[okfn-discuss] possible OKD conformant licence: The MirOS Licence

Tue Jul 8 10:04:25 UTC 2008

On 07/07/08 15:15, Patrick Anderson wrote:
> On Mon, Jul 7, 2008 at 7:39 AM, Jonathan Gray <j.gray at cantab.net> wrote:
>> Welcome to the list Patrick!
> 
> Thanks.
> 
>> The GPL/AGPL aren't included because they are covered by the open
>> source/free software definitions.
> 
> Does this mean all licenses covered by those definitions are
> automatically INCLUDED, or that they are automatically EXCLUDED?

To my mind they would be automatically *included*, i.e. any licence 
compliant with the Open Source Definition would be compliant with the 
Open Knowledge Definition. However, I note that

a) I am not absolutely certain of this. There might be edge cases where 
  there would be a difference (e.g. I do not know what the deal would be 
under the OSD if you distributed code that was openly licensed but the 
documentation (or artwork) distributed with the code was not openly 
licensed -- under the OKD this would clearly not be an open 'work').

b) Importantly the OKD talks about works as primary and licenses as 
secondary. Specifically it states [1]

"A work is open if its manner of distribution satisfies the following 
conditions:"

These conditions extend beyond specific requirements on the licence 
specifically item 1 (Access) mandates that the work must be made 
available (this is similar to the idea behind 'Affero' type clauses).

[1]:<http://www.opendefinition.org/1.0/>

That said, the conditions imposed on licenses by the OKD are 
(intentionally) essentially identical to those for code in the OSD.

>> The OKD focuses on knowledge - which
>> includes content and data.
> 
> Yes, luckily I understand that much ;)

:)

> I should have been more explicit.  I am talking about applying the GNU
> [A]GPL to data.  This is already done to a small degree.
> 
> For instance, I follow the development of many Free Software games,
> and know that some of the groups are moving from the CC licenses to
> the GPL for the artwork (3D models, sprites, textures, etc.) because
> it just makes things easier to use a GPL compatible license, and the
> GPL protects the community from those that would otherwise make that
> code or data proprietary.

I am not quite sure what the GPL delivers here above CC by-sharealike. I 
understand that there may be then some compatibility issues as well as 
difficulties where code blurs into content (an issue you mention below)

> Some wikis use the GPL for the 'content'.  i.e.: http://EmacsWiki.org

They may do though I'd suggest it was a better idea to use a licence 
that was focused on content rather than code. (Why you might ask: 
because there will be clauses that are specific to that particular area 
-- e.g. GPL makes frequent reference to 'source code' and 'object code' 
which does not, necessarily, make much sense for a e.g. play or a 
sound-recording).

> Even a small amount of music and video is available under the GPL.
> 
> 
>> From the terminology section:
>>
>> "Software is excluded despite its obvious centrality because it is already
>> adequately addressed by previous work."
> 
> Yes, but I am not talking about software.
> 
> I am talking about applying these licenses to data even though they
> are traditionally used for that other thing called software.
> 
> 
> By the way, some computer languages (such as Lisp) do not
> differentiate between code and data.  I wonder why we feel we must
> enforce that mostly artificial difference...(?)

This is an *excellent* point and something I have also frequently 
thought about -- in my view the distinction between code and data is 
rapidly disappearing. For example in the Open Economics project [2] 
we've been accumulating economic type data. At present we store csv 
files usually with python code alongside to load these up into python 
data structures. However I could just store the data in a python file as 
python data structures to start with ...

[2]:<http://www.openeconomics.net>

However one answer to your question about the distinction is legal: 
there may be different IP rights (or the same IP rights applied 
differently) in different areas. For example, the EU, has a sui generis 
"database right" specifically for 'databases (whereas for normal code it 
is copyright -- though I would note that in many common-law 
jurisdictions data collections have also traditionally been protected by 
copyright). To see how difficult things can get see this article [3] 
which was about copyright in Huffman encoding tables.

[3]:<http://www.law.ed.ac.uk/ahrc/SCRIPT-ed/vol3-2/khong.asp>

Regards,

Rufus Pollock