In arguing elsewhere for a strict implementation of the SVG 1.1 F.6 Elliptical arc implementation notes I noted two points that 'everyone' gets wrong:

• rx/xy values forced to absolute value, and
• nonzero flag values forced to '1'.

Now I see that the grammar for SVG2 has dropped the insistence on rx/ry being nonnegative numbers. This is wonderful as:

• previously the grammar contradicted the implementation notes,
• the browsers got it right anyway, and
• everybody else (many software libraries/modules) got it wrong as they read only the grammar and not the implementation notes.

Now 'everybody' might have a reason to believe the implementation notes when they say:

If rx or ry have negative signs, these are dropped; the absolute value is used instead.

 
However, for my second nit, I see that SVG2 C.4.2. Out-of-range parameters still insists that

Any nonzero value for either of the flags fA or fS is taken to mean the value 1.

while the grammar still insists on flag::=("0"|"1").

Now in fact browsers do insist on flag values of '0' or '1'. And the SVG 1.1 test suite insists on flag values of '0' or '1', which is what the browsers probably copied.

The C.4.2 note carried over from SVG 1.1/1.0

Any nonzero value for either of the flags fA or fS is taken to mean the value 1.

should be dropped as it was never implemented by at least the major browsers, which chose to support the stated grammar and the expected test results.