- From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
- Date: Wed, 4 Jul 2012 12:32:04 +0200
- To: www-style@w3.org, www-svg@w3.org
Lev Solntsev: ... >> Obviously for a SVG document with the full profile one can as well >> reference a view element with id='xywh' with a viewBox(40,0,20,20) >> to be able to note this: >> background-image: image('sprites.svg#xywh') > >Unfortunately, it doesn't work in that way in <img> tags or >background-image:url() attribute in current implementations. > Well, implementations have lots of bugs and gaps, not just for CSS or SVG. There are much more of those bugs and gaps in implementations than bugs and gaps in recommendations ;o) This is only an indication, that implementations need to be improved dramatically. On the other hand - typically the 'bug probability' of documents of authors is heigher than for implementations displaying such a document, an indication as well, that the majority of content needs to be improved too ;o) It does not necessarily improve the situation, to specify yet another method, incompatible and less advanced than that, what is already recommended. >Also it isn't clear, will the image be clipped by the corresponding >viewport, for example if it is centered. I assume it shouldn't and that >differs from “Image Fragments”. This may depend on the overflow property, clipping and masking etc ... http://www.w3.org/TR/2011/REC-SVG11-20110816/masking.html#ClippingPaths http://www.w3.org/TR/2011/REC-SVG11-20110816/masking.html#AutoClipAtViewportNotViewBox Fortunately in SVG the author can decide what is intended. Unfortunately the clip property is corrupted in CSS 2.1. SVG 1.1 still follows CSS2.0, but if viewers have a bug following CSS2.1, obviously they will fail with clip, therefore better to use clipPath, clip-path etc. Olaf
Received on Wednesday, 4 July 2012 10:32:36 UTC