From mark at novemberborn.net Fri Dec 16 19:30:55 2005 From: mark at novemberborn.net (Mark Wubben) Date: Fri, 16 Dec 2005 20:30:55 +0100 Subject: [sIFR3-dev] Hello World Message-ID: Just testing :) -- Mark Wubben http://novemberborn.net/ From mark at novemberborn.net Sun Dec 18 23:30:26 2005 From: mark at novemberborn.net (Mark Wubben) Date: Mon, 19 Dec 2005 00:30:26 +0100 Subject: [sIFR3-dev] About tuning Message-ID: <8496076A-1CA9-4AD3-88AB-A87F5651A68D@novemberborn.net> Hey! The first mail to the list, here we go! I've been looking at the way to best do font tuning in sIFR 3. See for example (should work in modern browsers) and (should work in older browsers such as Safari 1.2 and Opera 7.x). What I found is that it's possible to calculate the line- height of the to-be-replaced-element and the number of lines the text is using. With this information, and the width of the element, it'll be possible to render the font with the same pixel size as the line- height as specified in the browser CSS (as long as font-size and line- height are specified and equal). This is different from sIFR 2 which used the vertical dimension of the element minus padding, but had no clue how many lines the text occupied. This meant that the actual rendering would vary between pages or even elements. While I haven't looked at the ActionScript code for determining the font size yet, I'm sure this is a huge part of why sIFR wouldn't always be predictable. The results of both tests should be equal: equal lines, equal line- height. I've tested it in Firefox 1.5, Safari 2.02 and Opera 8.5 on OS X with different font sizes (and yes, it scales!). Could somebody test it in IE and older versions of the browsers mentioned above? Kind regards, and thanks for subscribing, -- Mark Wubben http://novemberborn.net/ From branstrom at gmail.com Tue Dec 20 01:35:19 2005 From: branstrom at gmail.com (=?ISO-8859-1?Q?Fredrik_Br=E4nstr=F6m?=) Date: Tue, 20 Dec 2005 02:35:19 +0100 Subject: [sIFR3-dev] About tuning In-Reply-To: <8496076A-1CA9-4AD3-88AB-A87F5651A68D@novemberborn.net> References: <8496076A-1CA9-4AD3-88AB-A87F5651A68D@novemberborn.net> Message-ID: <22410ce30512191735u31689d90l5799a78d4ab33e2a@mail.gmail.com> The links get wrapped and broken. :/ Confirmed working in Opera 8.51 and 9.0 Preview on Windows. And Firefox of course. IE6 Windows has some fun with line-height.html: 1. - *#one* - vertical padding: 0px and 0px - vertical border: 1px and 1px - line-height: 12px - offsetHeight: 20px - 1 line(s) - working height: 18px 2. - *#two* - vertical padding: 0px and 0px - vertical border: 1px and 1px - line-height: 12px - offsetHeight: 36px - 2 line(s) - working height: 34px 3. - *#three* - vertical padding: 1px and 1px - vertical border: 1px and 1px - line-height: 12px - offsetHeight: 68px - 5 line(s) - working height: 64px 4. - *#four* - vertical padding: 2px and 2px - vertical border: 1px and 1px - line-height: 12px - offsetHeight: 100px - 7 line(s) - working height: 94px 5. - *#five* - vertical padding: 0px and 0px - vertical border: 1px and 1px - line-height: 12px - offsetHeight: 52px - 4 line(s) - working height: 50px deducing-line-height.html on IE6 Win: 1. - *#one* - vertical padding: 0px and 0px - vertical border: 1px and 1px - line-height: 18px - offsetHeight: 20px - 1 line(s) - working height: 18px 2. - *#two* - vertical padding: 0px and 0px - vertical border: 1px and 1px - line-height: 18px - offsetHeight: 36px - 1 line(s) - working height: 34px 3. - *#three* - vertical padding: 16px and 16px - vertical border: 1px and 1px - line-height: 18px - offsetHeight: 68px - 1 line(s) - working height: 34px 4. - *#four* - vertical padding: 32px and 32px - vertical border: 1px and 1px - line-height: 18px - offsetHeight: 100px - 1 line(s) - working height: 34px 5. - *#five* - vertical padding: 0px and 0px - vertical border: 1px and 1px - line-height: 18px - offsetHeight: 52px - 2 line(s) - working height: 50px Seems to work, right? It's a bit higher (saw 16, 32 and 48 in other browsers), but that's because they interpret the ems differently. Am I right? On 12/19/05, Mark Wubben < mark at novemberborn.net> wrote: > > Hey! The first mail to the list, here we go! > > I've been looking at the way to best do font tuning in sIFR 3. See > for example < http://tests.novemberborn.net/sifr3/experiments/tuning/ > line-height.html> (should work in modern browsers) and tests.novemberborn.net/sifr3/experiments/tuning/deducing-line- > height.html> (should work in older browsers such as Safari 1.2 and > Opera 7.x). What I found is that it's possible to calculate the line- > height of the to-be-replaced-element and the number of lines the text > is using. With this information, and the width of the element, it'll > be possible to render the font with the same pixel size as the line- > height as specified in the browser CSS (as long as font-size and line- > height are specified and equal). > > This is different from sIFR 2 which used the vertical dimension of > the element minus padding, but had no clue how many lines the text > occupied. This meant that the actual rendering would vary between > pages or even elements. While I haven't looked at the ActionScript > code for determining the font size yet, I'm sure this is a huge part > of why sIFR wouldn't always be predictable. > > The results of both tests should be equal: equal lines, equal line- > height. I've tested it in Firefox 1.5, Safari 2.02 and Opera 8.5 on > OS X with different font sizes (and yes, it scales!). Could somebody > test it in IE and older versions of the browsers mentioned above? > > Kind regards, and thanks for subscribing, > > -- > Mark Wubben > http://novemberborn.net/ > > > > _______________________________________________ > sifr3-dev mailing list > sifr3-dev at lists.novemberborn.net > http://lists.novemberborn.net/mailman/listinfo/sifr3-dev > -- Fredrik Br?nstr?m branstrom at gmail.com fredrik at brnm.se http://fredrik.brnm.se -------------- next part -------------- An HTML attachment was scrubbed... URL: From mark at novemberborn.net Tue Dec 20 01:46:41 2005 From: mark at novemberborn.net (Mark Wubben) Date: Tue, 20 Dec 2005 02:46:41 +0100 Subject: [sIFR3-dev] About tuning In-Reply-To: <22410ce30512191735u31689d90l5799a78d4ab33e2a@mail.gmail.com> References: <8496076A-1CA9-4AD3-88AB-A87F5651A68D@novemberborn.net> <22410ce30512191735u31689d90l5799a78d4ab33e2a@mail.gmail.com> Message-ID: <169F3AC8-AD01-4610-BF47-50FCB959A776@novemberborn.net> On Dec 20, 2005, at 2:35 AM, Fredrik Br?nstr?m wrote: > IE6 Windows has some fun with line-height.html: Those don't look right at all. Will have to look into that... > > Seems to work, right? It's a bit higher (saw 16, 32 and 48 in other > browsers), but that's because they [IE] interpret the ems > differently. Am I right? Not sure. I hope they interpret them the same! -- Mark Wubben http://novemberborn.net/ From branstrom at gmail.com Tue Dec 20 06:43:25 2005 From: branstrom at gmail.com (=?ISO-8859-1?Q?Fredrik_Br=E4nstr=F6m?=) Date: Tue, 20 Dec 2005 07:43:25 +0100 Subject: [sIFR3-dev] About tuning In-Reply-To: <169F3AC8-AD01-4610-BF47-50FCB959A776@novemberborn.net> References: <8496076A-1CA9-4AD3-88AB-A87F5651A68D@novemberborn.net> <22410ce30512191735u31689d90l5799a78d4ab33e2a@mail.gmail.com> <169F3AC8-AD01-4610-BF47-50FCB959A776@novemberborn.net> Message-ID: <22410ce30512192243x71e376e2tc8ccada715ad57b3@mail.gmail.com> I have a distinct memory that they don't, IE usually renders text a bit bigger or smaller than Firefox and Opera, don't remember There are tiny differences between even Firefox and Opera, IIRC. I'll test it tomorrow or whenever I get the time. IE6 is a PITA... You should have a PC for testing, you're the kung-fu scripter and direct access to the rendering engine that's the most likely to cause a fuss at any given time is pretty crucial if we want sIFR to behave on 85% of Windows machines. On 12/20/05, Mark Wubben wrote: > > On Dec 20, 2005, at 2:35 AM, Fredrik Br?nstr?m wrote: > > > IE6 Windows has some fun with line-height.html: > > Those don't look right at all. Will have to look into that... > > > > Seems to work, right? It's a bit higher (saw 16, 32 and 48 in other > > browsers), but that's because they [IE] interpret the ems > > differently. Am I right? > > Not sure. I hope they interpret them the same! > > -- > Mark Wubben > http://novemberborn.net/ > > > > -- Fredrik Br?nstr?m branstrom at gmail.com fredrik at brnm.se http://fredrik.brnm.se -------------- next part -------------- An HTML attachment was scrubbed... URL: From marcos.peebles at gmail.com Wed Dec 21 12:38:32 2005 From: marcos.peebles at gmail.com (Marcos Peebles) Date: Wed, 21 Dec 2005 13:38:32 +0100 Subject: [sIFR3-dev] works ok with IE 6 on XP Message-ID: Seems like IE 6.0.2900 works fine, tested with XP looks the same as FF 1.5 (exactly the same) Cheers On Dec 20, 2005, at 2:35 AM, Fredrik Br?nstr?m wrote: > I have a distinct memory that they don't, IE usually renders text a bit > bigger or smaller than Firefox and Opera, don't remember There are tiny > differences between even Firefox and Opera, IIRC. I'll test it tomorrow or > whenever I get the time. > IE6 is a PITA... You should have a PC for testing, you're the kung-fu > scripter and direct access to the rendering engine that's the most likely to > cause a fuss at any given time is pretty crucial if we want sIFR to behave > on 85% of Windows machines. On 12/20/05, Mark Wubben wrote: > > On Dec 20, 2005, at 2:35 AM, Fredrik Br?nstr?m wrote: > > > IE6 Windows has some fun with line-height.html: > > Those don't look right at all. Will have to look into that... > > > > Seems to work, right? It's a bit higher (saw 16, 32 and 48 in other > > browsers), but that's because they [IE] interpret the ems > > differently. Am I right? > > Not sure. I hope they interpret them the same! > > -- > Mark Wubben > http://novemberborn.net/ From mark at novemberborn.net Wed Dec 21 21:25:01 2005 From: mark at novemberborn.net (Mark Wubben) Date: Wed, 21 Dec 2005 22:25:01 +0100 Subject: [sIFR3-dev] About tuning, part 2 Message-ID: Hi, This is another mail about the tuning, the first thread can be found in the archive [1]. If you go to you'll find three test cases. These show how to calculate the line- height and number of lines text is taking in Mozilla, Safari, Opera and Internet Explorer. IE requires quite a different approach, and this can be found in "deducing-line-height-ie.html" (the first link in the directory listening, why they names are truncated I don't know). Please have a look at the tests and let me know how well they match up between browsers. Thanks! [1]: http://barclay.textdrive.com/pipermail/sifr3-dev/2005-December/ 000001.html -- Mark Wubben http://novemberborn.net/ From marcos.peebles at gmail.com Thu Dec 22 14:02:43 2005 From: marcos.peebles at gmail.com (Marcos Peebles) Date: Thu, 22 Dec 2005 15:02:43 +0100 Subject: [sIFR3-dev] About tuning, part 2 Message-ID: Seems all well, they look very much alike I posted 9 screenshots here: http://www.piezoworks.be/gallery/thumbnails.php?album=2 Cover FF 1.5, IE 6.00 and Opera 8.51 all in WP SP2 Cheers From branstrom at gmail.com Fri Dec 23 00:21:35 2005 From: branstrom at gmail.com (=?ISO-8859-1?Q?Fredrik_Br=E4nstr=F6m?=) Date: Fri, 23 Dec 2005 01:21:35 +0100 Subject: [sIFR3-dev] About tuning, part 2 In-Reply-To: References: Message-ID: <22410ce30512221621g33efe0f8m1e780f7f9d0a7e5b@mail.gmail.com> They look the same save for a few pixels, yes, but that's not what we're after here. We want the numbers below to be the same in every browser. (Sorry for the duplicate mail Marcos, forgot to Cc the list.) On 12/22/05, Marcos Peebles wrote: > > Seems all well, they look very much alike > > I posted 9 screenshots here: > http://www.piezoworks.be/gallery/thumbnails.php?album=2 > > Cover FF 1.5, IE 6.00 and Opera 8.51 all in WP SP2 > > Cheers > _______________________________________________ > sifr3-dev mailing list > sifr3-dev at lists.novemberborn.net > http://lists.novemberborn.net/mailman/listinfo/sifr3-dev > -- Fredrik Br?nstr?m branstrom at gmail.com fredrik at brnm.se http://fredrik.brnm.se -------------- next part -------------- An HTML attachment was scrubbed... URL: From mark at novemberborn.net Fri Dec 23 01:00:48 2005 From: mark at novemberborn.net (Mark Wubben) Date: Fri, 23 Dec 2005 02:00:48 +0100 Subject: [sIFR3-dev] About tuning, part 2 In-Reply-To: <22410ce30512221621g33efe0f8m1e780f7f9d0a7e5b@mail.gmail.com> References: <22410ce30512221621g33efe0f8m1e780f7f9d0a7e5b@mail.gmail.com> Message-ID: <7CD6FF2B-BFD6-40FC-B274-C006FABD441A@novemberborn.net> On Dec 23, 2005, at 1:21 AM, Fredrik Br?nstr?m wrote: > They look the same save for a few pixels, yes, but that's not what > we're after here. We want the numbers below to be the same in every > browser. Heh yeah, already mailed Marcos about that :) It seems that the default 1em font-size in IE is two pixels smaller than that in MOSe (at least on OS X), but that shouldn't be much of a problem. If someone could test this in an old Konqueror or Safari 1.0 I think we're golden. -- Mark Wubben http://novemberborn.net/ From branstrom at gmail.com Fri Dec 23 01:17:58 2005 From: branstrom at gmail.com (=?ISO-8859-1?Q?Fredrik_Br=E4nstr=F6m?=) Date: Fri, 23 Dec 2005 02:17:58 +0100 Subject: [sIFR3-dev] About tuning, part 2 In-Reply-To: References: Message-ID: <22410ce30512221717h7542e897ka17acf6c1ba40368@mail.gmail.com> IE6-Win shows: http://filer.brnm.se/i/Shot0034.png On 12/21/05, Mark Wubben wrote: > > Hi, > > This is another mail about the tuning, the first thread can be found > in the archive [1]. > > If you go to > you'll find three test cases. These show how to calculate the line- > height and number of lines text is taking in Mozilla, Safari, Opera > and Internet Explorer. IE requires quite a different approach, and > this can be found in "deducing-line-height-ie.html" (the first link > in the directory listening, why they names are truncated I don't know). > > Please have a look at the tests and let me know how well they match > up between browsers. > > Thanks! > > [1]: http://barclay.textdrive.com/pipermail/sifr3-dev/2005-December/ > 000001.html > > -- > Mark Wubben > http://novemberborn.net/ > > > > _______________________________________________ > sifr3-dev mailing list > sifr3-dev at lists.novemberborn.net > http://lists.novemberborn.net/mailman/listinfo/sifr3-dev > -- Fredrik Br?nstr?m branstrom at gmail.com fredrik at brnm.se http://fredrik.brnm.se -------------- next part -------------- An HTML attachment was scrubbed... URL: From branstrom at gmail.com Fri Dec 23 01:33:40 2005 From: branstrom at gmail.com (=?ISO-8859-1?Q?Fredrik_Br=E4nstr=F6m?=) Date: Fri, 23 Dec 2005 02:33:40 +0100 Subject: [sIFR3-dev] About tuning, part 2 In-Reply-To: <7CD6FF2B-BFD6-40FC-B274-C006FABD441A@novemberborn.net> References: <22410ce30512221621g33efe0f8m1e780f7f9d0a7e5b@mail.gmail.com> <7CD6FF2B-BFD6-40FC-B274-C006FABD441A@novemberborn.net> Message-ID: <22410ce30512221733g269ac2c2na3f4589e2e7c7b32@mail.gmail.com> You mean bigger? Because it's two pixels bigger than Firefox's and Opera's on Windows; http://filer.brnm.se/i/Shot0035.png http://filer.brnm.se/i/Shot0036.png On 12/23/05, Mark Wubben wrote: > > It seems that the default 1em font-size in IE is two pixels smaller > than that in MOSe (at least on OS X), but that shouldn't be much of a > problem. If someone could test this in an old Konqueror or Safari 1.0 > I think we're golden. > > -- > Mark Wubben > http://novemberborn.net/ > > > > -- Fredrik Br?nstr?m branstrom at gmail.com fredrik at brnm.se http://fredrik.brnm.se -------------- next part -------------- An HTML attachment was scrubbed... URL: From marcos.peebles at gmail.com Fri Dec 23 04:07:22 2005 From: marcos.peebles at gmail.com (Marcos Peebles) Date: Fri, 23 Dec 2005 05:07:22 +0100 Subject: [sIFR3-dev] About tuning, part 2 In-Reply-To: <7CD6FF2B-BFD6-40FC-B274-C006FABD441A@novemberborn.net> References: <22410ce30512221621g33efe0f8m1e780f7f9d0a7e5b@mail.gmail.com> <7CD6FF2B-BFD6-40FC-B274-C006FABD441A@novemberborn.net> Message-ID: I'll check with Safari (@ my office) and let you know... Sorry bout my intitial confusion ... Cheers On 23/12/05, Mark Wubben wrote: > On Dec 23, 2005, at 1:21 AM, Fredrik Br?nstr?m wrote: > > > They look the same save for a few pixels, yes, but that's not what > > we're after here. We want the numbers below to be the same in every > > browser. > > Heh yeah, already mailed Marcos about that :) > > It seems that the default 1em font-size in IE is two pixels smaller > than that in MOSe (at least on OS X), but that shouldn't be much of a > problem. If someone could test this in an old Konqueror or Safari 1.0 > I think we're golden. > > -- > Mark Wubben > http://novemberborn.net/ > > > > From branstrom at gmail.com Fri Dec 23 05:13:38 2005 From: branstrom at gmail.com (=?ISO-8859-1?Q?Fredrik_Br=E4nstr=F6m?=) Date: Fri, 23 Dec 2005 06:13:38 +0100 Subject: [sIFR3-dev] About tuning, part 2 In-Reply-To: References: <22410ce30512221621g33efe0f8m1e780f7f9d0a7e5b@mail.gmail.com> <7CD6FF2B-BFD6-40FC-B274-C006FABD441A@novemberborn.net> Message-ID: <22410ce30512222113o2975082eg457dc155e83fb12e@mail.gmail.com> Would be great! Set up a VNC server so you can access that computer from home, too, if you have the administrative rights and the opportunity. That's OK, I was a little confused at first too ^^ On 12/23/05, Marcos Peebles wrote: > > I'll check with Safari (@ my office) and let you know... > > Sorry bout my intitial confusion ... > Cheers > > On 23/12/05, Mark Wubben wrote: > > On Dec 23, 2005, at 1:21 AM, Fredrik Br?nstr?m wrote: > > > > > They look the same save for a few pixels, yes, but that's not what > > > we're after here. We want the numbers below to be the same in every > > > browser. > > > > Heh yeah, already mailed Marcos about that :) > > > > It seems that the default 1em font-size in IE is two pixels smaller > > than that in MOSe (at least on OS X), but that shouldn't be much of a > > problem. If someone could test this in an old Konqueror or Safari 1.0 > > I think we're golden. > > > > -- > > Mark Wubben > > http://novemberborn.net/ > > > > > > > > > _______________________________________________ > sifr3-dev mailing list > sifr3-dev at lists.novemberborn.net > http://lists.novemberborn.net/mailman/listinfo/sifr3-dev > -- Fredrik Br?nstr?m branstrom at gmail.com fredrik at brnm.se http://fredrik.brnm.se -------------- next part -------------- An HTML attachment was scrubbed... URL: From mark at novemberborn.net Wed Dec 28 22:07:15 2005 From: mark at novemberborn.net (Mark Wubben) Date: Wed, 28 Dec 2005 23:07:15 +0100 Subject: [sIFR3-dev] Compatibility Message-ID: Hey hey, I hope you had a wonderful Christmas! With the new tuning approach [1] it's time to look at browser compatibility. The latest browsers, Firefox 1.0, Safari 1.3, Opera 8.x and IE 5.0 work fine with tuning. It's older browsers [2] which need more help: the user will have to specify the extra horizontal and vertical spacing [3] on the elements. This, in turn, means that sIFR will be more difficult to configure for these browsers. Therefore I'm planning to introduce two modes: normal and backwards compatibility mode. The latter will require the manual setting of the spacing and work in older browsers, the former won't. The idea is that it becomes very easy to deploy sIFR, and if necessary it can be enabled for older browsers. Since sIFR is fully degradable this won't be a problem: users with older browsers will see the normal text. I'm also thinking of not supporting IE/Mac altogether. Not because I don't think it'll work -- after all, sIFR 2 works in IE/Mac -- but because this browser acts funky with Flash movies and CSS. It's just not worth the effort. I'm wondering, what are your requirements for IE/Mac support? Do you need to have a consistent design across browsers, and will you not be able to use sIFR at all if it does not work in IE/Mac? (Please remember that it is disabled by default in sIFR 2). A final issue is Flash support. sIFR 3 will require Macromedia Flash 7 or above (while of course gracefully degrading if you use an older version or Quicktime Flash). It is my understanding however that Flash 8 filters are not compatible with Flash 7 (Am I correct here?). In order to work around this three Flash modes will be introduced: 1. Flash 7 (default; no filters) 2. Flash 8 3. Flash 7 & Flash 8 (requires two differently compiled movie files) The thinking here is that you start working with one movie first, and then, if necessary, compile other movies to use Flash 8 techniques. On the face of it, all these modes look complicated. The theory is that you start simple, and if necessary you can complicate matters. I'd love to hear your thoughts on this. Regards, [1]: http://tests.novemberborn.net/sifr3/experiments/tuning/ (still needs testing in Safari 1.0 - 1.2, Opera 7.x and Mozilla 0.9 - 1.3!) [2]: Safari 1.0 - 1.2, Opera 7.x and Mozilla 0.9 - 1.3 [3]: padding and borders -- Mark Wubben http://novemberborn.net/