The Problem(s) with open-source law

Shawn, I think the real issue here is how closely to hew to the "open source" production model. Most of the difficulties here could be characterized as "difficulties in working collaboratively in law." From that point, the idea of "open source" is really only an inspiration rather than a plan of approach. We are asking how lawyers and laymen can collaborate on large scale to produce legal work product. We are hoping to achieve some of the advantages of production that anarchic approaches like free software have produced in other domains.

(One of the many advantages of "free software" over "open source" as notation for what we do is that it's much less subject to metaphor abuse. People come to me all the time with schemes for "open source" this and that, which need to be evaluated on their merits without regard to how computer software is made and distributed, or even without regard, as here, to the principled distinction between zero and non-zero marginal cost.)

Because legal activity is fact-dependent, and—as Jerome Frank pointed out so long ago—facts are uniquely combined and only partially discoverable in each individual social situation with which legal process deals, legal products have non-zero marginal cost. Even if they are merely bitstreams, the must be adapted somehow each time they are used.

So we are really asking here, in my opinion, not how law can be produced the way free software is produced, but rather how lawyers can collaborate so as to raise productivity substantially, that is, so as to make much more legal service available to more clients at prices they can afford to pay, while improving the quality of the services rendered in absolute terms.

I think this question does bear close relationship to what we've been talking about. It's more, however, like how to collaborate to produce major motion pictures without coercively controlling distribution. Lawyers need to be able to discover opportunities to collaborate with distant or previously-unknown colleagues, to evaluate the proposed collaborators involved in the opportunity, to jointly access and manage privileged material, to generate work product for deployment by primary counsel, and to construct billing systems that allocate and distribute shares of fees received. They also need to be able to think collaboratively, using the Web as a place to evolve rather than just communicate ideas.

As you will have seen by now, I also think this questions bears a close relationship to what we've been doing, too. I believe that this course is part of my effort to evolve the arrangements you're talking about. My 1L course, which you've also taken, is designed as the introduction to thinking like that kind of lawyer. Each of my other courses is designed to use technologies of collaboration to induce skills that contribute to being that kind of lawyer, intellectually. Colleagues like Conrad Johnson, Mary Zulack, and Jane Spinak are offering other courses and teaching within the existing program in ways that help with other aspects of training such lawyers. As more teachers are themselves brought to relearning how to practice law, their teaching too will change. Students, however, should both demand and become faster change themselves, because you need to master the existing (new, rapidly-changing) world in order to begin your careers. The thought-road you're on is the crucial one. Don't be decoyed by a metaphor.

No, Devin, the GPL is merely one set of terms available for software that's free to copy, modify, and share. Most free software licenses are "permissive," which means that they do not require modified versions to be redistributed under the same terms as the modifier received in the root stock she modified. This property of requiring downstream preservation of license terms, often called "copyleft," is the conceptual innovation of the GPL. Licenses that don't use copyleft, like the Apache Software License, or the BSD and MIT/X11 licenses, permit the downstream modified versions to be licensed however the modifier wishes. This is still FOSS. So are the hybrid models in between, like the Eclipse license, which can best be described as "permissive as to executables, copyleft as to source code." All have their place in the FOSS ecology. This might be a parallel to the discussion Shawn begins here, modulo the general warning about comparisons.

Also, neither widely-used version of the GPL requires distribution of source code to "anyone who asks." Under GPLv3, source code must be provided to anyone who has a copy of the executable, or binary, version of the code. It is permissible, though not usual, to require proof of possession of the binary. Under GPLv2, source may be provided only to persons provided the binary, if the source is provided "along with" the executable. Under GPLv2, executables can be provided without source code, provided there is a written offer for the source code included with the executable. In this instance, only, source code must be provided to anyone who asks. Conceptually the case is an edge condition, because distribution of source "along with" binaries is the preferred and most common means of complying with the license. But commercially it's very important, because product manufacturers whose products embed GPL'd software often do not feel inclined to put a CD of source code in the product box. They feel it confuses their customers. So they prefer to make a written offer, which leads to an obligation to produce source code to all requesters.