[okfn-discuss] New project - Open Data licence

Tue Sep 4 16:32:20 UTC 2007

peter murray-rust wrote:
> At 07:35 30/08/2007, Jordan Hatcher's lists wrote:
[snip]
>> The licence itself doesn't do too much in terms of restrictions
>> except for a Share Alike requirement.  It has been suggested to me
>> that SA would be particularly bad for the sciences and data
>> integration projects.  I very much would like to hear your thoughts
>> on what you'd like to see with the licence.
> 
> It is probably going to be very hard work to get something 
> "absolutely right" for the sciences. "data" can overlap with 
> materials, with "text" and with "code". Simplistically I have taken 
> the view that "data are not copyrightable" and this phrase was also 
> used by Elsevier in a reply to me about re-using their data (although 
> I strongly suspect that other officers in Elsevier will change that 
> simple view).

Data may not be copyrightable but I wonder what Elsevier would say as to 
whether they are database-right-able? I hope they simply meant they 
waived all IP rights in the data in toto but one would need to be careful.

> I can suggest the following axioms:
> * there is a very large class of scientific data for which the 
> creators wish to claim no royalty, no reach-through restrictions and 
> no explicit permissions. This is what I am most concerned with 
> personally. The authors simply wish re-use, attribution and 
> non-corruption. There are, however, cases where authors welcome 
> community enhancement of the data.
> * the current position, relying on "non-copyright" and "BBB OA", 
> though logically supportable is not workable. There are too many 
> variations of opinion and logic. We therefore need *something* that 
> authors can add to their data which must be transmissible to 
> derivative works if required.

Nicely put Peter.

> I suggest that for the immediate future it would be valuable to have 
> something *very* simple - i.e. at least as simple as BSD - which asserts:
> 
> "we authors state that this is Open Data and freely re-usable for 
> whatever purpose. It may not be constrained by third-party copyright 
> or other restrictions." In practice we have simply labelled our 

In the Guide to Open Data Licensing [1] there is a section on Licensing 
which contains an item we entitled the "The Zero Conditions Option" [2] 
which begins: "The most basic option for providing your data openly is 
to attach a license that simply waives any database rights you might 
have." (Note that we should probably replace database rights here with 
'any IP rights').

Following this there is a license doing just this based on the MIT open 
source license which begins:

"Permission is hereby granted, free of charge, to any person obtaining a 
copy of this data and associated documentation files (the "Database"), 
to deal in the Database without restriction, including without 
limitation the rights to use, copy, modify, merge, publish, distribute, 
sublicense, and/or sell copies of the Database, and to permit persons to 
whom the Database is furnished to do so, subject to the following 
conditions:"

More details at:

[2]:<http://okfn.org/wiki/OpenDataLicensing#head-e2b9239f9664fd717c057aa21abd17e68dadd731>

[1]:<http://okfn.org/wiki/OpenDataLicensing>

> CrystalEye as "OpenData" using the OKFN logo with a hyperlink.
> 
> If you wish to work out a consistent scientific licence you are 
> probably going to have to talk to at least CODATA and Science Commons 
> and this inevitably won't be short. I suggest that in the interim we 
> create a simple placeholder "Open Data" with a 1-line statement of 
> intent, and links to the OKFN site.

I think this is a great idea. We could base this on the MIT/BSD style 
data license alluded to above or perhaps Jordan's TCL redraft will be 
suitable ...

~rufus