Based on my reading, the Creative Commons licenses actually put the honus on the creator to define how the work is to be attributed.
“You must attribute the work in the manner specified by the author”
At a point when I was getting many requests for republishing a parody image I’d created (thanks to a cease-and-desist letter), I ended up writing up an attribution policy that spelled out how I’d prefer having my work attributed to me. It’s over-earnest I admit, but I wrote it more for people who’d never heard of the CC licenses.
In my opinion, you’re going above and beyond the call of duty by 1) including the authors name, 2) linking to the authors page 3) linking to the permalink page for the photo, and 4) linking to the license. Adding © + CC is also a nice touch. Personally I believe that having the photo link to the permalink page for the source image is sufficient.
As far as GPLed software, how exactly does it “enforce generational data integrity” (except through multiplicity)? Perhaps I haven’t read the fine print lately, but I thought the essence of the GPL was that you could use and modify GPLed code as long as you release the modifications under the GPL. Based on that logic, I could take WordPress, strip it down for the purposes of publishing a photoblog only, every WordPress user could switch to Habari, and then all that’d be left of WP is my “resized” photoblog software.
It seems like maybe you’re getting at the need for a long term CC-licensed content archive?
honus > onus
Your Flickr attribution looks perfect to me.
There’s an experimental, buggy license chooser — part of the experiment is encouraging people to specify a name and URL to use for attribution, and including that in metadata (which you will probably not like as it doesn’t produce valid HTML).
I don’t think CC is trying hard to be vague about attribution — it is licensor-specified and may be appropriate to the medium, which covers a lot of ground — but not strongly encouraging specificity from licensors has probably been a missed opportunity so far — people love to get linked.
Regarding requiring source material, wouldn’t it be pretty hard to define what what would fulfill such a requirement for many media?
(I work for CC but am not a lawyer.)
I had asked the question about the ‘kind’ of attribution that is expected by a CC license in the commons mailing list and I got the same reply as stated by Justin in the comment above. Based on that feedback, when I released all my images in my photoblog under NC-SA, I specifically stated the nature ot attribution I expected (http://insight.srijith.net/copyright.shtml)
The fact that you do not see this being explicitly mentioned in most CC licensed work kind of reinforces a nagging feeling I have been having for a lot of licensed work in general (GPL/BSD/CC … whatever) – a *lot* of people who use any of these licenses to release their work do not really understand the complete working and consequences of the licenses. It applies more to GPL/BSD kind of licenses than to CC but not by so much!
As for the ‘GPL=CD, CC=tape’ observation, again Justin seems to have takes the words out of my fingers (!).
— Srijith 
You’d never pin this down in a license. Source code is source code. The GPL couldn’t require a game developer’s bible or the higher-resolution artwork that was scaled down for the game itself.
But just as it’s possible to be a rotten GPL citizen, it’s possible to be more generous contributor to the Creative Commons.
I have promotional music CDs (and downloads) that provide isolated tracks for the vocals and each instrument to encourage private (or bootleg) remixes. These are “all rights reserved” materials, but the same encouragement can be given through Creative Commons-licensed releases.
> As far as GPLed software, how exactly does it “enforce generational data integrity” (except through multiplicity)?
You’re right that my analogy isn’t perfect. But consider this scenario: someone writes a program, releases it under the GPL, works on it and loves it for a while. It gets packaged and redistributed in various Linux distributions. Then the original author gets tired of it and deletes the “upstream” web site and all the original code. Now all that’s left is the various repackaged versions. But those versions that live on Debian’s and Canonical’s and Redhat’s servers include the original source code — that’s required by the license, that everybody who distributes binaries must distribute source code. So if some other developer wants to pick up where the original author left off, they can do so with the original source code — not a hacked up compiled-and-decompiled version. Creative Commons licenses don’t require distributors to maintain perfect copies, and I’m concerned that this bodes poorly for reusability / remixability in the long run. Maybe I’m concerned over nothing — that’s certainly one opinion — but that’s the issue I’m trying to raise.
— Mark
>Commons licenses don’t require distributors to maintain perfect copies
Neither does the GPL. Starting with your example, let’s say that Debian decided to modify the program before repackaging it to make it work better with Debain. It is possible that due to the modifications, it only runs (or only builds) on Debian. They don’t have to provide the original source, they only have to provide the modified source. If fact, they do this all the time (try building deb sources on other distros). There is nothing wrong with it, but if the upstream developer were to go away and mysteriously no one had an archive of the original source, the data fidelity has degraded.
I will agree that fidelity loss is more likely to happen with a CC licensed photo than a GPL licensed piece of software, but that has less to do with the respective licenses than it does with ease of modification. It is simple to modify a photo (especially resizing), and not so simple to modify source code (at least, not if you actually want it to work afterward). Because it is easier to do, more photos are redistributed modified, but that isn’t a fault of the CC license.
— Tomp
I like your attribution style, I’m going to start using it for the flickr pictures I use in my blog entries. It’s more concise and obvious than what I’ve been doing. Especially since half of the pictures I chose for the title, and that’s not ‘near’ the actual picture.
I really like Sam’s SVG-isms, but flickr shore is quickr.
w/r/t ‘copied low res versions’; I’m too lazy to even copy them. I link directly to the small version. If they disappear, oh well. I hope my blog can survive without them … and it’s almost more of an embarassment for the owner, who removed it for some reason.
Like Tomp said: pictures aren’t like source code, in important ways.
— Tim Bray
Debian provides the original upstream source (the orig.tar.gz file), along with the patches for the changes they’ve made.
I think this is more of a social and technical matter than one of licenses. Even within a closed environment, it’s often quite hard to find the original photoshop file for an image. Programmers have a culture of preserving code, and much better tools for doing so.
For me, modifying an image is pretty difficult. If I want it to work afterwards.
— Daniel
The degradation a very good point, though it seems like it isn’t really a licensing issue until we see CC Attribution-Preserve-Original. Also Cool URIs shouldn’t need archives.
An extreme case is placeholder code or images – using someone else’s material during development, later replacing it with your own version. The replacement might not be any kind of derivative (e.g. a completely different “photo of bath”), so there’s no legal obligation to follow the original license.
But perhaps there is a compelling social obligation (especially given the potential loss of fidelity). You’ve got me thinking I should always drop a copy of the original of any 3rd-party material I use (no matter in how throwaway a fashion) into my own public repository, with link-rich acknowledgements, down to: X, Y & Z aren’t part of this end product, but helped in its development.
— Danny
Regarding requiring source material, wouldn’t it be pretty hard to define what would fulfill such a requirement for many media?
It is hard for original works, but it seems fairly straightforward for derivative works- it is whatever went into the front-end of your work process that you created the derivative from. One could require hosting/redistribution of those whenever you make a CC-SA derivative, I suppose, but that seems to raise the barrier to SA-ing a fair amount.
For what it is worth, Tomp, Debian and Red Hat both frown on distributing modified sources and prefer to ship pristine tarballs + patches, and then combine those during the build process, but you’re correct insofar as that is motivated by maintainability and is not a requirement of the license.
>Debian provides the original upstream source (the orig.tar.gz file), along with the patches for the changes they’ve made.
Hmm, so they do. You will have to excuse my ignorance, I haven’t used Debian in years. (IIRC the apt-src tool does the patching automatically, so I only saw the patched code back then.) The fact that they do provide the original upstream source doesn’t change the fact they the GPL does not require them to do so. They would not be violating the license if they decided to only distribute patched source, but as you correctly point out, they don’t do that because of their culture. My example still works if you replace “Debian” with “Some Theoretical Distribution”.
>For me, modifying an image is pretty difficult. If I want it to work afterwards.
Resizr will let you resize images with just a few clicks. Granted, there are many modifications you could make to an image that are much harder, but this one is so easy anyone can do it, and it happens to be the same modification that sparked this topic.
— Tomp
The biggest problem with requiring source is that source for music and movies is huge compared to the flattened and compressed presentation version.
Moreover, what is the preferred format for making modifications if an intermediate author wants to change tools? Suppose I make a diagram using OmniGraffle. The source code is obviously the .graffle file in app-specific proprietary format. Suppose someone takes a PDF produced by OmniGraffle and edit it in Illustrator. Now you cannot go back to OmniGraffle editability. Data has been lost. What should be distributed to ensure downstream freedom?
GFDL and GSFDL make this more interesting, because an OmniGraffle file obviously is not Transparent but the original author / copyright holder him/herself is the one granting the license, so who can enforce the Transparency requirement (legally rather than socially) against the original author?
With CC-by-sa this is not a license problem. It is just a practical problem.
Despite these problems, I think the GSFDL draft is the most interesting recent development in the non-software copyleft space.
http://www.advogato.org/article/851.html
For the CC founders and many of CC’s advocates, FOSS’s success is a source of inspiration. However, despite CC’s stated desire to learn from and build upon the example of the free software movement, CC sets no defined limits and promises no freedoms, no rights, and no fixed qualities. Free software’s success is built upon an ethical position. CC sets no such standard.
— Noah Slater
I think you are missing something.
Requirement to provide originals is on most cases just too big burden, concerning audio and video especially.
So when original author does not want or cant provide originals (cc is an altruistic act on most cases), assuming they are required, what kind of punishment you have in mind? For releasing some photos for free and get punished after HDD blew?!? If not, what’s the point of that requirement?
— Peter
‘Source’ is a variable term, best defined as ‘the preferred format for making modifications’, which is what you’re really getting at. So if you are really serious about this, then when you take a high-rez version and modify it, what you really ought to be redistributing is a GIMP or Photoshop file where you’ve made your modifications on separate layers (ie. one that still contains all the original bits).
However, in practice almost no-one does this except design shops that have some kind of asset management system to track versions, formats, and derivatives (some of this ‘preserving-the-original-data’ capability is in photo management programs like iPhoto, though).
Meanwhile, you may have noticed that the GFDL has a similar problem. I’ve seen several GFDL licensed books that are only available as PDFs, and while it is apparently *possible* to output PDFs that are amenable to direct human editing, I have yet to see one that satisfies this requirement.
Furthermore, for images the GFDL explicitly identifies one of the ‘transparent’ formats as JPG, which of course has it’s own ‘generational degradation’ problems.
In any case, this set of problems you’ve identified are not unique to the CC licenses.