[Osmf-talk] New license change proposal status

Richard Fairhurst richard at systemeD.net
Sat Dec 5 10:54:06 UTC 2009

Elizabeth Dodd wrote:
> On Fri, 4 Dec 2009, Matt Amos wrote:
>> i'm not saying that ODbL is ideal - i never have. i'm saying that CC
>> BY-SA is totally, utterly, irrecoverably broken and ODbL is the best
>> alternative we have at the moment.
> I'm not going to accept "this is the best at the moment"
> I'm not going to relicence, because in my jurisdiction a database licence is
> completely untested.

I'm disappointed by that.

It's not true to say that a database licence per se is untested in your, 
or indeed any jurisdiction. Tele Atlas offers you a database licence 
when you buy data from them. Same goes for Navteq. Same goes for 
Google's Terms of Use. They, just like ODbL, are enforced by contract 
where copyright is not sufficient; and certainly contract law is shown 
to work in Australia.

The Nine Network case, which I presume is the High Court ruling you're 
referring to, only deals with copyright protection. It essentially 
reaches the same conclusion as Rural v Feist in the US; facts are not 
worthy of copyright, and nor is trivial arrangement (TV listings and a 
phone book respectively). ODbL was written in full light of Rural vs 
Feist so should withstand Nine Network v Ice TV just as well.

To delve a little deeper:

There is one particular difficulty in jurisdictions like the US and 
Australia, where ODbL needs to be enforced via contract rather than 
copyright. Users are often asked to assent to contracts via 
"click-wrap", i.e. you have to click a button before you get access to 
the data. OSMF feels (probably rightly) that this is too onerous, and 
maybe unworkable, for an open licence. Therefore OSMF's suggested 
implementation of ODbL is closer to "browse-wrap": here is the licence, 
by using the data you assent to it. This is weaker, of course.

I believe that click-wrap and browse-wrap are indeed both untested in 
Australia (I may be wrong). They have, however, been the subject of 
several cases elsewhere.

My reading is that browse-wrap can now be viewed as acceptable if the 
licensor makes clear efforts to show the licence terms to the user. 
Pollstar v Gigmania (US), for example, concluded that "the browser wrap 
licence may be arguably valid and enforceable" back in 2000. In that 
particular case, though, it wasn't, solely because Pollstar had not 
presented it in a sufficiently obvious way to the user. (FWIW I suspect 
our current presentation of the licence on the osm.org homepage would 
also fail this test.) 
http://www.wilmerhale.com/files/upload/pollstar_gigmania.pdf if you want 
to read more.

More recently, Register v Verio (US) also upheld browse-wrap licensing 
in a manner very applicable to OSM: repeated extraction of data from a 
database protected by browse-wrap terms of use. (There's an Australian 
viewpoint on this at http://www.findlaw.com.au/article/12352.htm .)

Certainly browse-wrap is good enough for Google Maps, for example - you 
assent to the contract (Terms of Use) they offer by browsing a map which 
has 'Terms of Use' in the bottom right.

So, what of Australia?

The Copyright Law Review Committee of 2002, commissioned by the 
Attorney-General [1], takes largely the same approach as Pollstar v 
Gigmania. It notes the importance of alerting users to the terms of a 
browse-wrap agreement. It then says:

"...notice may yet be sufficient if the terms are not unusual and/or 
unusually onerous. For example, it might be argued that the use of terms 
which exclude or modify the copyright exceptions, although once unusual, 
is increasingly becoming standard industry practice of which users could 
reasonably be expected to be aware. It might also be argued that such 
terms are not unusually onerous in the same way as those dealt with in 
the ticket cases, which have generally involved terms limiting liability 
for damage/loss to property or injury to persons."

So to recap - and thank you for reading this far:

ODbL's principal provisions, those of share-alike and attribution, are 
ones "which exclude or modify the copyright exceptions". An Australian 
copyright review concluded that these are likely to be enforceable in 
your country if reasonably prominent notice is given, and the cited 
article on Register v Verio agrees. OSMF has thus far shown some 
promising awareness of giving such notice, and I hope it will be kept 
under consideration.

In contrast, Nine Network vs Ice TV drastically curtails the scope of 
CC-BY-SA's applicability (as a copyright licence) to OSM data. A 'bare' 
derived street database, with just geometries, road names and 
classifications - the most commercially valuable dataset OSM has right 
now - would almost certainly be judged "obvious and prosaic", and 
therefore not merit protection.

ODbL therefore, I believe, provides much stronger protection for our 
data in Australia.



More information about the osmf-talk mailing list