December 20, 2002  ·  Lessig

Matt Croydon wonders about how CC licenses will interact with software. In a careless earlier version of this, I said they won’t. Sam Ruby suggests the most I could mean by that is that our energy will be directed elsewhere. Indeed, that’s the most I mean. We share RMS’s concern that there is a proliferation of licenses in software. And our view was that there was a dearth for other creative content. Thus we start outside the software world. But creative reuse of creative content is what CC is all about. My apologies for any confusion.

  • http://weblog.infoworld.com/udell/ Jon Udell

    Hi Larry,

    Should I not be using a cc license for the LibraryLookup project, then, in your view?

  • http://weblog.burningbird.net Burningbird

    Jon beat me because that was one of the questions I was going to ask.

    More, this opens the door to trying to interpret what you mean be software. For instance, in the RSS-dev list, there has been considerable discussion about incorporating CCLs into RSS feeds. Would we consider an RSS feed to be software? Would it make sense to put CCL information about a resource into a feed that contains an excerpt of that resource? Or is the CCL attached to the resource all that’s needed.

    If a CCL attached to a RSS feed differs from that attached to the resource directly, I’ve heard that we’re to interpret this to mean we pick which CCL to apply. But RSS is usually machine generated — wouldn’t want assume, then, that the RSS feed could be error, and one should support the CCL within the resource directly.

    If the CCL is attached to the RSS feed itself, exactly what does unaltered copies of the work mean in this instance? Display the markup? Or the items? One could say I should use my commonsense to interpret this, but I’ve not found that law and commonsense to synonymous.

    If a CCL is propagated with the RSS feed, do aggretators (tools that redisplay the info or publish it such as O’Reilly’s Meerkat, and News is Free), need to also display the CCL?

    And when CCL is attached to resources that are in some variation of XHHTML, exactly what does it apply to? The content? Or the look and feel of the site? All of the above as a package? Or do we need to have CCL for page design, and CCL for content. Or is one CCL embedded in a page tantamont to saying “this applies to everything — design and content”?

  • http://guests.evectors.it/zoe/ ZOE

    What about the source code itself? Could the source code be considered as “creative content” and the resulting binary (aka software) a “derivative work” in the sense that the “Work was translated, recast, transformed, and adapted” to… run as an executable binary…

    Bottom line: could a piece of code be considered in the same way as writing an article or book? And the software resulting from such a source code be viewed as “Derivative Work” in the sense of paragraph 4.c of by-nc: “binary version of the Work by Original Author” or “binary version based on original Work by Original Author”?

    Does that make any sense?

  • http://guests.evectors.it/zoe/ ZOE

    Or perhaps software could be viewed as a “performance of the Work”? The “Work” being the source code…

  • Lessig

    Great questions all, and I’m sorry I wasn’t clear before. We don’t mean to forbid the use with software. We just didn’t write the licenses with software in mind. But if the license fits, wear it. Obviously, adding this into RSS is a critically useful adaptation, and we encourage it.

  • http://weblog.burningbird.net Burningbird

    Appreciate your additional, comment but unfortunately, this raises more questions.

    In a post I wrote on this (at http://weblog.burningbird.net/fires/000766.htm ), I brought up the fact that, for weblogs at least, individual posts could have different licenses attached, granting each different rights as to republication.

    However, RSS feeds are template based and automated. So the RSS template would be setup to grant public domain access to the text of an entry, automatically, but the actual entry itself may be much more restrictive.

    A person republishing the text does so based on the license in the RSS feeds, which is counter to the intent of the person who wrote the text.

    Can you not see potential problems in this in court. In fact, couldn’t this open the door up for potential lawsuits against the tool maker as well as the person who re-published the work? (Worst case scenario.)

    Additionall, you say it’s a critically useful adapation. Why? Saying so on your part doesn’t explain to us why this is so, and this type of explanation could help us understand how to effectively apply license in other circumstances.

    And you also say the CC encourages incorporation of license data into RSS feeds. Why?

    I know there’s discussion about being able to search on this license information on the web, looking for public domain material. However, this seems a case of putting the reuse capability before the content.

    Wouldn’t one search for content first, and then look at license? And the license would be attached to the content.

    Also, RSS feeds have no persistence. An item rolls off the feed just as quickly as it rolls on. RSS feeds really wouldn’t be effective sources of persistent data for future discovery.

    I’m sorry, and I appreciate your answer. But all you’ve done is raise more questions.

  • http://www.scripting.com/ Dave Winer

    Shelley, RSS feeds are persistent if you want them to be. I maintain an archive of Scripting News in RSS, and soon all users of Radio will have the feature. I understand that people are working on the same thing for Moveable Type.

  • http://www.scripting.com/ Dave Winer

    I’ve been talking with Larry about the line between software and other forms of expression for quite a few years. It’s interesting to see the discussion swing around to that with smart passionate people like Zoe, Shelley, Matt and Sam in the loop. I think we’re on to something very important. More here.

  • http://weblog.burningbird.net Burningbird

    Dave, persisting syndication feeds makes not sense. Where is the line, then, between the original source and RSS, if you replicate all the text in the original source and then persist it?

    Why have a weblog then? Might as well just have a RSS generator.

    Regardless, this could go outside of the scope of this original comment, so is best debated elsewhere. And still doesn’t disregard my earlier concerns about including license information in the feeds.

  • http://www.scripting.com/ Dave Winer

    Shelley, the RSS and HTML renderings of my writing are equivalent. One is pretty lean and mean, and the other is all pretty and nice to look at. I don’t see the problem. Makes perfect sense. About debating elsewhere, I think that’s up to our host. Thanks.

  • http://a.wholelottanothing.org/ mathowie

    Shelly, I’m not Larry, but here’s my take on all the questions you’ve asked here.

    Would we consider an RSS feed to be software?

    It’s fuzzy, and the licenses weren’t optimized for software, but I tend to think of RSS files as pretty close to the spirit of the licenses for text files, when applied to the text content contained within them. Jon Udell’s js is a little closer to software and fuzzier still.

    Would it make sense to put CCL information about a resource into a feed that contains an excerpt of that resource? Or is the CCL attached to the resource all that’s needed.

    It’s optional to put the license info in a feed, but seems fine to do that. If I were a publisher looking to reprint something of yours I read in my RSS reader, if I didn’t see the CC license info on the RSS file, I would likely check our site as well. It’s not a requirement but seems nice.

    If a CCL attached to a RSS feed differs from that attached to the resource directly, I’ve heard that we’re to interpret this to mean we pick which CCL to apply. But RSS is usually machine generated — wouldn’t want assume, then, that the RSS feed could be error, and one should support the CCL within the resource directly.

    You’re talking about a situation that is best solved by not applying differing licenses. If I’m a musician and I put out different versions of our band’s MP3s (of the same song) under different licenses, there’s going to be confusion. Ideally, people should apply the same license to their RSS file as their weblog. If not, you might want to ask our lawyers directly if one gets to choose which license they prefer.

    If the CCL is attached to the RSS feed itself, exactly what does unaltered copies of the work mean in this instance? Display the markup? Or the items?

    The spirit of the text file license coverage is about the content within, not the markup surrounding it, be they HTML or Word macros. I would interpret this as the item content.

    If a CCL is propagated with the RSS feed, do aggretators (tools that redisplay the info or publish it such as O’Reilly’s Meerkat, and News is Free), need to also display the CCL?

    Practically speaking, it’s a brand new feature for RSS and I wouldn’t expect readers or aggregators to have support for it. I think it’d be nice if they redisplayed it, but it’s up to the aggregator writer to implement.

    And when CCL is attached to resources that are in some variation of XHHTML, exactly what does it apply to? The content? Or the look and feel of the site? All of the above as a package? Or do we need to have CCL for page design, and CCL for content. Or is one CCL embedded in a page tantamont to saying “this applies to everything — design and content”?

    For HTML, it applies to the content on the page. Ideally, if you wanted to apply a license to everything on a page, you could state that in a way that makes it clear to visitors what your intention is, but the licenses were designed to cover the text, and not everything else on the page.

    There’s a important point about the “distance” between a license and the intended content to be licensed. The farther that distance, the more potential for confusion. Imagine a page of your songs in MP3. You could put one license on the page saying all songs are under it, but you could also put the license URL directly into the ID3 tags on the songs. From the perspective of a new visitor or bot, you could see how the distance shrinking to embedded on the file makes that very clear, versus a blanket license on the page.

    If you embedded RDF directly into your HTML files, and into each and every image using GIF, PNG, and JPEG comment metadata, it might be possible to shrink that distance between the licenses and the content to make it easier for machines to understand what is licensed. Otherwise, you’ll have to explain your intentions on your site. Without an explanation, the assumption is the text of your site is licensed.

    And you also say the CC encourages incorporation of license data into RSS feeds. Why?

    I think it’s just to help with the dissemination of information. If I have a photo blog with the images licensed, and I have an RSS feed of the last five photos, I might as well carry that license over to make it clear to people how they can resuse my photos.

    If I can let people know about my intentions in two places instead of one, why not do so?

    Wouldn’t one search for content first, and then look at license? And the license would be attached to the content.

    If you had a specific re-use in mind (“I need to get a photo for this collage I’m using in my college art show”), and you knew of a license that allowed for your intended reuse, why wouldn’t you search first on the licenses?

    If you look at a place like Getty Images, you can search on the license types, which I do all the time. If I need an image for a client’s site, I often restrict my searches to only Photodisc’s content (Getty is an umbrella site to about 5 big stock photography houses) since I know all the images I get back I can afford (photodisc images are around $30 for web use, while another Getty image by Tony Stone Images can go for upwards of $1,000 per use, per year).

    If you searched by content first, yes the license would be attached, and that’d be one way of doing it. But if you had a specific resuse in mind it should be possible to search by license first. I don’t see any harm in offering two methods of finding things in a search engine.

    Also, RSS feeds have no persistence. An item rolls off the feed just as quickly as it rolls on. RSS feeds really wouldn’t be effective sources of persistent data for future discovery.

    Right, which is why I consider them as temporary duplicate representations of content. If I really want to license your essay about RDF, I’ll go to your site and find it. I wouldn’t use RSS as a first point to search for archived content because that’s not what it was designed for.

  • http://a.wholelottanothing.org/ mathowie

    I forgot to note above that I’m not a lawyer, this is all just my personal opinion and Larry’s free to correct me on any of my points.

  • http://weblog.burningbird.net Burningbird

    Fair answers, Matt. And I appreciate both them and your time. (Since you’re not a lawyer, does this mean I won’t be billed 250.00 rate for the answers? *grin*)

  • http://www.scripting.com/ Dave Winer

    Matt, don’t depend on RSS being a temporary storage format. I already said it once in this thread, but I think you might have missed it. It’s going to be an interchange format, and a backup format, and the basis for blog browsers.

  • http://advogato.org/person/mbrubeck/ Matt Brubeck

    Jon, your license claims to be for a “method for using a Javascript bookmark.” The CC licenses are copyright licenses — they allow you to give away some of your exclusive rights under copyright law. You can’t copyright a “method;” you can ony patent it. Copyright doesn’t give you any exclusive rights to your method, so unless you patent it you have no authority to place restrictions on the method’s use. You cannot legally restrict me from using the “method” given by your code. You can only control the use of the exact code you have written, and works that a court decides are “derivative” of that code. If I wrote a similar bookmarklet with significantly different Javascript, it would not fall under your license even if it used the same method.

    Also, your use of the license has bearing on any future patents. Whether you license the work or not, its publication prevents valid patents of any identical methods “invented” after yours.

  • http://advogato.org/person/mbrubeck/ Matt Brubeck

    In my last paragraph, “has bearing” should be “has no bearing.”

  • Jon Udell

    Matt – Excellent response, thanks. So, it sounds like I should either withdraw or rephrase that embedded statement.

    Regarding the effect of publication on future patent claims on a method: you say the license has bearing, but how so? Doesn’t the mere fact of publication take care of that?

    If I retain the CC license, how would you recommend that I use it?

    Thanks.

  • Jon Udell

    “bearing” -> “no bearing” … OK, sorry, missed that.

    So, same question. Keep or toss the license? And, if kept, how?

  • http://advogato.org/person/mbrubeck/ Matt Brubeck

    Jon: You say in your blog that you considered a public domain dedication, but were concerned about patents. The public domain dedication would still be an excellent choice, and does not change your protection against patents.

    If you want to retain the “ShareAlike” (copyleft) clause in the license you chose instead, you may want to use a license that is designed to protect software. The CC licenses don’t make any source/object distinction, for example. Unlike the GNU GPL/LGPL, they wouldn’t prevent someone from distributing derivative works in a compiled or obfuscated form. I imagine this is part of what Larry meant when he said that the CC licenses are focused on non-software content.

    (Aside: The ShareAlike clause does prevent derivative works from being distributed with “technological measures” that prevent the recipient from exercising rights granted by the license. But it does not seem to require that derivative works be distributed in an easily-modified format. The Creative Commons itself is distributing a Flash animation under a ShareAlike license, without providing the “uncompiled” source file for easy modification. For audio and video works, it’s not clear exactly what the preferred “source” format would be. If these issues are a concern, use a copyleft license with source distribution requirements like the GPL.)

    Regardless of the license you choose, remember that it restricts only your copyrighted work, and works that are considered derivative under copyright law.

  • http://www.dealchimp.com/ njm

    With RSS, if the CCL is attached to and refers to the resource and not the feed itself, does the feed retain the grants/restrictions attributed to the resource? Even if attached to the feed, does the license conflict with implied consent of the data considering that RSS feeds are assumed to be aggregated and potentially published?