > > for each HTML specificnation tag i see ? which converts to '<LF> using > system font > is this correct? Actually it is the entity &#10003; in the source code. Already when viewing source it converts to a check-mark i FFox, a bug I suppose. Opera gets it right in its view-source! If it converts to line-feed that sounds like another bug to me. In what Browser? -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From webmaster at keryx.se Sun Jun 28 10:45:59 2009 From: webmaster at keryx.se (Keryx Web) Date: Sun, 28 Jun 2009 19:45:59 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> Message-ID: <4A47AC57.8070308@keryx.se> On 2009-06-27 20:47, Simon Pieters wrote: > On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: >> http://keryx.se/resources/html-elements/ > > <base> is now allowed in XHTML5 too. Missed that. Thank you. > Maybe it would make sense to have one column for HTML5 and one for XHTML5? Maybe for my next *major* update. > <basefont> is deprecated in HTML4. Ooops! > <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. Thats why it is marked "s" = "sometimes". > <eventsource> was dropped. I thought it just moved to a separate spec? Note to self: "Must pay better attention!" > (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. OK. I just took for granted that it would include all ruby markup from XHTML 1.1. And that the rest just wasn't spec'd yet. Has this been discussed? Where? (I have not got time to follow every conversation, so sometimes I miss stuff. That's why I am asking for help ;-) ) > The "?" label seems unused. Yes. Since my last revision it is. > I guess <html>, <listing>, <xmp> should be labeled "B". Will fix. > <noscript>, > <ins>, <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, > <embed>, <applet> are "I". Will fix. > <table> is marked "B" but is not > display:block. I guess none of the frameset-related elements should have > "B", since framesets can't be expressed with CSS. Perhaps I should be more clear about what I mean with "blockish"... This table is goes back more than a few years and is the result of talking to my students. It is based on my personal quick way of explaining this to complete newbies: 1. Block elements will cause line breaks and must be floated or positioned in order to have contents on their sides. 2. Inline elements may never contain block elements. Both rules apply to table. Anyway, I think it is probably wise to be more precise. I will change it. Today I also teach CSS from day one, making it more clear what I mean anyway. > The <ruby>-related > elements also have special values for 'display'. I'm not sure about > <layer> and <nolayer> -- I think they're just unknown (and hence inline) > in browsers these days. I thought they were ignored completely...? At least they should be, IMHO. > <figure> is an "N" element. Ooops! > <isindex> is a bit weird since the parser > expands it into <form>...<input>. Weird indeed, but an input control will be shown. Sounds replaced to me! > <bb> can be either normal or replaced, > I think. Anyone who can conform this? > I think <details> is always "N". Anyone who can conform this? > <command> is "I". I thought some control would be shown. in my rightmost column I have said tha "Button, radiobutton or checkbox" will be shown. > HTML 5 > doesn't say that <meter> or <progress> are replaced elements. That was deduction from my part. How else would they display if not through some widget? > <img> and > <object> can be either "R" or "N". You mean, usually "R", but "N" when providing text-fallback? > <audio> can be either "R" or "I". The > "I" for <source> is left-aligned while the others are centered. Sounds like it's too complex to describe with just one letter... > <map> is > a normal inline element. As indeed it is in my table. > I wouldn't say <frameset> is an "N" -- maybe > leave the cell empty. Same for <frame>. Sounds reasonable. Will change. > I think <xml> is invisible in IE. A quick test proves this to be the case. >> It now has an HTML 5 doctype and some experimental CSS transformations >> (Only for the Fox 3.5 so far). > > Nice. Hixie: some tables in the spec could benefit from the same styling. > It is a very fragile technique that requires fixed widths for table columns and is almost impossible to get pixel perfect right. I think it is a very clear use case for something that should be very easy to accomplish in CSS, and I have sent an e-mail to the CSS WG. So far I have received no answer from anyone, though. -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From major at minet.sk Sun Jun 28 10:55:45 2009 From: major at minet.sk (Juraj Major) Date: Sun, 28 Jun 2009 19:55:45 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A4600FF.1090104@keryx.se> References: <4A4600FF.1090104@keryx.se> Message-ID: <4A47AEA1.3090508@minet.sk> Hello! I think this issue about CDATA > Opera may ignore CSS if MIME is true XHTML is not right. <style><![CDATA[ #foo { css rules } ]]></style> These css rules are ignored if MIME is text/html. It's all right with */*+xml. From webmaster at keryx.se Tue Jun 30 02:56:23 2009 From: webmaster at keryx.se (Keryx Web) Date: Tue, 30 Jun 2009 11:56:23 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A47AC57.8070308@keryx.se> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <4A47AC57.8070308@keryx.se> Message-ID: <4A49E147.40209@keryx.se> On 2009-06-28 19:45, Keryx Web wrote: > On 2009-06-27 20:47, Simon Pieters wrote: > ... I think I have addressed all of your concerns. An updated version is now up at: http://keryx.se/resources/html-elements/ (I tried to thank you on twitter, but could not find you...) -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From webmaster at keryx.se Tue Jun 30 03:17:00 2009 From: webmaster at keryx.se (Keryx Web) Date: Tue, 30 Jun 2009 12:17:00 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A49E147.40209@keryx.se> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <4A47AC57.8070308@keryx.se> <4A49E147.40209@keryx.se> Message-ID: <4A49E61C.2080008@keryx.se> On 2009-06-30 11:56, Keryx Web wrote: > I think I have addressed all of your concerns. An updated version is now > up at: http://keryx.se/resources/html-elements/ FYI, here is the changelog: More detailed explanations. Print stylesheet, version is shown. <base> allowed in XHTML 5 (spec change) <basefont> is deprecated in HTML 4.01 (fix) <ventsource> has been dropped *as an element* (spec change) <html>, <listing>, <xmp> are B (block elements) (fix) <a> Added "May contain block elements in (X)HTML 5." <noscript>, <ins>, <del> <object>, <noembed> and <canvas> are now marked I* (display: inline, but may contain block elements) (pedagogic change) <embed>, <applet>, <meter>, <progress> and <keygen> are now marked I (inline) (fix) <keygen> is now marked R (replaced) (fix) Ruby elements are not marked inline anymore (blank) (pedagogic change) Ruby elements <rbc>, <rtc> and <rb> marked as ? for HTML 5 (Not in spec, maybe never will be?) <isindex> Is more described as they are treated by an HTML 5 parser. (pedagogic change) <layer>, <nolayer>, <ilayer>, <xml>, <multicol>, <spacer> have no information any longer about inline/block normal/invisible/replaced, since current behavior differs from the original intention (pedagogic change) Expanded info about <isindex> (Block, text appears) (pedagogic change) <details> is N (normal), not I/N/R (pedagogic change) <thead> repeats on each page when printing (fix) Table is now T (not B) (pedagogic change) -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From ian at hixie.ch Wed Jun 10 00:05:08 2009 From: ian at hixie.ch (Ian Hickson) Date: Wed, 10 Jun 2009 07:05:08 +0000 (UTC) Subject: [html5] a question about canvas element's shadow effect In-Reply-To: <2E9A38A0754E4E22B85E659E028883ED@FNSTcyp> References: <2E9A38A0754E4E22B85E659E028883ED@FNSTcyp> Message-ID: <Pine.LNX.4.62.0906100701551.1648@hixie.dreamhostps.com> On Wed, 20 May 2009, caoyipeng wrote: > > canvas element's shadow effect use Gaussian blur to fulfill. > > In Gaussian blur, when do convolution operation, Gaussian template's size(or > Gaussian kernel width and height) affect the blur effect > > ,is there default value about Gaussian template's size(such as the size is > (6?+1)*(6?+1),? is the standard deviation). The bigger the better. This is just a quality of implementation issue. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From ian at hixie.ch Thu Jun 11 12:16:46 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:16:46 +0000 (UTC) Subject: [html5] Microdata and Meta In-Reply-To: <b6bb4d890905250914x41f3c2a1q12a5d58cd89a22ad@mail.gmail.com> References: <b6bb4d890905250914x41f3c2a1q12a5d58cd89a22ad@mail.gmail.com> Message-ID: <Pine.LNX.4.62.0906111915370.1648@hixie.dreamhostps.com> On Mon, 25 May 2009, Sean B. Palmer wrote: > > I know the Microdata stuff isn't stable yet, but I was just wondering > if the following is a valid and sensible thing to do: > > <!DOCTYPE html> > <head item> > <title>Microdata Example</title> > <link itemprop="about" href=""> > <meta itemprop="example" content="value"> It's valid, but I would urge people to just use normal <meta> elements instead, as in: <!DOCTYPE html> <head> <title>Microdata Example</title> <meta name="example" content="value"> -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From ian at hixie.ch Thu Jun 11 12:17:50 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:17:50 +0000 (UTC) Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> Message-ID: <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> On Mon, 25 May 2009, Sean B. Palmer wrote: > > HTML 5 currently says that img/@width and so forth use CSS 2.1 pixel > values. Pixel values in CSS 2.1 are of course a relative rather than an > absolute unit. In other words, CSS px is a non-linear alias for radians; > 675.522px, for example, is a quarter of a radian. > > But non-vector image formats like JPG and PNG have discrete pixels, and > screens have pixel displays with certain ppi values. So has anybody > looked at whether browsers actually implement px as a relative unit? And > how do they render images? I'm presuming that they don't resample to fit > the relative measurement. It's hard to say today because monitors are all assumed to be 96dpi. The idea of using CSS pixels everywhere is to make the spec be ready for when screens have higher resolutions. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From mrmazda at ij.net Thu Jun 11 12:56:58 2009 From: mrmazda at ij.net (Felix Miata) Date: Thu, 11 Jun 2009 15:56:58 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> Message-ID: <4A31618A.8030809@ij.net> On 2009/06/11 19:17 (GMT) Ian Hickson composed: > monitors are all assumed to be 96dpi. OS X makes this assumption. Windows makes this assumption by default, but enables application of alternative assumptions. One particular alternative, 120, is commonly applied prior to sale by vendors of laptops having materially above average DPI. Other operating systems vary whether and to what extent assumptions about display DPI are made, particularly when display DPI is in fact greater than 96. -- "Cast but a glance at riches, and they are gone, for they will surely sprout wings and fly off to the sky like an eagle." Proverbs 23:5 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://fm.no-ip.com/ From ian at hixie.ch Thu Jun 11 12:58:53 2009 From: ian at hixie.ch (Ian Hickson) Date: Thu, 11 Jun 2009 19:58:53 +0000 (UTC) Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <4A31618A.8030809@ij.net> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> <4A31618A.8030809@ij.net> Message-ID: <Pine.LNX.4.62.0906111958330.16244@hixie.dreamhostps.com> On Thu, 11 Jun 2009, Felix Miata wrote: > On 2009/06/11 19:17 (GMT) Ian Hickson composed: > > > monitors are all assumed to be 96dpi. > > OS X makes this assumption. Windows makes this assumption by default, > but enables application of alternative assumptions. One particular > alternative, 120, is commonly applied prior to sale by vendors of > laptops having materially above average DPI. Other operating systems > vary whether and to what extent assumptions about display DPI are made, > particularly when display DPI is in fact greater than 96. Even on Windows with a non-96dpi, browsers today still assume 96dpi, as far as I can tell. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.' From mrmazda at ij.net Thu Jun 11 13:23:17 2009 From: mrmazda at ij.net (Felix Miata) Date: Thu, 11 Jun 2009 16:23:17 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <Pine.LNX.4.62.0906111958330.16244@hixie.dreamhostps.com> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> <4A31618A.8030809@ij.net> <Pine.LNX.4.62.0906111958330.16244@hixie.dreamhostps.com> Message-ID: <4A3167B5.8030406@ij.net> On 2009/06/11 19:58 (GMT) Ian Hickson composed: > On Thu, 11 Jun 2009, Felix Miata wrote: >> On 2009/06/11 19:17 (GMT) Ian Hickson composed: >> > monitors are all assumed to be 96dpi. >> OS X makes this assumption. Windows makes this assumption by default, >> but enables application of alternative assumptions. One particular >> alternative, 120, is commonly applied prior to sale by vendors of >> laptops having materially above average DPI. Other operating systems >> vary whether and to what extent assumptions about display DPI are made, >> particularly when display DPI is in fact greater than 96. > Even on Windows with a non-96dpi, browsers today still assume 96dpi, as > far as I can tell. On Windows, some assume 96 (e.g. Safari, which I reported as a Safari bug "UI text is too small for high PPI display environment" https://bugreport.apple.com/cgi-bin/WebObjects/RadarWeb.woa/61/wo/APqcRdz7DDWD1jtFE8yeWw/3.79.28.0.9 and related Webkit bug http://bugs.webkit.org/show_bug.cgi?id=18001 ), while for others 96 is only a floor (e.g. Gecko). IE uses whatever DPI has been specified for the desktop generally. -- "Cast but a glance at riches, and they are gone, for they will surely sprout wings and fly off to the sky like an eagle." Proverbs 23:5 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://fm.no-ip.com/ From sirokai at gmail.com Fri Jun 12 08:31:10 2009 From: sirokai at gmail.com (Christian Montoya) Date: Fri, 12 Jun 2009 11:31:10 -0400 Subject: [html5] Absolute Pixels in HTML 5 In-Reply-To: <4A31618A.8030809@ij.net> References: <b6bb4d890905250924n27b1d457m4cda9418614b837b@mail.gmail.com> <Pine.LNX.4.62.0906111917010.1648@hixie.dreamhostps.com> <4A31618A.8030809@ij.net> Message-ID: <1a19d7d10906120831x36f24a31o5d795cdfd8c98492@mail.gmail.com> On Thu, Jun 11, 2009 at 3:56 PM, Felix Miata<mrmazda at ij.net> wrote: > On 2009/06/11 19:17 (GMT) Ian Hickson composed: > >> monitors are all assumed to be 96dpi. > > OS X makes this assumption. Windows makes this assumption by default, but > enables application of alternative assumptions. One particular alternative, > 120, is commonly applied prior to sale by vendors of laptops having > materially above average DPI. Other operating systems vary whether and to > what extent assumptions about display DPI are made, particularly when display > DPI is in fact greater than 96. 8 years ago I bought a Dell widescreen laptop, Inspiron 6000, and it was pre-configured at 120 DPI. I've been telling people ever since, 96 DPI is not a standard. Dell does this for a lot of widescreen laptops, and that's just 1 vendor. -- -- Christian Montoya mappdev.com :: christianmontoya.net From chris at martinilab.com Thu Jun 18 10:53:07 2009 From: chris at martinilab.com (chris at martinilab.com) Date: Thu, 18 Jun 2009 12:53:07 -0500 (CDT) Subject: [html5] toDataURL calling itself causes security error Message-ID: <1245347587.134132250@192.168.1.71> Clever me thinking I could come up with a undo script. I ended up with an error I can't get past. var img = new Image(); img.src = canvas.toDataURL(); ctx.drawImage(img,0,0); This only works once though. The next time I use line 2. img.src = canvas.toDataURL(); I get a security error even though img is created from the same script on the server. The only way around this (and very ineffeciently I light add) is to post save the canvas to an image on the server. This is happening on FireFox 3, I haven't checked Safari 4 though. Is this working as it should? Is it a bug? Is there a better workaround than posting to the server? Thanks! Chris Williams www.martinilab.com From excors+whatwg at gmail.com Thu Jun 18 11:02:29 2009 From: excors+whatwg at gmail.com (Philip Taylor) Date: Thu, 18 Jun 2009 19:02:29 +0100 Subject: [html5] toDataURL calling itself causes security error In-Reply-To: <1245347587.134132250@192.168.1.71> References: <1245347587.134132250@192.168.1.71> Message-ID: <ea09c0d10906181102j6eee5633n7e9ca87ce4ed7050@mail.gmail.com> On Thu, Jun 18, 2009 at 6:53 PM, <chris at martinilab.com> wrote: > [...] > var img = new Image(); > img.src = canvas.toDataURL(); > ctx.drawImage(img,0,0); > > This only works once though. ?The next time I use line 2. > img.src = canvas.toDataURL(); > > I get a security error even though img is created from the same script on the server. > [...] > This is happening on FireFox 3, I haven't checked Safari 4 though. > > Is this working as it should? ?Is it a bug? It's a bug according to the security rules in HTML 5 (or at least it was when I last checked), and was broken in all implementations (when I last checked, ages ago). Sounds like https://bugzilla.mozilla.org/show_bug.cgi?id=417836 fixes it, presumably for Firefox 3.5, but I haven't tested that. >?Is there a better workaround than posting to the server? Do you have to use toDataURL? e.g. could you construct a new temporary canvas element, then draw from the current canvas onto it, then draw back, instead of using an Image? i.e. something like var tmp = document.createElement('canvas'); tmp.width = canvas.width; tmp.height = canvas.height; ctx.drawImage(tmp, 0, 0); tmp.getContext('2d').drawImage(canvas, 0, 0); -- Philip Taylor excors at gmail.com From chris at martinilab.com Thu Jun 18 11:34:45 2009 From: chris at martinilab.com (chris at martinilab.com) Date: Thu, 18 Jun 2009 13:34:45 -0500 (CDT) Subject: [html5] toDataURL calling itself causes security error In-Reply-To: <ea09c0d10906181102j6eee5633n7e9ca87ce4ed7050@mail.gmail.com> References: <1245347587.134132250@192.168.1.71> <ea09c0d10906181102j6eee5633n7e9ca87ce4ed7050@mail.gmail.com> Message-ID: <1245350085.461817920@192.168.1.70> On Thu, Jun 18, 2009 at 6:53 PM, <chris at martinilab.com> wrote: > It's a bug according to the security rules in HTML 5 (or at least it > was when I last checked), and was broken in all implementations (when > I last checked, ages ago). Sounds like > https://bugzilla.mozilla.org/show_bug.cgi?id=417836 fixes it, > presumably for Firefox 3.5, but I haven't tested that. Good to know! > Do you have to use toDataURL? e.g. could you construct a new temporary > canvas element, then draw from the current canvas onto it, then draw > back, instead of using an Image? i.e. something like > > var tmp = document.createElement('canvas'); > tmp.width = canvas.width; > tmp.height = canvas.height; > ctx.drawImage(tmp, 0, 0); > tmp.getContext('2d').drawImage(canvas, 0, 0); That works great actually. Thanks so much! Chris Williams http://www.martinilab.com From webmaster at keryx.se Sat Jun 27 04:22:39 2009 From: webmaster at keryx.se (Keryx Web) Date: Sat, 27 Jun 2009 13:22:39 +0200 Subject: [html5] Trying to help, please help me Message-ID: <4A4600FF.1090104@keryx.se> Hello all! I have updated my Best Practice Sheet at http://keryx.se/resources/html-elements/ It now has an HTML 5 doctype and some experimental CSS transformations (Only for the Fox 3.5 so far). All feedback about the content is appreciated. I have also written the following on a blog of mine: http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html Title: Validation and doctype myths and (inconvenient) truths It's an effort to educate. Please let me know if anything can be improved. -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From zcorpan at gmail.com Sat Jun 27 11:47:59 2009 From: zcorpan at gmail.com (Simon Pieters) Date: Sat, 27 Jun 2009 20:47:59 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A4600FF.1090104@keryx.se> References: <4A4600FF.1090104@keryx.se> Message-ID: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: > Hello all! > > I have updated my Best Practice Sheet at > > http://keryx.se/resources/html-elements/ <base> is now allowed in XHTML5 too. Maybe it would make sense to have one column for HTML5 and one for XHTML5? <basefont> is deprecated in HTML4. <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. <eventsource> was dropped. (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. The "?" label seems unused. I guess <html>, <listing>, <xmp> should be labeled "B". <noscript>, <ins>, <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, <embed>, <applet> are "I". <table> is marked "B" but is not display:block. I guess none of the frameset-related elements should have "B", since framesets can't be expressed with CSS. The <ruby>-related elements also have special values for 'display'. I'm not sure about <layer> and <nolayer> -- I think they're just unknown (and hence inline) in browsers these days. <figure> is an "N" element. <isindex> is a bit weird since the parser expands it into <form>...<input>. <bb> can be either normal or replaced, I think. <command> is "I". I think <details> is always "N". HTML 5 doesn't say that <meter> or <progress> are replaced elements. <img> and <object> can be either "R" or "N". <audio> can be either "R" or "I". The "I" for <source> is left-aligned while the others are centered. <map> is a normal inline element. I wouldn't say <frameset> is an "N" -- maybe leave the cell empty. Same for <frame>. I think <xml> is invisible in IE. > It now has an HTML 5 doctype and some experimental CSS transformations > (Only for the Fox 3.5 so far). Nice. Hixie: some tables in the spec could benefit from the same styling. > All feedback about the content is appreciated. > > I have also written the following on a blog of mine: > > http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html > > Title: Validation and doctype myths and (inconvenient) truths > > It's an effort to educate. Please let me know if anything can be > improved. > -- Simon Pieters From mgainty at hotmail.com Sat Jun 27 19:02:13 2009 From: mgainty at hotmail.com (Martin Gainty) Date: Sat, 27 Jun 2009 22:02:13 -0400 Subject: [html5] Trying to help, please help me In-Reply-To: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> Message-ID: <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> Goddag Simon for each HTML specificnation tag i see ? which converts to '<LF> using system font is this correct? this is good information laid out in a very intuitive matrix Sverige Altid! Martin ______________________________________________ Verzicht und Vertraulichkeitanmerkung/Note de d?ni et de confidentialit? Diese Nachricht ist vertraulich. Sollten Sie nicht der vorgesehene Empfaenger sein, so bitten wir hoeflich um eine Mitteilung. Jede unbefugte Weiterleitung oder Fertigung einer Kopie ist unzulaessig. Diese Nachricht dient lediglich dem Austausch von Informationen und entfaltet keine rechtliche Bindungswirkung. Aufgrund der leichten Manipulierbarkeit von E-Mails koennen wir keine Haftung fuer den Inhalt uebernehmen. Ce message est confidentiel et peut ?tre privil?gi?. Si vous n'?tes pas le destinataire pr?vu, nous te demandons avec bont? que pour satisfaire informez l'exp?diteur. N'importe quelle diffusion non autoris?e ou la copie de ceci est interdite. Ce message sert ? l'information seulement et n'aura pas n'importe quel effet l?galement obligatoire. ?tant donn? que les email peuvent facilement ?tre sujets ? la manipulation, nous ne pouvons accepter aucune responsabilit? pour le contenu fourni. > To: webmaster at keryx.se; help at whatwg.org > Date: Sat, 27 Jun 2009 20:47:59 +0200 > From: zcorpan at gmail.com > Subject: Re: [html5] Trying to help, please help me > > On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: > > > Hello all! > > > > I have updated my Best Practice Sheet at > > > > http://keryx.se/resources/html-elements/ > > <base> is now allowed in XHTML5 too. > > Maybe it would make sense to have one column for HTML5 and one for XHTML5? > > <basefont> is deprecated in HTML4. > > <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. > > <eventsource> was dropped. > > (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. > > The "?" label seems unused. > > I guess <html>, <listing>, <xmp> should be labeled "B". <noscript>, <ins>, > <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, <embed>, > <applet> are "I". <table> is marked "B" but is not display:block. I guess > none of the frameset-related elements should have "B", since framesets > can't be expressed with CSS. The <ruby>-related elements also have special > values for 'display'. I'm not sure about <layer> and <nolayer> -- I think > they're just unknown (and hence inline) in browsers these days. > > <figure> is an "N" element. <isindex> is a bit weird since the parser > expands it into <form>...<input>. <bb> can be either normal or replaced, I > think. <command> is "I". I think <details> is always "N". HTML 5 doesn't > say that <meter> or <progress> are replaced elements. <img> and <object> > can be either "R" or "N". <audio> can be either "R" or "I". The "I" for > <source> is left-aligned while the others are centered. <map> is a normal > inline element. I wouldn't say <frameset> is an "N" -- maybe leave the > cell empty. Same for <frame>. I think <xml> is invisible in IE. > > > > It now has an HTML 5 doctype and some experimental CSS transformations > > (Only for the Fox 3.5 so far). > > Nice. Hixie: some tables in the spec could benefit from the same styling. > > > > All feedback about the content is appreciated. > > > > I have also written the following on a blog of mine: > > > > http://itpastorn.blogspot.com/2009/06/validation-and-doctype-myths-and.html > > > > Title: Validation and doctype myths and (inconvenient) truths > > > > It's an effort to educate. Please let me know if anything can be > > improved. > > > > > -- > Simon Pieters > _______________________________________________ > Help mailing list > Help at lists.whatwg.org > http://lists.whatwg.org/listinfo.cgi/help-whatwg.org _________________________________________________________________ Windows Live? SkyDrive?: Get 25 GB of free online storage. http://windowslive.com/online/skydrive?ocid=TXT_TAGLM_WL_SD_25GB_062009 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.whatwg.org/pipermail/help-whatwg.org/attachments/20090627/1ed05aa7/attachment-0002.htm> From webmaster at keryx.se Sun Jun 28 04:29:27 2009 From: webmaster at keryx.se (Keryx Web) Date: Sun, 28 Jun 2009 13:29:27 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <BLU142-W18C4247ABCCD53CDA861BBAE330@phx.gbl> Message-ID: <4A475417.5070408@keryx.se> On 2009-06-28 04:02, Martin Gainty wrote: > Goddag Simon > > for each HTML specificnation tag i see ? which converts to '<LF> using > system font > is this correct? Actually it is the entity &#10003; in the source code. Already when viewing source it converts to a check-mark i FFox, a bug I suppose. Opera gets it right in its view-source! If it converts to line-feed that sounds like another bug to me. In what Browser? -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From webmaster at keryx.se Sun Jun 28 10:45:59 2009 From: webmaster at keryx.se (Keryx Web) Date: Sun, 28 Jun 2009 19:45:59 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <op.uv6458c87a8kvn@simon-pieterss-macbook.local> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> Message-ID: <4A47AC57.8070308@keryx.se> On 2009-06-27 20:47, Simon Pieters wrote: > On Sat, 27 Jun 2009 13:22:39 +0200, Keryx Web <webmaster at keryx.se> wrote: >> http://keryx.se/resources/html-elements/ > > <base> is now allowed in XHTML5 too. Missed that. Thank you. > Maybe it would make sense to have one column for HTML5 and one for XHTML5? Maybe for my next *major* update. > <basefont> is deprecated in HTML4. Ooops! > <![CDATA[ ]]> is allowed in HTML5 but only in SVG and MathML contexts. Thats why it is marked "s" = "sometimes". > <eventsource> was dropped. I thought it just moved to a separate spec? Note to self: "Must pay better attention!" > (X)HTML5 doesn't have <rbc>, <rtc>, <rb>. OK. I just took for granted that it would include all ruby markup from XHTML 1.1. And that the rest just wasn't spec'd yet. Has this been discussed? Where? (I have not got time to follow every conversation, so sometimes I miss stuff. That's why I am asking for help ;-) ) > The "?" label seems unused. Yes. Since my last revision it is. > I guess <html>, <listing>, <xmp> should be labeled "B". Will fix. > <noscript>, > <ins>, <del>, <meter>, <progress>, <keygen>, <object>, <canvas>, > <embed>, <applet> are "I". Will fix. > <table> is marked "B" but is not > display:block. I guess none of the frameset-related elements should have > "B", since framesets can't be expressed with CSS. Perhaps I should be more clear about what I mean with "blockish"... This table is goes back more than a few years and is the result of talking to my students. It is based on my personal quick way of explaining this to complete newbies: 1. Block elements will cause line breaks and must be floated or positioned in order to have contents on their sides. 2. Inline elements may never contain block elements. Both rules apply to table. Anyway, I think it is probably wise to be more precise. I will change it. Today I also teach CSS from day one, making it more clear what I mean anyway. > The <ruby>-related > elements also have special values for 'display'. I'm not sure about > <layer> and <nolayer> -- I think they're just unknown (and hence inline) > in browsers these days. I thought they were ignored completely...? At least they should be, IMHO. > <figure> is an "N" element. Ooops! > <isindex> is a bit weird since the parser > expands it into <form>...<input>. Weird indeed, but an input control will be shown. Sounds replaced to me! > <bb> can be either normal or replaced, > I think. Anyone who can conform this? > I think <details> is always "N". Anyone who can conform this? > <command> is "I". I thought some control would be shown. in my rightmost column I have said tha "Button, radiobutton or checkbox" will be shown. > HTML 5 > doesn't say that <meter> or <progress> are replaced elements. That was deduction from my part. How else would they display if not through some widget? > <img> and > <object> can be either "R" or "N". You mean, usually "R", but "N" when providing text-fallback? > <audio> can be either "R" or "I". The > "I" for <source> is left-aligned while the others are centered. Sounds like it's too complex to describe with just one letter... > <map> is > a normal inline element. As indeed it is in my table. > I wouldn't say <frameset> is an "N" -- maybe > leave the cell empty. Same for <frame>. Sounds reasonable. Will change. > I think <xml> is invisible in IE. A quick test proves this to be the case. >> It now has an HTML 5 doctype and some experimental CSS transformations >> (Only for the Fox 3.5 so far). > > Nice. Hixie: some tables in the spec could benefit from the same styling. > It is a very fragile technique that requires fixed widths for table columns and is almost impossible to get pixel perfect right. I think it is a very clear use case for something that should be very easy to accomplish in CSS, and I have sent an e-mail to the CSS WG. So far I have received no answer from anyone, though. -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From major at minet.sk Sun Jun 28 10:55:45 2009 From: major at minet.sk (Juraj Major) Date: Sun, 28 Jun 2009 19:55:45 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A4600FF.1090104@keryx.se> References: <4A4600FF.1090104@keryx.se> Message-ID: <4A47AEA1.3090508@minet.sk> Hello! I think this issue about CDATA > Opera may ignore CSS if MIME is true XHTML is not right. <style><![CDATA[ #foo { css rules } ]]></style> These css rules are ignored if MIME is text/html. It's all right with */*+xml. From webmaster at keryx.se Tue Jun 30 02:56:23 2009 From: webmaster at keryx.se (Keryx Web) Date: Tue, 30 Jun 2009 11:56:23 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A47AC57.8070308@keryx.se> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <4A47AC57.8070308@keryx.se> Message-ID: <4A49E147.40209@keryx.se> On 2009-06-28 19:45, Keryx Web wrote: > On 2009-06-27 20:47, Simon Pieters wrote: > ... I think I have addressed all of your concerns. An updated version is now up at: http://keryx.se/resources/html-elements/ (I tried to thank you on twitter, but could not find you...) -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/ From webmaster at keryx.se Tue Jun 30 03:17:00 2009 From: webmaster at keryx.se (Keryx Web) Date: Tue, 30 Jun 2009 12:17:00 +0200 Subject: [html5] Trying to help, please help me In-Reply-To: <4A49E147.40209@keryx.se> References: <4A4600FF.1090104@keryx.se> <op.uv6458c87a8kvn@simon-pieterss-macbook.local> <4A47AC57.8070308@keryx.se> <4A49E147.40209@keryx.se> Message-ID: <4A49E61C.2080008@keryx.se> On 2009-06-30 11:56, Keryx Web wrote: > I think I have addressed all of your concerns. An updated version is now > up at: http://keryx.se/resources/html-elements/ FYI, here is the changelog: More detailed explanations. Print stylesheet, version is shown. <base> allowed in XHTML 5 (spec change) <basefont> is deprecated in HTML 4.01 (fix) <ventsource> has been dropped *as an element* (spec change) <html>, <listing>, <xmp> are B (block elements) (fix) <a> Added "May contain block elements in (X)HTML 5." <noscript>, <ins>, <del> <object>, <noembed> and <canvas> are now marked I* (display: inline, but may contain block elements) (pedagogic change) <embed>, <applet>, <meter>, <progress> and <keygen> are now marked I (inline) (fix) <keygen> is now marked R (replaced) (fix) Ruby elements are not marked inline anymore (blank) (pedagogic change) Ruby elements <rbc>, <rtc> and <rb> marked as ? for HTML 5 (Not in spec, maybe never will be?) <isindex> Is more described as they are treated by an HTML 5 parser. (pedagogic change) <layer>, <nolayer>, <ilayer>, <xml>, <multicol>, <spacer> have no information any longer about inline/block normal/invisible/replaced, since current behavior differs from the original intention (pedagogic change) Expanded info about <isindex> (Block, text appears) (pedagogic change) <details> is N (normal), not I/N/R (pedagogic change) <thead> repeats on each page when printing (fix) Table is now T (not B) (pedagogic change) -- Keryx Web (Lars Gunther) http://keryx.se/ http://twitter.com/itpastorn/ http://itpastorn.blogspot.com/