> Acceptable software licenses for those should be defined, either
> anythinghttps://opensource.org/licenses  approved, or we could take
> a stand against all or almost all viral licenses.  I certainly do,
> the moment I see GPL/LGPL or worse I move it into "perhaps examine
> for ideas category.  We are not a popular and big enough community
> to allow members of it to try to force copyleft on everyone else.
>
> An exception is the Eclipse license; while the boundary isn't
> precisely defined, it's clear enough for me, if you modify the stuff
> released under it, it's copyleft, you have an entirely reasonable
> reciprocal duty to the community that provided it to you.  But
> anything that for example links or uses it remains your own IP.  E.g.
> patch the core if Clojure and you have a duty to make this available,
> but your independent code written in Clojure is yours.

General notes on licensing:

- There is no such thing as a clear license. The war stories are many.
   Fans of particular licenses like to advertise how clear it is. None
   of those people are lawyers nor ran a big software project for years.

- The rational choices are:

   1. If you want copyleft, pick the most popular copyleft one.

   2. Pick the most popular license in the community you work in.

   3. Since all licenses are unclear anyway, might as well pick the
      shortest widely used license, which is probably ISC.

- We'll never reach consensus on copyleft vs non-copyleft, it ties
   very deeply into personality, political outlook and past experiences.
   A practical choice is to use a short permissive license for things
   that need to be shared between copyleft and non-copyleft camps.

- In many parts of the world an author can't deliberately place works
   into the public domain, so some kind of license is needed.