license policy for contributed code

[AAMargolin]

It's been pointed out by multiple people that we need to define an (open source) license policy that will apply to code contributed to GA4GH. Lack a of such a policy has prevented some would be contributors from adding to the GA4GH code base, as their (internal) policy requires such a license policy be defined.

This thread is intended to open discussion on this topic, and hopefully move towards agreement.

As usual, I'm less expert than others, but I'll start the ball rolling by suggesting Creative Commons, which is perhaps the least restrictive and consistent with GA4GH philosophy.

[diekhans]

What's the difference between a license policy and a license?

The license has always been Apache 2.0. What else do they need to know? I don't think we should have code with other licenses in the tree.

https://github.com/ga4gh/schemas/blob/master/LICENSE

AAMargolin [email protected] writes:

It's been pointed out by multiple people that we need to define an (open source) license policy that will apply to code contributed to GA4GH. Lack a of such a policy has prevented some would be contributors from adding to the GA4GH code base, as their (internal) policy requires such a license policy be defined.

This thread is intended to open discussion on this topic, and hopefully move towards agreement.

As usual, I'm less expert than others, but I'll start the ball rolling by suggesting Creative Commons, which is perhaps the least restrictive and consistent with GA4GH philosophy.

— Reply to this email directly or view it on GitHub.*

[haussler]

Ccing Peter Goodhand, as his team has done a lot of work on this. -D

On Fri, Jul 10, 2015 at 3:18 PM, Mark Diekhans [email protected] wrote:

What's the difference between a license policy and a license?

The license has always been Apache 2.0. What else do they need to know? I don't think we should have code with other licenses in the tree.

https://github.com/ga4gh/schemas/blob/master/LICENSE

AAMargolin [email protected] writes:

It's been pointed out by multiple people that we need to define an (open source) license policy that will apply to code contributed to GA4GH. Lack a of such a policy has prevented some would be contributors from adding to the GA4GH code base, as their (internal) policy requires such a license policy be defined.

This thread is intended to open discussion on this topic, and hopefully move towards agreement.

As usual, I'm less expert than others, but I'll start the ball rolling by suggesting Creative Commons, which is perhaps the least restrictive and consistent with GA4GH philosophy.

— Reply to this email directly or view it on GitHub.*

— Reply to this email directly or view it on GitHub https://github.com/ga4gh/schemas/issues/359#issuecomment-120540799.

[shadowtramp]

I'll oversimplify it a little, but in few words the license does specify the condition and scope of use for the intellectual property, in our case avro schema and reference code. The contribution agreement specify conditions and responsibilities of parties, eg. GA4GH and contributors, regarding the contributions made by third parties to GA4GH git repos. That is GA4GH needs something in the line of agreements, created by apache foundation: https://www.apache.org/licenses/icla.txt for individual contributors, https://www.apache.org/licenses/cla-corporate.txt for corporate contributors.

For me, for instance, having such agreement in place will tremendously simplify the procedure for contribution.

[awz]

@shadowtramp those CLAs are very standard and look good to me too. An alternative is used by Docker: https://github.com/docker/docker/blob/master/CONTRIBUTING.md

Both of these have a similar legal effect, as I understand it.

@AAMargolin Creative Commons recommends against using their licenses for software (with some caveats). See: https://wiki.creativecommons.org/wiki/Frequently_Asked_Questions#Can_I_apply_a_Creative_Commons_license_to_software.3F

[awz]

@diekhans "The license has always been Apache 2.0. What else do they need to know? I don't think we should have code with other licenses in the tree."

I worry though that without htslib--the reference implementation for SAM/CRAM/VCF which are now GA4GH standards--we just don't have a usable thing. The license of htslib is very permissive (and that might be fine) but it is an immediate exception to the rule that the "tree" is pure Apache 2.

I see this as going two ways: i) we either re-license everything to Apache 2 or ii) we have a more inclusive policy of which licenses are OK for reference-implementations of GA4GH standards.

[diekhans]

Ah, I always forget about HTSLIB, as it's not in the ga4gh github organization and was kind of grandfathered in.

However, I don't think we have a problem. There is only one reference implementation and it's Apache 2. Htslib is used by the reference server as an external dependency via pysam, like many other libraries.

The only issue I see is if someone want to add a new repo with a different license, which we should strongly discourage.

Alexander Wait Zaranek [email protected] writes:

I worry though that without htslib--the reference implementation for SAM/CRAM/ VCF which are now GA4GH standards--we just don't have a usable thing. The license of htslib is very permissive (and that might be fine) but it is an immediate exception to the rule that the "tree" is pure Apache 2.

I see this as going two ways: i) we either re-license everything to Apache 2 or ii) we have a more inclusive policy of which licenses are OK for reference-implementations of GA4GH standards.

— Reply to this email directly or view it on GitHub.*

[diekhans]

@shadowtramp how about doing a PR on the contributing doc to clarify this in a way that makes your life easier? We want to encourage contribution. Seems like Apache has this worked out, so lets just use these.

shadowtramp [email protected] writes:

https://www.apache.org/licenses/icla.txt for individual contributors, https://www.apache.org/licenses/cla-corporate.txt for corporate contributors.

For me, for instance, having such agreement in place will tremendously simplify the procedure for contribution.