Re: Some hard numbers on licenses used by elisp packages

[Jonas Bernoulli <jonas@bernoul.li>]

>   > I have looked at some files and made some improvements.  We are now down
>   > to roughly 200 packages that still need to be checked.
>
> That's much better than it previously appeared.

And it is getting better.  Patience seems like the best strategy right
now.

* I have improved the heuristics etc.  More licenses and permission
  statements are now being detected and identified.  I also make an
  effort to differentiate between variations of a license.  This
  obviously can still be improved, but that isn't urgent.

* I have looked at nearly every package with an "unknown" license, and
  have split that set into two new sets: "none" and "failure".

  - failure: The licensing terms are specified somehow.  But my tools
    cannot detect that.  In some cases that is a failure of my tools,
    in other cases it is a failure to specify the terms in a way that
    could possibly be detected automatically.

    I don't think it is important to get this number down at this point.
    Doing so is certainly possible, but teaching the tools to match more
    also comes at the risk of false-positives.

  - none: Looking at the package for half a minute did not reveal any
    license of permission statement.  Like my tools, I myself can make
    mistakes, so this is an upper bound.  A small number of these
    packages (let's guess ten) probably do actually feature some kind of
    permission statement, which I failed to see.

* I have contacted nearly all of the authors of packages that appear to
  not specify a license.  These packages are being identified below by
  the pseudo license "pending".

  Responses - all positive - have started to dribble in.  Several dozen
  packages now have proper licensing terms thanks to these efforts.  But
  most of the authors I did not contact until last night, so a few dozen
  additional quick responses are to be expected within a week or so.

  What is needed now is patience.  As I have said elsewhere in this
  thread I have years of experience contacting authors of elisp packages
  to ask them to fix something that they might consider to be a very
  minor issue.  Usually a small number of authors (up to 10%) respond
  and acts very quickly, within hours or days.  A large number of
  authors (up to 60%) respond within one to three months without
  requiring any reminders.  The rest requires reminders.  However these
  reminders should not be send out until a few months later.  In my
  experience, sending out reminders too early, reduces the likelihood of
  something being done.

The newest statistics:

| License       | Count | Percent |
|---------------+-------+---------|
| GPL-3+        |  2295 |      62 |
| GPL-2+        |   617 |      17 |
| ((pending))   |   128 |       3 |
| MIT (x11)     |   117 |       3 |
| MIT           |   102 |       3 |
| GPL-3         |    98 |       3 |
| ((failure))   |    66 |       2 |
| BSD-2-clause  |    65 |       2 |
| GPL-2         |    40 |       1 |
| unlicense     |    30 |       1 |
| Apache-2.0    |    26 |       1 |
| BSD-3-clause  |    26 |       1 |
| public-domain |    23 |       1 |
| ((none))      |    16 |       0 |
| as-is         |    15 |       0 |
| WTFPL         |    11 |       0 |
| GPL-1+        |     4 |       0 |
| ISC (and)     |     3 |       0 |
| Artistic-2.0  |     2 |       0 |
| CC-BY-3.0     |     2 |       0 |
| GPL           |     2 |       0 |
| ISC           |     2 |       0 |
| LGPL          |     2 |       0 |
| AGPL-3        |     1 |       0 |
| AGPL-3+       |     1 |       0 |
| BSD           |     1 |       0 |
| BSD-3         |     1 |       0 |
| EPL           |     1 |       0 |
| EPL-1.0       |     1 |       0 |
| ISC (and/or)  |     1 |       0 |
| LGPL-3        |     1 |       0 |
| LGPL-3+       |     1 |       0 |
| LGPL-3.0      |     1 |       0 |
| MPL-2         |     1 |       0 |
|---------------+-------+---------|
| total GNU     |  3063 |      83 |
|---------------+-------+---------|
| total         |  3703 |     100 |

  Best regards,
  Jonas