This next presentation was by Daniel Carchidi and Lindsaey Weeramuni, both at MIT, talking about copyright and how the Open Course Ware system deals with it. There were two parts to this. First was an overview of copyright in general, and the second focused on OCW specifically. My impression listen to this, and which was confirmed later during the QA period, was that OCW takes a very conservative approach to copyright. This is quite intentional on their part. It reflects the priorities of MIT for that particular project, and not any position on copyright in general. [My thoughts and comments are enclosed in square brackets.]

Most of the about 30 people in the room had heard of or visited the MIT Open Courseware site. [an interesting example of the marketing value of sharing content - in theory the contributors have given for free some content on which  in theory they could maybe have made some money, but in practice both the individual providers and the institution have gotten a great deal of good publicity out of it. I don't know the finances well enough to know if the value are on par with each other, but my off-hand guess is that the publicity is worth as much or more than the content itself.]

Dan leads off the presentation with an overview of Open Courseware. In 1999 there was wide-spread discussion [references would be nice here - sounds familiar, but actual links would be even better] that distance learning would make traditional, classroom education obsolete. In response to that, the MIT college president formed a committee to examine the situation. The committee produced two main results. First, they felt that MIT had missed the wave in terms of commercial exploitation of distance learning. Second, and perhaps more importantly, they concluded that the MIT experience didn’t transfer to distance learning. That is, there was value to the experience of being at MIT that could not be provided via an on-line experience. The president then asked, essentially, “what now?”, and the committee went back to work. After further consideration, they decided that the MIT course materials should be provided for free on the internet [I would love more details about this process - What objections were raised and how were they countered? What were the perceived costs and benefits, and how has reality matched? What lead them in that direction to begin with? etc.] With that, MIT started the Open Courseware project, with a goal of publishing material for all their core courses.

Some useful clarifications / re-iterations:
- OCW is NOT an MIT education. It’s missing the experience in the classroom and lab, and the presence of teachers and fellow students
- OCW is NOT a certificate or degree program
- OCW is permanent [at least, as much as any info is]
- OCW is open to the world
- OCW contains all course material (lecture notes, lab descriptions, problem sets, tests, etc.)

There is a LOT of material in OCW now. [I'd last looked through it several years ago and found it a little sparse, though impressive even so. Now it is truly amazing.] Just last fall they hit their publication goal/milestone of having every core course in OCW.

At this point Lindsey stepped in for Dan and talked about the intellectual property issues of OCW. Prior to her current position at MIT, she worked in the publishing industry, so she has a good perspective from both sides (the side asking to use material, and the side that controls the material).

OCW uses one of the Creative Commons licenses for all their material. She didn’t spend too much time on Creative Commons itself - that topic is easily an entire presentation in itself [It's definitely worth spending some time reading the Creative Commons site. It's an increasingly widely used system for licensing works, and it's useful to understand the various options and what it can and can't do for you.]. MIT uses the non-commercial, attribution, share-alike (by-nc-sa) options for their license. In brief, people are ENCOURAGED to use, modify, re-mix, etc. the content, as long as they’re not selling the results, as long as they refer to MIT as the originator, and as long as they like-wise share their modifications if they publish them. [That's my pithy paraphrase of her high-level summary - if you actually want to use Creative Commons you must read the actual license. There are some complexities and specific, legal wordings that are completely glossed over here. It's also worth noting that Creative Commons is not something that circumvents copyright or replaces copyright. Instead it relies existing copyright law to create licenses that let people use material is specific ways, just like any other license. It just happens that the specific ways are very unrestrictive compared to most other licenses traditionally used in the publishing world.]

MIT has had a lot of success moving protected material [highly restricted] into the Create Commons environment. After they hit their initial publication goal they did a comprehensive review of their material. They found that about 1/3 of their courses still had some copyright issues (by the restrictive standards MIT is using for OCW - more on that later). Most of those problems were in the lecture notes for a course - that it, the material that the instructor would use when teaching (a more detailed breakdown is available in their slides). Conversely, 2/3 of their courses were completely OK. An important conclusion of that review is that it’s possible to have a very rich repository of publicly available material that respects copyright.

Then Dan gave a bit of Copyright 101. [for full details on the law, check out the law itself] Copyright protects works of authorship (see the slideshow), i.e. ‘the creative product’. In their work with faculty, these are the lecture notes, problem sets, tests, etc. Things that are NOT covered include ideas, concepts [patents cover those two], facts, raw data [including research data, I guess; explains why researchers are often a bit paranoid about sharing / publishing their source data in addition to their results / conclusions], names, and titles. [To protect the latter sorts of things, look into trademarking - it may or may not apply, but copyright certainly won't.]

The people who get the protection [copyright is, at least originally, about protecting a publisher from economic losses, though these days it's cast as more about controlling how content is used - the two subjects are related, but they are by no means the same] are the author, or the employer in a work-for-hire situation (where the contract assigns the rights to the employer), or anyone to whom either of the first two have assigned rights. A copyright holder has exclusive rights to reproduce, to distribute, to display, to create derivative works, AND to grant licenses to others. In OCW the creators still hold copyright to their work, but grant license to MIT and the OCW system (via the Creative Commons  licensing tool set). Among other things, the law means that works are NOT freely available UNLESS it is EXPLICITLY in allowed - that is  failure to claim copyright explicitly does not give up rights, and simply citing the creator is NOT enough without the copyright holders permission.

Copyright is an evolving set of laws. Since 1976, any tangible publication [how does electronic publication fit into that? I assume it's covered in some particular clause of the law, but I'm not willing to dig through and find it just at the moment] automatically has copyright protection. Prior to that the work had to be registered with the Library of Congress. Copyright has an expiration [at least in theory - in 1998 it was extended for 20 years] and works published before 1923 are in the public domain. This set of older publish material is drawn upon heavily by history courses, and to a lesser extent in goverment courses.

The was definitely some early resistance by faculty to contributing material to OCW, but over time people have become enthusiastic about it. About 90% of MIT faculty participate by contributing material to the project. Students also contribute. There was a nice quote of a student about how pleased she was to have her work published in the OCW system. [How does OCW publish count, professionally speaking? presumably it's less valuable than a journal publication, but at the very least it's publicity. I imagine hiring and tenure committees will have to tackle this question and make some kind of policy at some point. Some discussion in the NERCOMP 2008 blog also touches on this idea]

MIT threw a lot of resources at the OCW project. It’s very team oriented. There are OCW team (6 people) who deal with manging the program, the publication managers and the dept liaisons (11 people) who work with faculty, a production team (4 people), and an intellectual property team (2 people) that makes sure the published material adheres to OCW’s policy. [numbers cone from the OCW team web site] Plus they consult a fair amount with the general counsel, and have two sizable advisory boards. In addition to the human resources, it’s clear that the institution embraces this project at the highest level, which makes it possible on a cultural as well as a technical level.

There was a nice few slides covering where and what kind of IP issues they encountered in their review of the repository. The most common areas of infringement are in the lecture notes, and the most commons types of infringement are images, text excerpts, and multi-media (clips, simulations). There are potentially very difficult issues in resolving those infringements. For example, much open source software is licensed using the GNU General Public License (GPL), which is incompatible with the by-nc-sa Creative Commons license that OCW uses [this is a fairly widely discussed issue - in brief, the GPL allows commercial uses, while the CC non-commercial, well, does not]. There are three main approaches they use to resolve IP issues (in no particualr order):
- remove the offending material and replace it with a specially commissioned replacement (MIT has graphic artists on hire for this kind of work)
- negotiate for usage rights with the copyright holder
- ??? [see the slideshow (coming soon...) - the talk was very fast at this point and I didn't catch the last method]
One general rule is that if material can’t be sourced, it can’t be used.

There were quite a few questions from the audience:
Q - You deployed students to take notes in the courses. Is this typical?
A - It depends on what already exists. If there are no lecture notes then the students’ notes form the core.

Q - How do you handle publishing links?
A - Published links are only to publicly available systems. We can put plain text addresses [URLs] which users can go to and/or research on their own, but we don’t do direct links to non-public systems.

Q - How do you deal with custom programmed materials (which don’t work too well with CC)?
A - If it can’t be released using a CC license, it’s not released.

Q - What the background of the Intellectual Property managers who are making the decisions?
A - Lindsey has extensive experience in the commercial publishing world. Her assistance has an art background, but has received IP tutoring.

Q - How much of the repository is actually captured lectures?
A - It depends on the courses. There are about 1K hours of video, ~30 full courses. The focus has been on curriculum, not actual presentation.

Q - How do you deal with text books being posted?
A - Most of the time, you can’t. At best, you can put a link on the reading list section, with links to amazon or the publisher so users can buy it themselves.

Q - Any thoughts on leveraging social systems to adapt and modify the OCW system?
A - There’s now a Open Courseware consortium that contains a lot of institutions world wide. Now that MIT has met their base goal, it’s thinking more about what to do next. E.g. making high school focused adaptations, with feedback from HS teachers.

Q - Has the OCW movement changed the way copyright holders view their IP at all?
A - I’m glad you asked. We just a few days ago made a partnership with Elsevier, one of the big names in journal publication. We got permission to use data from any of their journal articles, with particular limits.

Q - What about fair use, especially with respect to Elsevier material (the negotiated deal sounded like it was all pretty much covered by fair use)? Do you never invoke fair use?
A - It’s a hot topic among consortium members. From the beginning we made a decision NOT to rely on fair use. We decided to be especially conservative because of the high profile - we didn’t want to expose MIT to any negative repercussions. We think fair use is important and valid, but not quite appropriate for OCW. In OCW we’re not just using the material ourselves, we’re publishing it for others to use. [This is a really rich and interesting topic. It would have been great for MIT to push this a bit, but I can totally understand why they didn't. I hope that now the MIT has paved the way with OCW in general that some other group tries to do more with fair use, because if we don't exercise it, we could lose it]

Q - What does ‘maintain’ a course mean?
A - Because of the structure of the publication process, we have a relationship with the professors: we know if their teaching a new course, what changes have been made, etc. Very relationship based.
Q - A follow up, if a course is no longer active, is it removed?
A - They’re kept in the archive.

Q - What about ITunes?
A - it’s another distribution channel. A big part of our audience (of a certain age) is introduced to professors via a lecture on iTunes.

Q - What are you mechanisms for dealing with infringement at MIT (outside of OCW)?
A - OCW is not involved. The general counsel handles it. Hopefully the OCW experience provides cultural knowledge/precedent.

Q - This presentation has been pretty black and white. What/where is the ambiguity?
A - A lot is actually case dependent, but we couldn’t get into the details in time we had

Q - How did you convince faculty to participate?
A - Shame :) Some departments resisted in entirety, at least at the beginning. Over time as they’ve seen success and faculty benefit they’ve come round. There were lots of objections (e.g. no one will come to class), but we’ve generally found the experience did not bear out. There have been some tricky cases involving publishing (e.g. textbook) vs open release (would one be in competition from other). [this has interesting parallels to objections to lecture capture at Williams - hopefully we'll eventually see success too]

Q - What are some benefits to the faculty?
A Lots of feedback from the users. Lots of citations in other publications for things published in OCW. Lots of visibility and publicity. Great organized and recorded courses. It counts as dissemination for NSF grants. Students use it to shop for courses. Students use it for review notes. They can see what their peers are doing. It improves teaching.

And then time was up. I think this discussion could have gone on a while, and I’d love to hear more about / from the OCW Consortium at some point.

I had a chance yesterday afternoon to hear Dr. Chris Kelty talk a bit about peer review processes and how they interact with internet publication. He approached this in the context of open educational resources (OER) in general, and Connexions specifically.

In brief, OER refers to educational materials (lesson plans, lecture notes, examples, labs, etc.) that have been explicitly placed in the public domain. OER is a fascinating topic in it’s own right. A simple Google search gives lots of decent places to learn more about it, and the Capetown Declaration is also worth reading if you’re interested in that sort of thing.

Peer review enters the discussion when trying to validate the resources being shared. The traditional publishing method (send the paper to an academic in the same field) breaks down in the world of internet publishing. The first problem is sheer volume - there’s more material being produced than there are qualified reviewers. The second problem is context - traditionally an article reviewed for a journal would be published in a journal (where it works well) and not a textbook (where it would be bad), but an article published on the internet can be accessed devoid of context.

Kelty mentioned some existing internet systems that deal with similar problems - Wikipedia’s content review/development, EBay’s trust system, and Amazon’s reviews. These crowd-sourced systems work reasonably well, but they have no way of leveraging existing expertise. Any internet review system for academic work should be able to take advantage of the expertise in the traditional system.

Connexions takes an interesting approach that pulls some ideas from various areas:

• from open source software : versioning and forking and public licensing (though Connexions uses a Creative Commons license rather than a software oriented license)
• from Wikipedia : the idea that anyone can contribute and that contributions are unfiltered. That also arises in part from DMCA regulations, in which an editor is responsible for content while a service provider has some protections
• from Amazon, the idea of reviews as a separate object with their own meta-data (pseudonymous author, rating, context, etc.)
• from EBay, the idea of trusted entities

The result is a system where any one can contribute to the repository and anyone wanting to use the repository can filter it based on which reviewers they trust in the context in which they want to use the content. It’s designed specifically for OER content, but I don’t see any reason the same principles couldn’t be applied to a system for journal-style content as well.

It’ll be very interesting to see where this all goes. I’m especially curious to see how Connexions will deal with an information hostile environment where users deliberately try to add wrong information or skewed information, and to see if and how people try to game the system.

I’m in the midst of reading Botany of Desire by Michael Pollan and just this morning over tea was reading about the New Leaf potato, corporate farming, mono-culture, and the farming practice of saving seeds to plant the following year. The potato is a chimera that has a Bt gene in it to create a toxin to repel and kill insects. For a variety of reasons I’m glad that project has been shut down. But, this post isn’t really about potatoes.

In one of the paragraphs where Pollan talks about the culture of mono-culture and the practice and business of farming he mentions the practice of brownbagging seed. Brownbagging is saving seed from a harvest to plant the next season. From the farmer’s standpoint it makes perfect sense, both economically (not spending money on seed is good) and ecologically (developing locally adapted varieties). From a corporate standpoint it’s a disaster. Imagine a company that sells a thing, only once you’ve sold it the person who bought it can easily make as many copies as they want for free…. hmmm…. sounds familiar… Seed companies and data companies (software, music, etc.) are in much the same boat. Once they’ve sold their product the purchaser can make and distribute (practically, if not legally) free copies of that product.

The development of hybrid crops with don’t ‘come true’ from seed make agribusiness corporation happy - if farmers want the super-sweet, super early sweet corn they have to buy the seeds fresh from the company every year, because the previous seasons seeds won’t produce plants like their progenitors. A more recent GMO innovation was the termination gene, where the modified plant won’t produce fertile seed at all, thus making sure that a farmer that want anything at all to grow, let alone that special variety, must buy the seed fresh each year. Agribusinesses have also had a number of court rulings in their favor in recent years, which make it illegal for proprietary varieties to be propagated (by seed or cutting) with out permission from the varieties’ owners. In many ways the seed industry is taking pages out of the data industry’s playbook. The various work to make sure saved seeds aren’t useful is akin to the copy-protection schemes the software industry has developed, and the anti-piracy and IP laws of the data world are very similar to the ones in the agricultural world (in effect, if not wording).

However, the most interesting point to my mind is that for centuries prior to the current technologies there were still companies that made money selling seeds. They certainly weren’t as dominant as the modern corporations (e.g. Monsanto), but they pulled in enough to make a reasonable living. How did they make money selling something that people could get for free? Here I can only speculate, as I have to admit I don’t have the time or interest to do in depth research. First, I imagine they spent a lot of time finding, developing, and marketing new varieties. Second, they weren’t selling seed per se, but the service of dealing with the hassle of saving seed - a farmer has a limited amount of time and energy to run a farm, and time and resources spent on seed saving can’t be used elsewhere. Saving seeds is pretty easy, so the latter point can’t account for too much of the price of a seed, but it’s certainly a factor. Third, they sold to new farmers who didn’t yet have a seed store of their own. Of the three I imagine the first was the most important. Why did farmers buy the seed? Because they didn’t have that new variety, because they didn’t want to deal with seed saving on their own, or because they didn’t have any seed (because their own rotted or was eaten by rats, or because they’re starting a new crop, or farm, or whatever). Despite the imperfect analogy, this suggests to me that even in the absence of copy protection and anti-piracy laws, software companies, music companies, and any other companies in the business of selling particular arrangements of 1’s and 0’s could still make money. Enough to be world dominating hegemonies? Probably not. But enough to keep the owners and employees comfortable? Probably so.

This isn’t exactly a new revelation. Companies involved in open source have been using this model for a while. The software (seed, data, whatever) is essentially free, but the service of improving and delivering it, and the comfort of reliable support, lets these companies make money anyway.  The reason I wrote this up was not because the idea is so new, but because I was struck by the parallels between data-oriented and agricultural businesses. What is a seed if not a biological program for creating a plant, and with it more seeds? Dealing with the issues of a freely reproducible product is an older business problem than is usually thought, and by no means insurmountable. In the world of data, copying the product is even easier than with seeds, but in the business practices of yore there may a commercially viable approach that doesn’t rely on clunky copy protection and heavy-handed, difficult-to-enforce, and unpopular laws.

Fairly recently I wrote a couple of programs for / with some professors here at Williams to assist them with their research and teaching. Not too surprisingly they and I want to be able to share that work with others. If it were just a research paper that would be no problem - there are well established models for ownership, credit, and distribution. With software it gets trickier.

Academia has had a long and happy tradition of sharing resources and information (with great influence in that regard on Net culture). The initial informal sharing systems weren’t well suited to the business world and the internet as it’s grown, and thus arose a number of other schemes (the GPL and Creative Commons being probably the two biggest players these days). Both those seems to work pretty well for people who want to share their work. There are slight changes / additions / options I’d like to see for them (essentially, a protection period that’s slightly more restrictive, after which they revert to the more open licensing), but over all they make me quite happy.

However, they’re only useful if a person actually has the right to license something. In the business world typically whatever a developer creates is owned by the company. Academia doesn’t quite work that way: researchers typically own (IP-wise) the results of their work, professors write textbooks that are theirs and theirs alone, scientific papers are owned by the individual rather than the institution, etc. However, all that precedent is for professors (and students, for that matter). Staff are often placed in a separate category (I think on the theory that we’re paid to do (and sometimes create) things for the institution). I don’t think it’s quite fair, but that’s usually the way it is. So, if I’ve created a program on the job for a professor (or anyone else, for that matter), my best non-lawyer guess is that the college technically owns it. So what does that mean from a licensing perspective? Specifically, who has the right to release (partially or wholly) that program into the public domain?

For lots of academic programming this question never comes up - IT departments share code and technique all the time with out anyone worrying about it. There are two big differences in my situation. First, the audience / user-pool of the application is not other IT people - the sharing is not in the course of working on problems common to all parties but distribution to end users. Second, there’s at lease a remote possibility that this application would be bundled with a textbook, thus allowing a publisher to make money from it (in which case I’d certainly like some kind of compensation/cut - not a huge amount, but something would be nice). So, I talked to our CTO and explained that we had a potentially complicated IP law issue and asked her to contact the college lawyers about it.

I haven’t heard much back yet (it’s been about 6 months since the initial conversation - they’ve been busy with RIAA obnoxiousness), but here’s what I’m hoping will happen:

• technically the college owns the rights to the work (I’d like the creator to own it and while I think that’s possible, I think not likely), BUT
• the college assigns those rights to the developer / creator(s) as an incentive (akin to options and bonuses in the business world); the revenue from such a work is miniscule from the colleges perspective, but possibly significant to the individual, and revenue aside the credit counts for a lot
• regardless, the college has one or more licenses it endorses (probably GPL, CC, or some based on those), from which creators could select appropriate ones for their work

We’ll see what happens….