[Mageia-dev] Mirror layout

[Romain d'Alverny]

On Fri, Dec 10, 2010 at 17:04, Wolfgang Bornath <molch.b at googlemail.com> wrote:
> 2010/12/10 Romain d'Alverny <rdalverny at gmail.com>:
>> Let's try this: what if we consider, at first, that software patents
>> were a non-issue? (that is, we just consider they are all invalid as
>> such).
>
> >From a mirror maintainer's view:
> If I am in a country like France (not acknowledging SP) I can ignore
> safely the issue
> If I am in a country which does acknowledge SP I will have to decide
> for myself if I want to take that risk. THis may be an easy decision
> in some countries but in others (like the USA) it may be a hard
> decision, also depending on who I am (a private person or a large
> institution/organisation).
>
> >From the users's view:
> Living in a country without SP (or not caring about the issue) it is
> the easiest way. I can use automagical setup of media and need not
> worry about an extra repo to set before I can watch my DVDs :)
> Again in a country like USA I have to decide for myself - this would
> be a nightmare if the "whatever-dubious" software is included in the
> normal repos. I would have to find out by myself which is "ok" to use
> and which is not.

Ok, but you still take into account SP in your answer. :-p (we would
have come to that, but the idea was to think about it from a naïve,
software-patent-free perspective).

So... let's take the European Union case, namely, no SP (see next §).
We just don't have to split the media. Easier for everyone in this
territory.

Well, thing is, even though there is a European policy about that...
it's been unbalanced and challenged for years. In the end, it's just a
mess and law is lagging behind; if not just broken. So it's going to
be, anyway, a battle of positions before something clear and
"definitive" comes out.

Still. What if? Where is it an issue to distribute/mirror then? Only where:
 - SP do exist by law;
 - and specific SP are registered on pieces of software distributed/mirrored;
 - and these SP are not invalidated (de facto or obviously) by some prior art;
 - and those SP are likely to be enforced (that is, practically, there
is a minimum incentive for a patent holder to raise his hand; or, that
it is worth it to enforce it).

Coming back to my previous post, maybe we should just try something
simple and go from that.

Shall we take a stance despite SP (for they are just not relevant),
worldwide or more reasonably have a simple, risk-management based
attitude to alter our media/mirror policy only if something happens?

That would relate somehow to Debian policy in this regard, as misc
mentionned in the "Why validate software patents ?" thread on
mageia-dev, Dec. 8th, ~12:58. And that sounds a decent attitude; see
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=365390#20 or
http://lists.debian.org/debian-project/2005/01/msg00316.html .

So why not take it the same way, not have a distinct media and see
what happens on notices?

Yes, this would raise mirroring issues (that would occur anyway), that
we can minify with a policy to manage that globaly; that when an issue
is raised:
 - try to fix it for the whole project, not just for some area, so the
weight of the whole community (not only Mageia for that matter) is
behind the case, and not only a local chapter;
 - remove only software that gets a full score through this list (to
be updated):
   * has an identifiable patent registered on it (exact code/method,
exact patent description);
   * has been notified about by the holder (or representative);
   * whose holder is identifiable;
   * whose holder does not provide a free use license;
   * which patent:
     - has not expired;
     - is not already invalidated;
     - has no obvious prior art;
     - is being actively enforced;
   * other?
 - such software or components would be made separately available only
(so we still manage that it depends on the territory).

So we start with an empty "tainted/whatever" media. And again, see
what happens. That's an option.


Romain