From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>

Date: Thu, 21 Jan 2010 16:47:44 +0100

To: www-svg@w3.org, fhemsher@gmail.com

Message-Id: <201001211647.44544.Dr.O.Hoffmann@gmx.de>

Date: Thu, 21 Jan 2010 16:47:44 +0100

To: www-svg@w3.org, fhemsher@gmail.com

Message-Id: <201001211647.44544.Dr.O.Hoffmann@gmx.de>

I calculated such tangents manually and tested, if they exist. While tangent(fraction of length) requires some numerics in general, using the usual parametrisations for path fragments it is in most cases simple to calculate tangent(path parameter). However in some cases there is no tangent (SVG has specific rules for this) - and in my tests it turns out, that all or most viewers have problems to determine correctly, that there really is or is not a tangent, therefore it could be helpful as well to have some implementation note with details on how to determine the tangent in such non trivial cases. In these cases the path has a tangent, but this cannot be determined directly with the common parametrisation, this need some more mathematics. To care about this could improve the quality of implementations at all already for those cases, where the tangents are already needed in the current specifications - alignment of markers and stroke-linecap. Currently there is one tricky option to get a graphical representation of tangent(fraction of length) by using animateMotion and keyPoints and related timing (there needs to be no visible motion to get it). to put the tangent object at the intended position of a path with proper alignment, without knowing the position and aligment itself (what does not necessarily mean, that the alignment or position of objects with animateMotion is directly exposed into the DOM - is this available? (DOM is not my field of experiments) if yes, this could be the current tricky workaround). And if it is possible to align it without knowing position and aligment - maybe this informations is not needed at all for many applications. And this does not mean, that viewers are doing it correctly for such problematic and exotic cases I can determine. OlafReceived on Thursday, 21 January 2010 15:51:47 UTC

*
This archive was generated by hypermail 2.3.1
: Wednesday, 8 March 2017 09:47:19 UTC
*