Re: text-intro-06-t reference image

On Thursday, December 9, 2010, 9:36:48 AM, Cameron wrote:

CM> http://dev.w3.org/SVG/profiles/1.1F2/test/harness/htmlObjectMiniApproved/text-intro-06-t.html

CM> Due to my unfamiliarity with the script, and the shaping that’s going
CM> on, I cannot tell if the pass criteria actually match the reference
CM> image.  Could someone confirm for me?

The reference image is using the wrong font, for a start. Apparently Batik has an issue seeing installed system fonts (this may be a Mac-specific problem). 

Second, the test claims to be testing ligature formation but is also testing text-align. Opera 10.63  passes, but the text is rather squashed up because the original alignment points were put in based on an incorrect assumption of where the text should go. Batik seems tu use the opposite meanings for start and end (and the alignment points reflect that).

I see that for SVGT1.2 we simplified the test to just test ligature formation

http://dev.w3.org/SVG/profiles/1.2T/test/svg/text-intro-06-t.svg

That is one option. Or we could keep the three copies, but move the alignment points further apart so the text does not overlap.

The reference image, and Opera, are constructing the ligatures correctly.

The pass criteria refer exclusively to text-align, while the test description refers to the mandatory ligatures part.

In my opinion the 1.2T test is better, with the text being larger. I suggest that the current test be copied over to a new, unapproved test which tests both alignment and ligatures, and the 1.2T test body content be used here for the approved test. The 1.2T pass criteria are better, too.


-- 
 Chris Lilley   Technical Director, Interaction Domain                 
 W3C Graphics Activity Lead, Fonts Activity Lead
 Co-Chair, W3C Hypertext CG
 Member, CSS, WebFonts, SVG Working Groups

Received on Thursday, 9 December 2010 16:28:44 UTC