getScreenCTM & parent.Offset

-getScreenCTM

On getScreenCTM, the spec states:

Returns the transformation matrix from current user units ... to the parent user agent's notice of a "pixel".

But most UA's are behaving (more accurately) like what is written in 1.2:
Returns the transformation matrix from current user units ...  used by screenX/Y of the MouseEvent interface.
Should we presume the latter is correct?
-offsetTop/offsetLeft
Webkit return these properties on the SVG element, which are actually HTMLElements, so shouldn't have these properties.  I think thought that within our conversations around SVG being treated like inline, they probably should in HTML (they do in XHTML in webkit, but haven't tested this in HTML)


Your help is appreciated.