Hey, Jon, others- >From reading the spec doc, it really feels like you need to flesh out the goals and use cases a bit more before you consider the requirements and implementation details. *Why* do you want licenses in the OS? Is it to track the license of everything installed on the OS? Give a list of licenses for users to choose from when creating content? Do license validation? It seems like you need to think through those issues a bit more before fleshing things out more. It is possible you guys have had this discussion elsewhere and I've missed it, but if so,
On 5/1/06, Jon Phillips <[EMAIL PROTECTED]> wrote: > http://wiki.creativecommons.org/Licenses_in_Operating_Systems_Specification > > Here are the associated challenges: > http://wiki.creativecommons.org/Include_licenses_in_operating_systems It is, by the way, completely unclear what the difference between these pages are, at least from the URL ;) Some more specific thoughts: * "What licenses should be allowed?" Why would any valid licenses be disallowed? If the licensor of the code/content/whatever thinks it is a license, then it will need to be represented, no? * "Should licenses be weighted?" You mean, should users (or the system) be able to indicate that some licenses are preferred in certain situations? Or do you mean something else? * "How will they connect with mime-types and how will a system know which filetypes connect with which licenses?" This seems like overthinking to me- I'm trying hard to think of a situation where (1) the user knows enough to meaningfully discern between license types but (2) would be overwhelmed by the list of licenses. Seems unlikely. [That and most licenses don't claim to be valid only for specific mime-types.] * LICENSE_PROVIDER-LICENSENAME-FORM-VERSION.OPTIONAL_FILE_ENDING You might look to the java-style org.creativecommons or org.fsf for the 'license provider' portion of this. * 'Forms': This is probably obvious, but CC is the only license on earth that provides for all three of these forms. So make sure the system supports them, but make sure not to make them mandatory. * "Provide an installation system for this that is cross-platform" nononono! Very good minds, with more relevant motivations, have foundered on this issue. Just provide simple tarballs, or zipfiles, or whatever, that apps can install themselves if they need the licenses and discover they aren't present on the system. * "Is dealing with possible violation too DRM-like?" (1) Validation is not part of a data spec. :) Solve the basic data problem first; and give it a shot to make sure it is extensible if people want to put in the bits that would be necessary for validation. The fact that no one is even trying to do validation right now should be a suggestion about its relative priority :) (2) It is DRM-like, in the sense that both fair use and license incompatibility are usually judgment calls. If lawyers can't agree whether or not licenses are compatible (see: openssl license) then software certainly can't. So you have the option of making licensing decisions for the user (which is the path the DRM industries have taken in the fair use case) or you can do the more limited (but potentially still very useful) 'warn but don't prevent' approach- 'your code might not be legally acceptable- continue anyway? Y/N' Anyway, this certainly seems like a very worthwhile effort, and there are clearly some good thinking here, but I strongly encourage writing down your goals and use cases some more before going further with the details. HTH- Luis _______________________________________________ cc-devel mailing list [email protected] http://lists.ibiblio.org/mailman/listinfo/cc-devel
