On Mon, 2012-09-10 at 17:57 +0200, Krzysztof Kosiński wrote:
2012/9/9 Tavmjong Bah <tavmjong@...8...>:
I've posted drafts of my talks for SVG Open 2012 (now known as "The Graphical Web") to my website. Comments welcomed. I give both talks Tuesday morning in Zurich.
The talks use JessyInk and require you to hit Page Down twice to get to the first slide. Firefox messes up some figures while Chrome doesn't do indenting text spaces well.
Inkscape Update:
http://tavmjong.free.fr/SVG/SVGOpen2012/INKSCAPE_2012/svg_2012_inkscape.svg
The presentations look very nice. Here are my comments:
You made a spelling mistake in my name :) It's Krzysztof not Kryzystof.
Oops.
Regarding the wishlist:
- Filter primitive subregions are rather useful for feFlood and
feImage and obviously necessary for feTile, they're just not implemented yet. This is down to the fact that the person who originally wrote the filters did not really understand the canvas object hierarchy or know C++ very well (e.g. he reimplemented an auto-growing vector!) - they are somewhat crudely bolted on. For example, they do not appropriately react to the 'update' callback. Poor handling of filterRes is also a consequence of this.
Reminds me that feImage needs to get fixed (coordinates are used wrongly). I had fixed it right before the big Cairo merge. I'll have a look after the SVG working group meeting.
- enable-background is strangely defined but implementable. Some
rationale on why it is defined this way would be useful.
- If we want feDropShadow then why remove feMerge? On the other hand,
combining feComposite and feBlend is a good idea, I don't see why they are separated. Preferably, feComposite should be expanded with the operations of feBlend and feBlend should left as a deprecated element.
- Warped text is too specific. This is achievable by combining normal
text with envelope deformation. There's actually an envelope deformation LPE already present in Inkscape. If it doesn't work with text, it's a technical problem in Inkscape.
Yes, if you convert it to paths. This is about keeping the text as text. It comes out of a question raised about how text should be deformed when on a path.
- Hatched fills are OK, but I'm not sure how this would be implemented.
I think you should also mention that at the moment 0.49 is not usable for images which contain bitmaps, as Cairo still does not implement bitmap downscaling.
Will do so. (Any workaround planned?)
Good luck :)
Thanks
Regards, Krzysztof
Tav