[tex-hyphen] A "short" summary of the licence-related meeting

[Mojca Miklavec]

Hi,

This is just a short summary of the meeting we had with Google (&
Mozilla) representatives. We might write a better summary along the
way, but here are the major points.

Recently a Google employee (Xiangye) contacted most of the pattern
authors, asking them for permission to change the licence, so that
they could include the patterns into Android (and potentially other
products in the future). It would be great if the problem of the
constant need to keep relicensing patterns would be solved somehow, so
that individual project wouldn't have to hunt down every single author
every few years.

Why specific licences are not acceptable to Google?
- LPPL: too ambiguous, confusing and too exotic; it's not that one
particular clause would be problematic, but the Google employee claims
that the text is open to too many different interpretations (and with
no previous cases in courts); citation: "imposes a lot of confused
definitions of derivative work"
- GPL: theoretically fine as long as the data is dynamically loaded,
but they want to avoid problems
- LGPL: version 2 is conditionally ok, but version 3 is not
- CC0 (this was a proposal of the "least common denominator" that
should be compatible with all other licences): they want to avoid it
due to one particular clause

Google's preference for the licence would be MIT (or BSD, Apache,
Unicode). Nobody is aware of any conflicts of the MIT licence with
other licences (at least not from the opensource world; there might be
some commercial licences that are not compatible).


---------------------------
Inclusion of patterns into CLDR - Unicode Common Locale Data Repository
---------------------------

We were discussing inclusion of patterns into CLDR from two aspects:
- technical
- legal

>From the legal point of view any patterns that would go to CLDR would
require author's extra permission. (Thus it might be more convenient
to keep all patterns at a single standalone place and submit just the
complying patterns to CLDR.)

>From the technical point of view it would make sense to also collect
dictionaries as opposed to just hyphenation data (wherever that is
even feasible).


In the long-term perspective it could make sense if pattern authors
would agree with giving the copyright to the Unicode consortium or
TUG, so that if a need arises to do something special with the
licence, Unicode or TUG act in author's behalf (still giving authors
full credit, but potentially allow relicensing to a more suitable
license). This is what has been done for OpenOffice years back for
example. (Apart from lack of authors' consent, one of the potential
problem might be lack of lawyers in TUG; but TUG would otherwise
probably be the most suitable organisation for that.)


---------------------------
Feedback from patterns authors to Google's requests
---------------------------

- Many authors are no longer reachable.
- Many authors agreed to a licence change in response to Google's
enquiries, but didn't take any action yet (either didn't bother
enough, didn't have time or didn't know exactly what to do).
- From those who replied, there was an almost even distribution of
different licences that authors picked, so at the end Google might end
up with many different (albeit complying) licences anyway.

Mojca