william(at)elan.net wrote:
Why was this text in there? I don't understand what makes
"first mechanism" any different then "2nd mechanism"?
If you want something like "positional modifier = modifying
one mechanism", then it's either "before" or "behind" the
affected mechanism. IIRC Mark had a reason to select "behind",
and then "before the first is an error" is a consequence of it.
If you'd select "before" then "behind the last" would be an
error (or at least utter dubious).
Maybe it was an arbitrary editorial decision, or it was for the
symmetry "global modifiers SHOULD be at the end _behind_ all
mechanisms" (?) I forgot the reasoning for "behind", but I
liked it when I first saw it here.
I understand your point perhaps then we need to define that
new type of "global positional" modifiers which positions
would affect both mechanisms and other modifiers.
Possible for spf2.0 or v=spf2 instead of other scope ideas
like spf2.0/scope,scope combined with %e tricks. But not in
"PRA with v=spf1" solutions.
The "global positional" is graph theory, if you have say four
mechanisms m1 m2 m3 m4, and you want m1 m2 m3 in scope s1, and
m2 m3 m4 in s2, then you can't do it with "global positional".
Something like "sc=s1 m1 sc=s2 m2 m3 m4" is what you want for
s2, but it adds m4 to s1, which is not what you wanted.
Same problem as with tags in XML:
<table> m1 <form> m2 m3 m4 </form> </table> is okay. but a
<table> m1 <form> m2 m3 </table> m4 </form> is not okay. Bye.