 And we're live. Hi, I'm Asa Greenwood, a scientist at MIT Media Lab, and I'll be teaching the computational law course at MIT for our winter session next week. And with us today by express invitation, I'm delighted to introduce Bill Rosenblatt, who is truly an expert in DRM and in rights expressions languages. We met at the Open Music Initiative, got to talking about computational contracts and licensing and the essence of how to express legal terms in a way that's machine-readable and processable. And honestly, I was so impressed and also found the topic so very, very timely and relevant to the course that I asked if you'd be willing to give us a guest lecture. And you kindly agreed. So thank you, Bill. And I appreciate your diligence in putting the slides together to really conform them to the topics in our class this year as well. And if you'd be so kind as to maybe introduce yourself with a little more background, and then we can dig right into your presentation. Great. Well, first of all, can you hear me all right? Yep, your audio is good. OK, great. So, Asa, thank you so much for inviting me to do this. It's a real pleasure. It's been a pleasure to get to know you and your very, very interesting program at MIT. I'm someone who turned down MIT as my undergrad institution. So it's a little bit of shame, I suppose, or come up that I'm doing this. But in any case, my name is Bill Rosenblatt as you mentioned. And I am, let's go to, it'll explain a little bit of my background on the first slide. So I'm just going to go right to that, just waiting for it to come up. Or did I, ah, OK, let's, there we go. OK, I hit Share and nothing is happening. And my browser is hung. OK, well, we can just give it a moment. Hangout on there is great in terms of the price it can't be beat, it's free. But sometimes it can be a little bit laggy. So this is not laggy. The app's frozen. I'm getting the blue circle of death and the window is grayed out. OK, let me take a quick look. We're actually in the same office right now. Just one sec. I'm trying to killing it. Where's the hangout? Oh, I see. Weird. OK, well, you're still online. Yeah, I wanted to maybe take a look and come back in. And we're really back in. So you can see we're, ah, OK, but you're back, Bill. And do you want to give the screen share or try? OK, looks good. I can see your presentation. Yep, I can see your presentation. Oh, wait. OK, so we're looking at the presentation. And the title of my presentation is Rights, Expression, Languages, or Copyright Law, Rumbles with Technology in a Dark Alley. This picture here is from the Broadway show West Side Story. The two gangs, the two rival gangs, meet in the back alley and have a gang fight. So just to give you a little bit more about my background, I am a consultant. I've run my firm, Giant Steps Media Technology Strategies since the year 2000. My background is in computer science and software engineering. Professionally, I've been an IT executive in the publishing and media industry. I have been a technology market strategy guy for Sun Microsystems. And I've also been a book editor, a book author, and a radio producer for a period of several years. And I wrote a book on DRM back in the year 2001, which introduced me to copyright law in addition to other activities that I was doing in the publishing industry where publishing industry was trying to get to grips with this thing called the internet that was coming down and was going to destroy everyone's business and so on. I've run a conference called Copyright and Technology, which actually is taking place next week in New York. And I've spoken on this subject matter at the World Economic Forum in Davos. I've guest lectured at various places, Columbia, Carnegie Mellon, NYU, and so on. And I've been an expert witness in various copyright and patent litigations that have to do with copyright in the digital age, digital media technologies, things of that sort. So basically, I'm not a lawyer, but I play one on TV sometimes. So what I'd like to do is start with something. I'm sorry, interrupt. For some reason, we're getting a little, oh, there we go. OK, no, we're good. I just want to make sure that your slides were showing and they're good now. Please continue. OK, so I wanted to start by mentioning a theory that Larry Lessig brought up in his first book, Code and Other Laws of Cyberspace, which is really a book that I think a lot of the people taking this class are going to be familiar with or at least have seen. And I think it's an extremely important book, regardless of what your views are on copyright in the digital age. And one of the things that I think was great about this book is that he characterized these four forces of regulation. And that's when you're talking about the online world or the tech world or you're talking about the real world. There are these four forces that each contribute to the way the world operates and is governed or regulated. And those four forces are the law, which is laws, as we know them. The market, which is just the way businesses behave, the way economics behaves. Architecture, which is the technology or physical aspects of the realm that you're looking at regulating or not regulating, as the case may be. And then norms, which are behaviors, what we're used to doing as humans and what we think is acceptable versus not acceptable. And as we have learned, these four factors differ greatly between the real world and cyberspace, or what used to be called cyberspace back in 1999 when Lessig wrote this book. And the idea of rights expression languages is really to try to, in some sense, bridge the gap between these two worlds and to readjust the balance among these four forces of regulation. And this may sound hazy right now, but hopefully it will be clearer as we go on here. So we're talking about rights expression languages which are in the realm of copyright. And so I thought for those of you who aren't already familiar, it would be useful to go over the basics of what rights are given to creators of copyrighted works under U.S. copyright law. And of course, this is all a U.S.-based presentation. Your mileage may vary outside the U.S. So there is the U.S. copyright law and it accords to the creator of a work, what we call a bundle of rights. And one of them is, in fact, it's so fundamental that many people ignore it, section 101 of the Copyright Act mentions that a work is a material object. So these rights apply to material objects. They do not apply to tiles or bits that fly around the internet and are independent of the medium that they may be stored on at any given time. And we'll talk a little bit about that more later. So then the copyright owner has exclusive rights. The right of reproduction, which is making copies, the right of distribution, which is sending those copies to someone else or sending the original copy to someone else. Creation of derivative works, which are things like film adaptations of books or things of that nature, edits of photographs, excerpts of written works and so on. Public display and public performance, depending on the type of work it is, also a copyrighter and exclusive right. And then recently there was an addition made to the Copyright Act, which according to certain additional rights for certain types of visual works, such as sculptures, paintings and limited edition photographs. The right to have the work properly attributed to the creator and what we call a moral right, which is the right to prevent reputational harm by damage or mutilation to the work or of the work. And then there's something that isn't a right per se, it's a limitation on rights. That we call the first sale doctrine and internationally it's called the doctrine of exhaustion, which is that once you obtain a work legally, such as you go buy a book or you buy a CD or someone gives you, legally gives you a CD or book or DVD or what have you, you can dispose of it in whatever way you want. And the entity that gave it to you, whether it's a publisher, retailer, friend, relative, whatever has no further control over that. So if you wanna resell it, you can resell it. If you wanna lend it, you can lend it. If you wanna bequeath it to your heirs, you can do that. That's a limitation on exclusive rights and copyright law. Paranthetically, there are classes of software to which it does not apply. Okay, so once we have these exclusive rights, we can also talk about how others can use content that you have created. Of course, if I obtain a piece of content, I can use it personally. That's the normal thing. I can license rights from the copyright owner that is rights beyond the ones that I would get normally. And the mechanism for doing that is normally a contract. And then there is something else called fair use, section 107 of the Copyright Act, which is really a defense to an allegation of infringement. It's your accuser of infringing my copyright. You say, no, I made a fair use of it and you explain in a court why you think you made a fair use of the work. It's really understood mostly as a defense to an allegation of infringement, although there was one recent appeals court decision in the so-called dancing baby case where they're starting to look at it more as an affirmative right, rather than as what we call an affirmative defense. And there's a whole bunch of stuff that I could talk about here. There are people who are far more expert in fair use law than I am, but the main thing that we need to understand here is that fair use is not decided by a machine or automated process. And I like to refer people to a seminal law review article from Pierre Laval, who is now a second circuit appeals court judge, the article was called toward a fair use standard. It's there in the footnotes, you can look it up. And the sort of money quote from that article is, we should not adopt a bright line standard unless it were a good one and we do not have a good one. So bear that in mind as we keep going. So what's a rights expression language? It is a machine readable description of rights that relate to the use of a copyrighted work. The idea of a rights expression language is to make rights and licenses understandable and actionable at internet speed as opposed to falling into what I call the trap door into the legal system, meaning hiring a lawyer to explain what rights you may or may not have and then perhaps taking a risk on exercising those rights and so on. And rights expression languages have two bubbles that are distinct yet overlapping. And we're gonna talk about each of those. One is to enforce rights and to act as part of a rights enforcement mechanism. And then the second is just to automate a licensing process for the sake of convenience and automation. Okay, so let's first talk about right rails rights expression languages for rights enforcement. And we are, of course, talking about DRM, Digital Rights Management, which is the field that I got into right at the beginning pretty much. Mark Stefik is one of the handful of people who can be said to have invented this field back in the mid-90s. And he walked into my office one day with his colleagues at Xerox. They were looking to commercialize technology that they had developed. And one of the things that he gave to me was the thing that it says in the second bullet point of this slide, a white paper called Letting Loose the Light, Igniting Commerce and Electronic Publication. This was written in 1995, originally just an internal Xerox PARC document, but it was published subsequently in a book called Internet Dreams where Mark Stefik was the editor. And this is sort of his one or perhaps one of a couple of chapters that he contributed to the book, a very interesting book published by the home team, MIT Press in 1996. He also took out a bunch of patents, the one relevant one of which is this patent that we call the 403 patent system for controlling the distribution and use of digital works having attached usage rights where the usage rights are defined by usage rights grammar. So what that basically means is a DRN system that uses a rights expression language. That's pretty much what that patent is said to cover. He wrote, he designed a rights language called Digital Property Rights Language, DPRL. And it was based on LISP. LISP, as many of you may know, was the Seminole Artificial Intelligence Language and Stefik has an AI background. In fact, I was familiar with some of his AI work that he did as a grad student at Stanford before he joined a part. So Xerox took this and attempted to commercialize it as a product called Content Guard, which was a DRM system. And in 1999, when I was at a publishing company, we were looking to perhaps adopt a DRM system and this is one of the products we looked at. So going to the next slide, this technology morphed into something called XRML, the Extensible Rights Markup Language. So these names are all based on XRML, which came out around 1997, 1998 timeframe as the sort of standard way of marking up data, structured data and structured content. It's one of the most influential technical standards in the modern era. And Xerox decided to glom on to XRML and recast DPRL as an XML derived language called XRML. And first they just released version two of DPRL and based it on XRML instead of Lisp, then they renamed it XRML as an expansion of DPRL 2.0. And for several years, they tried to sell this DRM system to cultures and record companies and whatnot. And it didn't really succeed commercially. So they dropped their products in 2002 to focus on licensing their portfolio of patents. I'm not going to go through this entire slide. You can read it later if you're interested in the history of this, but they submitted XRML to MPEG, which had a standards initiative in this space. And what MPEG ended up doing was adopting a variation on XRML as an MPEG standard, which went nowhere. No one really uses the standard. Hello? Without that, please continue. Keep going? Okay. So XRML was a big deal in the mid 2000s, but no one really used it for anything except Microsoft, which along with Time Warner became a part owner of Content Guard, which is why I call this the establishment rel. It's Xerox Time Warner Content Guard in a French company called Thompson that were the owners of this company. They subsequently sold it to a company called Pendrell, which is a patent licensing company. And that's kind of the end of its relevance, other than if you use certain Microsoft products where they use XRML for license management. So this is just some example code that I found in XRML that shows that you're paying $15. And for that, you get the right to play or view an ebook and the right to print it. So you can dig into the details. It's actually pretty self-explanatory if you're comfortable reading XML. So then we come to the other important rights language, which I call the Indie rights expression language, which is ODRL, Open Digital Rights Language. It was created a few years after DPRL and XRML, created in the early 2000s. And the primary creator of it was a guy in Australia called Renato Iannella, who worked for a system integrator called IPR Systems. And his number one collaborator was an Austrian professor at an Austrian business school named Susanna Gutt. So it's Austria and Australia. And they advanced this thing independently of any companies or standards bodies. They were kind of like the Indies. And they achieved a big success commercially in the early 2000s when ODRL, or a subset thereof, was used in a standard DRM for mobile telephone music services in Europe. So there's something called the Open Mobile Alliance, which is a mobile industry standards body. They adopted a subset of ODRL as a standard for creating mobile digital music services, digital content services in Europe. In the US, however, this failed to catch on because the US, as you know, is more litigious than anywhere else on earth. And there are a bunch of patents floating around, not just Xeroxes, but others. And everyone's afraid of getting nailed or sued for alleged infringement of those patents. And so there was a lot of caution around it. And as a result, momentum over the OMA DRM standard died down and ODRL sort of went into a dormancy state. So I'll hold that thought, we'll pick that up later. Okay, and Bill, here's just the information about ODRL and how it was designed. It was basically designed on the idea of permissions that can be assigned. If you look at this diagram from left to right, I've gone to the next slide. Daza, I'm not sure what to... Yep, I don't see that you're on the next slide. You bet. I would just say, I apologize, we're having some trouble with the presentation from your screen. What do you need to do to get to the next slide? So is that... You're on the next slide, okay, got it. Okay, good. And if you could just leave, let me know when to advance the slide. You see, this slide is taken directly from the ODRL documentation. You've got usages, display, print, play, execute. You have derivative work rights, if you will, excerpt, annotate, aggregate, modify. You've got transfer rights, which are your exhaustion rights or non-exhaustion rights, perhaps cell, lend, give, and lease. And then you have the distribution rights, which are things like move, duplicate, backup, delete, et cetera. So that's ODRL of 1.1, and that's where that was in the early to mid 2000s. So there were a couple other DRM related or enforcement related rights expression languages. One was from Real Networks, which back in the day was a serious invader in the streaming media space and a serious rival to Microsoft, which Microsoft eventually vanquished in competition. And they had their own DRM called Helix DRM for audio and video. And then the other important one, which still exists now is a EBX electronic book exchange, which was created by a startup called Glassbook that was acquired by Adobe. And it became part of Adobe's ebook ecosystem, which they call Adobe Content Server. And in the US, the primary two users of that technology are Barnes & Noble for their noob system. And if you've ever borrowed an ebook, quote, quote, borrowed, unquote, an ebook from a public library, you'd probably use a system called Overdrive, which also uses that technology. And it's also a very wide use in Europe, Asia, all over the world, still very much a going concern. Okay, so now we're gonna switch gears and talk about the other objective of rights expression languages, which is to automate licensing, simply to rely on legal system for enforcement of the licensed terms, if any enforcement, not on technological enforcement means, but simply using rights expressions as a means of automating licensing. And therefore, the primary ability for this type of rights language is in the B2B area, where you're not trying to issue licenses for random consumers, but to business partners, read distributors and aggregators, and that's the port and so on. I would say the analog to Mark Steffick's work in Xerox Park and his paper, Letting Loose the Light, is a professor at Villanueva Law School named Henry Parran who published a paper called Permission Headers and Contract Law, at a conference that took place at Harvard at the School in 1993. This conference called Technological Strategies for Protecting Intellectual Property in the Networked Multimedia Environment. But this is 1993 and they didn't say internet yet. So they had to say Networked Multimedia Environment. This is actually a very, very seminal conference. It's probably the most seminal conference in this field and the proceedings are still available online miraculously. You don't have to go to the internet archive to get them. Most of the papers in these proceedings are fascinating reads, if you're a nerd about this stuff like I am. So Henry Parran wrote this paper about how you can use machine readable headers in contracts to automate licensing. And that was the first appearance of this type of idea, even though he didn't implement anything, he didn't follow lies that he just wrote about the idea and how it worked legally and so on. And Parran was actually an MIT man. He had undergrad engineering and a Sloan MBA from MIT. Then there was a standard or purported standard called the Information and Content Exchange or ICE standard in 1998, which was led by a company that was well known at the time as a content management technology company called In Yet. And they had a bunch of other companies participating in this initiative. It was kind of an over-engineered, overly baroque standard. And their RSS came about at the same time as a much simpler equivalent. And RSS became huge where as ICE kind of faded into obscurity, the ICE idea was related to this vogue, then in vogue concept of content syndication, which was, you know, I'm Connie Nass and I'm going to set out these articles that people will pay 50 cents a piece for. And I'm just gonna rank in the cash because so many people will be interested in these piecemeal articles and photos, but that didn't work for a number of reasons. And so ICE was about automating that type of scheme and it kind of faded. So to talk more, so now we're gonna talk more about this type of rights language, which is arguably the most important application of rights languages. But first I'm gonna talk about creative comments. So creative comments is something that I'm sure you've heard about, you may be very familiar with, but for those of you who are not as familiar, it is a concept that was formed by Larry Lessig and a group of others, including Hal Abelson and MIT in the early 2000s. And the idea was this notion that copyright had gone too far in terms of the rights afforded to the creator. And compared to what copyright started out as when it's beginning in the United States. And they wanted to kind of roll back and rebalance copyright and give more rights to the user, less rights to the creator. And so the motto of creative comments is some rights reserved, you've heard the phrase all rights reserved, this is some rights reserved. The idea is to give the licensee more rights than the licensee would have through the copyright bundle in the copyright law. And if you're talking about the applicability, creative comments apply to any content at all physical or digital. But the idea is that you're, that it's assuming it's based on an assumption that the copyright bundle applies to piles of bits which isn't really true, but it's a useful, it's a useful assumption to make and it's not like the system breaks because it's not true. System is based on real law. And there are people who are more qualified to talk about the new ones than I am, but it is important to feel to emphasize that piles of bits unattached to any physical storage medium are not covered under copyright law. So, creative comments is a set of licenses and the licenses take the forms of legal text and machine readable code that you can put in your files. And if you're a creator and you want to provide a copyright, a creative comments license as I do for most of what I write, you get to choose a few attributes of your license. One is the attribution attribute which basically means you're requiring if you set that to yes or on and then you're requiring the licensee to attribute the word to you. There's something called share alike which means that you're requiring the licensee to offer the same rights that you are offering. Uncommercial means you're going to forget the commercial use of your content without your additional permission. And then the no derivative works attribute is whether you're going to allow people to modify or adapt the work or whether you're going to require them to redistribute it as his state. So that's creative comments. And the way creative comments was originally designed was that the machine readable piece of the license would be in comments in the HTML of a webpage. And Halle Abelson and others set out to improve that in around 2008. The idea was to make creative comments, license terms, more machine readable by using technology such as RDF which is an XML technology for essentially assigning semantics to content to structured metadata which indicates semantics content. Number of different formal taxes for expressing RDF and I leave that as an exercise to the reader as they say, I'm not going to go over that but there are a few different ways that the creative comments rights expression language can be expressed syntactically. And there's also a standard for multimedia files from Adobe called XMP which is basically a standard set of header made of data that you can put in your photos, your illustrations, your movies, et cetera, et cetera. And they provide the syntax for CC rel and that's usable in the Adobe XMP system. So just to be crystal clear about this, the idea of CC rel was not to turn creative comments into a rights enforcement DRM game. It was really meant to improve machine readability of creative comments, comments licenses and eliminate redundancy and hearing to what we call the dry principle don't repeat yourself principle. This is that much interest in CC rel nowadays. I feel there should be more interest in it. It's a beautiful piece of work. There's some sample code which basically says that this photo that's on Flickr has a creative comments attribution non-commercial license and the derivatives license. So if you look at this page here, the second page of the sample code, it says buy-nc-nd which is the attribution non-commercial and part of the machine readable piece of this is on towards the bottom of the second page where it says how you are supposed to refer to the creator in your attribution. It's this URL laughingsquid.com and the name of the photographer is Scott Biel slash laughingsquid. Okay, so that's a sample piece of code of CC rel. Moving on, another probably the most important rights language today for license automation is something called ODRL 2.x. So last, when we last saw ODRL, it was version 1.1. I think 1. something, it was the mid 2000s and it had faded into obscurity as people grew nervous about the patent scene and DRM was sort of becoming less used and certainly in a music space, it was being phased out. Now it's 2011 and Renato, you know, it comes back and he submits ODRL, he revamps ODRL version 2 of it and submits it to the worldwide web consortium with a view to it being used for license automation as part of something that he called the permissions and obligations working or PLE permissions and obligations, expressions. So he was very, very adamant to point out this is not about DRM, this has nothing to do with any patents that talk about the use of DRM with the rights language. This is just about expressing machine readable permissions and obligations. And about a year ago, almost a year ago, the W3C advanced that to recommendation which means it's an official W3C standard. So ODRL is now an official W3C standard and there is a certain profile of it, meaning like a subset or an application of it in the news industry where the main standards body is IPTC, which I believe stands for International Press Telecommunications Council called Wright's ML. So Wright's ML, first of all, this is a sort of roughly equivalent diagram to the diagram I showed previously about ODRL 1.1. This is the ODRL 2.2 information model and it shows you that you've got assets, you've got parties that are looking for rights to those assets. Excuse me. You have rules that describe permissions and duties and prohibitions. And you can sort of look at this and get an idea of how ODRL works. And then again, I've included a couple of pieces of example code. This is JSON code. And this first example is you can only play this movie in Germany after 2018. And you can look at this code in more detail later. Here's another one that says you can distribute this movie but only over a mobile network as opposed to a home broadband or what have you. So Wright's ML is in use today by the Associated Press and its affiliates, among others. And there's a Wright's ML website which has got some very nice sample code. This is one of the samples. This is similar to what the ODRL example that I showed. This picture can be displayed by the European Publishing Association licensee within Germany. And here we show it as a Wright's ML data structure template which is just a vanilla way of showing things like what permissions are being assigned, who is assigning them, to whom there's an assigner or a signee. What permission is it? It's the action of distribution. What are the constraints on the distribution? Spatial constraint that is Germany, D-E-U being a code for Germany. And then on the next slide we show the same example code in JSON. And they also offer it in XML-RDF, various flavors of XML-RDF. Oh, I want to back up and mention, let's see, where did I put this? Sorry, in the footnotes of the ODRL slide, a few slides ago, it mentions a slide share presentation from Stuart Miles of the Associated Press, which in my view is the best presentation I've ever seen on the practical use of a Wright's language for license automation. It shows you the practical value of it, how the Associated Press is really using this thing. And it's very, very informative. We want to take a look at that. So now we're talking about, we're on slides, whatever number this is. Wright, here are a few other examples of license automation and Wright's expression languages that are in use. One is the picture licensing universal system or plus Wright's language. This is a perfect example of reverse engineered acronym in the license image community, the stock image community. Prism, which is a standard that has to do with content that's nominally in the consumer magazine world, has a Wright's language. Prism is another reverse engineered acronym that I can't even remember what it stands for, publishing requirements for industry standard metadata. There you go. Metz Wright is a library world standard from the world government and academic libraries for communicating Wright's information about content among libraries. And then probably the most recent one is OxPL permissions language, which has to do with an onyx as a set of metadata for books, including physical books. And in fact, started out as a physical book metadata standard for things like how big is this book, how many come in a box, how many pages does it have and so on. And there are footnotes here with the URLs for all of these standards if you'd like to take a look at them. So let's step back a bit and talk about the relationships among these various things that we've been talking about. Wright's expression language is ERM licenses and copyright. And there's a lot that can be said about this, but here are a few points that I'd like to make that I'd like everyone to think about. Maybe we can discuss some of these things when we meet next week. Licenses are of course contracts. They are legally enforced rights and restrictions. ERM is a technologically enforced rights and restriction mechanism. The purpose of a RHEL is to enable precision in describing rights in both licenses and ERM systems. And RHELs can encode licenses that confer more rights than you get in a copyright bundle that's the basic idea of Creative Commons. What additional rights would you like to offer to the public besides what the copyright law tells you? They can also be used to describe licenses that confer less or more restrictive rights than copyright. And if you were to look at most end user license agreements for things like Amazon, Apple, whatnot, that will be the case. If you were to encode those EULA terms in a RHEL, they would be more restrictive than what we're typically. There are exceptions, but in general, these EULAs are more restrictive than what you would get with copyright if they have to do with content. And then once again, if we're talking about pure digital content that is not attached to a storage medium like a DVD, CD, some drive, et cetera, then copyright doesn't really apply. It's all about licensing. And this was just affirmed very recently in an appeals court decision written by none other than Judge Pierre Laval of the Second Circuit Transformative Use that involved a startup called BDG located across the river in the Prudential Center in Boston that was trying to build a marketplace for used digital music files. Fascinating subject, which I'd be delighted to discuss offline, very interesting case, very interesting startup and so on. And then the final thing that I'd like to mention in terms of what do these things have to do with each other, RELs, DRMs, licenses and copyright. You can have a REL and you can use it with a DRM but it won't make DRM capable of emulating copyright rights. It just, people talk about how RELs, sorry, about how DRMs are restrictive compared to copyright. I prefer to say that they're just not capable of emulating copyright with any degree of precision. And I'd like to amplify that point by talking about something that really didn't get enough notoriety in my opinion, something called the Digital Media Project was a standards initiative that was started in 2003 by a man named Leonardo Carri-Leone who is the founder of MPEG. So this is someone who has a great deal of experience in the tech standards area. He was trying to create an open interoperable DRM standard that respected and emulated real-world content usages including copyright law and all of its ramifications in multiple geographies around the world. This is an extremely, extremely ambitious project. And that's the most interesting thing to map out that project, which technically is still alive, it's kind of on low simmer. So something that they tried to do called traditional rights and usages or TRUs and this is actually a link which I'm not gonna show you now, but Dazza will have the URL and can share it. They did an exhaustive or maybe not exhaustive but prodigious amount of work in trying to describe precisely various different traditional rights and usages around content in different geographies around the world including fair use, fair dealing and things of that nature to try to catalog these and be able to represent them in machine readable ways. This would have been a very, very interesting project if explored to its limits, whatever they may have turned out to be, but it was never really completed. Anyway, that just sort of shows you what the limitations are with regard to Rails and the real world. I want to finish this talk by tying this to the Music Modernization Act, which we're hovering in this course and talk about rights expression languages could relate to the areas around Music Modernization Act. So I'm going to go over some material which is probably redundant to what other people are talking about, people like Kiki Nelton, George Howard and so on, but at the risk of repeating these people, I'll give you at least my version of the background of the Music Modernization Act. So there is a section of the copyright act called section 115, which establishes a compulsory license for something called composition mechanicals. I think I have skipped a slide. Sorry, yes, we're going to go back and talk about music rights and services. Yes, this little, sorry, I skipped that one. So in order to understand the Music Modernization Act, we need to understand a little bit about the basics of copyrighted music. And once again, other presenters may go over this. So at the risk of being repetitive, I'm going to go over this. Every music recording has two copyrights. There's a copyright on the composition, which you can think of as the cheat music with lyrics. And then there's the copyright on the recorded performance recorded by a performing artist or recording artist. And when you offer, when you have a digital music service that offers digital music, you need to have licenses for both things for the compositions and the sound recordings. So today's digital music services are what we call DSPs, digital service providers. What they offer is basically three things. They offer interactive streaming. We're talking about Spotify, Apple Music, Tidal, Deezer, Rhapsody, et cetera, et cetera. Their active streams means you, the user, choose whatever you want to listen to and then stream to you. There's something called conditional downloads, also known as tether downloads, which are you can download the music to your device and play it off your device until your subscription is canceled. So that's why it's conditional. And then there are permanent downloads where a user has to pay, let's say 99 cents a track to keep the download permanently. And typically that's DRM free. It's an MP3 or an MP4 EAC file. So these services have to pay a few different royalties. And again, this is a typical case. There are exceptions. There are no constraints to this. But they have to pay reproduction and distribution royalties to record labels for sound recordings. They have to pay something called a mechanical, which is a reproduction and distribution royalty, just conventionally called a mechanical for the composition, not the recording, but the composition to the music publisher. And that is a compulsory license under section 115 of the Copyright Act, which says that anyone can make a reproduction of the composition, which normally were in older times would mean do a cover version of that composition. That's guaranteed to you by law, but you do have to pay a royalty for that. You need to fulfill certain other conditions, mainly paperwork type conditions. And then there is something called public performance royalties that you pay to, again, in the typical case, to an entity called a pro, a PRO, such as ASCAP or BMI. And there is a couple other ones in the United States. And this is very much a national scale thing. And in some cases, you'll pay directly to a music publisher instead of to the PRO in the case of a bigger publisher, such as one of the major publishers. Okay, so now we're gonna talk about the Music Modernization Act, which was just enacted recently. And the problem with section 115, the compulsory license, is that it is not also a blanket license. You have to license each composition a track at a time. And what happens is, if you're a music service like Apple Music or Spotify, you get feeds of tens of thousands of tracks per day from record companies, from companies that aggregate indie artists and so on. And for each of those, you have to go find composition that's being played, which is not obvious in some cases. And then you have to secure a license for that composition, fulfilling what amount to paperwork requirements under section 115. And that is enormous pain in the neck because not only is it a lot of just paperwork, but there's also ambiguity about what composition is, because you may not necessarily get the information about what that composition is when you get a feed from a record company. There's data floating around the industry that's old, that's incorrect, that's missing, it's actually maybe even be disputed. And this is basically a huge burden on these music service providers. And it's led to lawsuits where this hasn't been, or allegedly hasn't been done correctly. Spotify has been sued, Apple has been sued. So the legislative solution that was enacted last year was one that does essentially two things. One is that it finally implements a blanket license instead of requiring track by track licensing is something that was attempted several years ago, but it didn't pass through Congress. And then the other thing that it does which is a little bit more controversial is it creates a single agency to process these mechanical royalties for every DSP instead of each DSP doing the job itself or more typically hiring a third party agency such as the Harry Fox agency to do that for them. There's also a clause in the act that's somewhat controversial that composition rights holders can't sue DSPs anymore for alleged lack of improper licensing, but we're not gonna talk about that, but it's got little to do with rights expression languages. So the license structure isn't really affected by the Music Modernization Act, but it's worth talking about anyway because the Music Modernization Act sort of wins out the benefit and encourages development of automation mechanisms for automating these licensing requirements. So let's look at what the license is. It's some rights that the publisher is granting to a digital service provider. It's a blanket statutory mechanical license. The performance license is still necessary, but it's not affected. This is independent performance right. And so the question becomes, how do you represent that in a license, in a, sorry, in a rights expression language? And one thing that exists already that can be drawn on is something called DDEX, which is a standard, it stands for Digital Data Exchange, a music industry standard for communicating information about sound recordings that is used in cases like a record label has just released an album and they wanna tell music services about it so they send a message in this format to the music services. And part of this message's impact is a set of codes for the rights that the record label is granting to the digital music service, which include on-demand stream, conditional download, and permanent download, and a bunch of others. But those are the three relevant ones that you can take from the DDEX standard as the rights that you are dealing with in a potential rights expression language implementation of these rights for the Music Modernization Act or for mechanical licensing. So here's an interesting idea. You can think of this as possibly being governed by two systems. One is a license automation system that governs the activity between the publisher, the music publisher, which is the blue box on the left, and the DSP, which is the orange box in the center. And then that in turn will indicate what happens between the DSP and users on the right in green. So the publisher, let's say, says, I'm granting you the DSP. We'll give some examples. I'm Warner Chapel Music. I'm granting Apple Music these rights, on-demand stream, permanent download, and conditional download. Then the DSP uses those rights in another automation system, which is a DRM system. And by the way, all these DSPs now use DRM for streaming and conditional downloads to govern what happens with the user's music on the user's devices. And there'll be a right for on-demand streaming that's pretty straightforward. The problem is that for this to be automatable with the rights expression language, the idea of conditional download needs to be made more specific than it is now. Now it just says, okay, conditional download. We need to know what does that mean? Okay, that means download how many times? Download until when? Play until when? Play how many times? You need to make that more specific in order for it to be automatable. So that's just sort of an interesting limitation on how these rights map to the real world implementations that you would want to make with the rights expression language. So I'm going to conclude by offering an exercise, which kind of ties us all together. And the exercise is to represent the Music Modernization Act relevant rights in a rights expression language type syntax. And the first one of these is the simplest, which is just to represent them as a rights ML style data structure, just to show who the entities are, who the license or licensees are, what kind of rights are being granted, what are the parameters of those rights and so on. For extra credit, you can implement it in ODRL using the syntax of your choice, JSON or one of the RDF syntaxes. And then for really extra credit, and this would be an industry needle moving kind of thing, represent all rights for interactive music DSPs as an ODRL profile. In other words, do what rights ML has done for the news industry for music. DDEX doesn't get us there. It's not really a rights expression language, even though it has some rights, metadata and code is included. A very viable basis for communicating rights among business entities in the music industry would be to take an ODRL, to take and create an ODRL profile for music rights. So that's your extra extra credit exercise created that covers streaming conditional downloads, permanent downloads, and then all the composition and sound recording rights that you require for these digital music services. So with that, I'm done. And I would like to say thank you very much for your attention and thank you very, very much to DASA for inviting me to do this because this is clearly a topic that's near and dear to my heart and I like talking about it and I hope that you enjoy learning about it. Well, thank you, Bill. That was extraordinary. Little technical snafus in the beginning with your presentation. So I... DASA, if you are talking, I can't hear you. And advance the slides. The slides are, of course, available on your session page. Can I just ask, Bill, as the students come up with their takes on the exercise, would you be willing to kind of give a little bit of feedback as we get toward the class so we can start with a bang? Absolutely. Okay, so then we'll take responsibility to package those up and share them with you. And I really look forward to hosting you in person at MIT on the first day of our computational law course this year, which is Tuesday, January 15th. And everybody out there in internet land would give a big round of applause for Bill doing the first proper academic presentation I'm familiar with, anyway, on rights expression languages in this context. It's much needed, overdue, and I'm very grateful for it. Thank you. Okay, thank you.