On 08/04/13 03:19, Luis Villa wrote:
I'd add to this summary that rules tend to be easier to
circumvent;
standards (because they are more flexible) often give the party charged
with enforcement more ways to act against bad actors (but may make even
the most basic enforcements more time-consuming and complex).
You will find exactly the same distinction in tax policy; different
countries go one way or the other. The UK is a rules-based system, which
is why our tax code is 11,520 pages:
http://www.taxpayersalliance.com/home/2011/02/research-length-tolleys-tax...
The rest of Europe, as I understand it, has more of a standards-based
system. If it walks like a duck,... Of course, that is open to the
criticism that people can't actually know how much tax they are supposed
to pay until they go to court, and/or that officials can overreach their
authority.
I think "appropriate" might not be quite right. Certainly
MPL aims at a
more rule-based approach (to enhance predictability), and so is probably
"appropriate." I've never thought about whether a standards-based
approach for GPL is "appropriate" because, at least until Thursday, I
had always assumed a standards-based approach for GPL was necessary,
given the vague and amorphous state of the law when trying to create a
strong copyleft. Having now pondered the issue, I think a standards-like
approach for strong copyleft is probably appropriate, given the likely
importance of enforcement to achieving the goals of a strong copyleft.
But I'm very open to persuasion on the point and don't have strong
commitments/feelings on the issue yet.
I think one question about whether you can even have a rules-based
approach depends on whether there are bright lines you can draw along.
The speed of a vehicle (at speeds a long way from the speed of light :-)
is an ascertainable fact; challenges to prosecutions often arise based
on the quality or otherwise of that ascertaining, but it is a fact.
Similarly, the concept of "file" is still with us in computing and so an
MPL file-level copyleft can be rules-based.
If a strong copyleft is to be rules-based, where are the bright lines?
Are we not back at the argument about where the boundaries of an
application lie?
Gerv