Re: [Gnu-arch-users] Front page to wiki now modifiable again

[Tom Lord]

    > From: Aaron Bentley <address@hidden>

    > No matter how much we trust the FSF, later versions of the GPL may have 
    > flaws that we consider unacceptable, whether by accident or
    > intent.  

They may.

    > See the current XFree86 fiasco for an example of a license
    > update that many consider a disimprovement.

I suspect that the FSF is not as free to screw up in such ways.

    > If you care about how your work is used, it 
    > is far smarter to license your work under a real, existing license.

Not necessarily.

The FSF _may_ screw up v3.   That event has probability P and a "cost"
(of some sort) PC.

Restricting your code to v2-only in a world shifting to v3 may
radically undermine the utility of your code in the future.   That has
probability Q and a "cost" of QC.

The only way to make a decision here is make a guess about which is
true:


        P * PC  ==  Q * QC

        P * PC  >  Q * QC

        P * PC  <  Q * QC


My opinion?

I don't see any way at all that V3 can be V2-compatible.  So, I
simplify my analysis by assuming that Q is 1.  I'm left asking

        P * PC  ==  QC

        P * PC  >  QC

        P * PC  <  QC


I very much like my code to be re-usable now and in the future.
Therefore, I think QC is pretty high:


        P * PC  ==  BIG

        P * PC  >  BIG

        P * PC  <  BIG

The FSF has a good track record managing the GPL.   On the other hand,
when they pointed to the Affero GPL as a candidate I did get worried
--- AGPL seems like a big screw-up to me.    So, flip a coin:

        .5 * PC         ==  BIG

        .5 * PC         >  BIG

        .5 * PC         <  BIG


Is PC "VERYBIG?"  I doubt it.   The FSF is socially bound by the
history of the GPL and the rhetoric around it and by it's legal status
as a public-interest NPO.   I'd cap PC as "less than or equal to BIG".

So, the betting money is on:

        .5 * LEBIG      <  BIG

and you should use "v2+".

-t