I recently noticed that Creative Commons has dropped all non-Attribution licenses from the “Choose A License” page. That means CC is trying to make all people include Attribution clauses despite well known and long acknowledged problems with them. In the late nineties, people began to realize that the attribution clause in the original BSD license was causing problems with projects that involved large collections of packages, such as GNU/Linux distros (more on this). Basically, the attribution clause required them to have hundreds of attributions on any materials they published, including any small advertisement. Imagine trying to include hundreds of attributions on a banner ad and you get the picture. Then factor in the amount of work that someone would have to do just to keep track of all these attributions.

The current BSD attribution clause is harmless because it clearly defines how attribution is given: just include the license file with the files that you distribute. When releasing software, there are many files that need to be included, and there are already well established techniques of managing files, so they is a minute burden. On the other hand, the Creative Commons’ attribution clause is very broad, making it worse than the original BSD license. (see section 4(b) of Attribution 2.5 Legal Code). The CC licenses are a notable improvement over completely closed licenses, but when compared to the proven free software licenses, there are problems, including some that have been recently solved.

Lawrence Lessig and company are most definitely lawyers. Lawyers generally believe that the law works while forgetting that the law is really expensive. So when it comes down to it, are you going to spend $5000 to $10,000 suing someone who didn’t give you attribution? If it was free content, they would still be able to distribute it for free. What I am saying is that yes, I believe that we should give credit where credit is due, but attribution clauses don’t really help with the problem while adding a real cost to using media with such a license.

The cost is that you have to track down and ensure the attribution information and make sure you get it correct. The CC licenses makes it even worse because it does not even specify how you must give attribution, but instead says that its up to the author. So if you were making something like Debian, you would have to track the specific attribution requests of thousands of different authors. Same goes for a track that uses a lot of samples, or a movie built from footage of many sources. That takes a lot of time, and the person has to be educated in the specifics of the license issues.

Another cost is for works whose author can not be found. If you stumble upon a 50 year old work with a CC-BY-SA license, and you want sample from it. You will need to track down that author in order to find out how you must attribute the work. Unlike the BSD licenses, the CC-BY-SA license does not specify how you should make the attribution, it leaves it up to the author. This is the exact same problem with current copyrights and will lead to lots of abandoned work being left in limbo. The CC license is an improvement over more restrictive licenses since the work would still be able to be viewed and distributed, it would not be legal to sample from it or modify it.

There are more and more examples of unforeseen problems that are cropping up as we find new uses for collections of human knowledge, and new methods for searching through them. One example that was recently brought to my attention is scientific data sets. As scientists post their data in common repositories, more people can search through them using and ever expanding selection of tools. Now consider a search that returns data from 50,000 different data sets. If those data sets were released with a license that used an attribution clause, you would now have to track down how 50,000 different authors want to be attributed. That is serious work which would seriously curtail research.

There are some attempts to deal with these issues in the CC-By clauses in the licenses by discussing how the attribution needs to be handled when included in a “Collective Work”, but it is still very murky, and there is not a “Human Readable Code” version of this text for those of us who just want to put our work out there rather than try to comprehend thick legalese. For more discussion on the problems of CC licenses, check out the reports Debian-legal workgroups on this topic:

debian-legal Summary of Creative Commons 2.0 Licenses
Debian Creative Commons Workgroup report
Debian and the Creative Commons

As for the other side of this debate, this blog post covers some of the possible advantages of attribution clauses:

Attribution and full-blooded open sourceness