ken@gvax.cs.cornell.edu (Ken Birman) (07/17/90)
Well, I don't know if ISIS and Mach are in the same category, but DOC recently relaxed its rules for export of systems like ISIS. On the phone, I was told that ISIS could now be exported without any restrictions at all. I am waiting to see this on paper, but it seems likely that the status of Mach will turn out to have changed (for the better) too. Note that Mach and ISIS are funded by the same program director (Brian Boesch at DARPA/ISTO). In fact, one would expect ISIS to be the more restricted technology of the two, since "fault tolerance" has a military ring to it... It is not clear to me that DOC would need to "discuss this with DARPA" unless DARPA feels that the technology should be classified or something. Certainly, there has never been anything in an ISIS funding contract compelling us to seek permission before exporting the system, and in fact DARPA has been one of the strongest forces in pushing for technology transition for both ISIS and Mach. Of course, the Mach contract could be worded differently, or DARPA/ISTO may have specifically asked Rick to talk closely with them on this issue. They have not done so with me, although we certainly keep Brian Boesch informed. Note that these days, DARPA doesn't even hold copyright on software like Mach or ISIS, or insist that the systems be public domain. Inquiries on export classifications are handled by the "ECCN classifications" office within the Department of Commerce: OTPA (ECCN Classifications), Po. Box 273, USDC, Washington DC. 20044. You need to send a cover letter describing your product or system, indicating especially if it is "freely available" (even with a fee; this means "on an equal basis for all who request it") in the United States, and include documentation giving the technical specifications of the system, like a user's manual. You can also talk to these people in person, as I did, but its better to have a response on paper. -- Ken Birman
mellman@motcid.UUCP (Tom Mellman) (07/19/90)
In article <43447@cornell.UUCP> ken@cs.cornell.edu (Ken Birman) writes: >Note that these days, DARPA doesn't even hold copyright on software >like Mach or ISIS, or insist that the systems be public domain. >-- Ken Birman Do I understand that correctly? Are you saying that projects that the public pays for can be copyrighted by private individuals for profit? Sounds like Reagan-era shanangans to me. That would call for a letter to my representative (for whatever that's worth). Thomas Mellman -- Reply-to: motcid!mellman@uunet.uu.net
ken@gvax.cs.cornell.edu (Ken Birman) (07/19/90)
In article <4195@graphite18.UUCP> motcid!mellman@uunet.uu.net writes: >Do I understand that correctly? Are you saying that projects that the >public pays for can be copyrighted by private individuals for profit? >Sounds like Reagan-era shanangans to me. That would call for a letter >to my representative (for whatever that's worth). Right. NSF and DARPA are no longer allowed to force software into the public domain -- nor do they hold copyrights on books or papers you write under a grant. Similarly, an artist retains rights to art produced under public funding. Universities sometimes set things up to hold the rights themselves, but Cornell does not do this unless you make unusual use of University resources (i.e. a gene database or something). Basically, a computer program is being treated as an intellectual work that can be protected by copyright just as for most other such works. I'm not a lawyer, but I think I have this straight. We are, however, required to disclose our software or whatever -- but this is not the same as not holding copyright on it. That is, I am required to permit you to have a copy of, say, ISIS in the state it was in at the time I went off NSF or DARPA funding, but because I would hold the copyright I can also limit, under copyright law, the uses to which you can put the system (or reproduce it, or modify it). Technically, I think the situation is that we have to disclose our work to the government, which must disclose it to you under the Freedom of Information Act. So, you can get at a copy. But, this doesn't change the copyright status of that copy, and so you are subject to any limitations I might want to put on it. Obviously, NSF and DARPA are not required to renew funding when they are unhappy about what a project ends up doing. However, both organizations tend to want to see "technology transition" plans these days. Another popular program is called SBIR; under this, all Federal funding agencies reserve of their funding to help small businesses develop products. You can get up to $550k specifically to develop a product -- of public money. Although I see that this bothers you, I guess that it bothers me less than to read about huge cost overruns on multi-billion dollar secret projects that were contracted out under sole-source contracts and never produce a single spin-off that might benefit the US economy or impact our day to day working environment... that stuff is usually not very public either, although I have no idea who gets the copyright on the software involved. My guess is that you haven't really given a lot of thought to this whole issue, since the current public policy is actually fairly consistent on intellectual property issues and also pretty reasonable. And, realistically, the government can't pay for development, distribution and support for large software systems indefinitely. Letting someone make money by extending and supporting something that originated through publically supported research makes a lot of sense, as long as the system is publically available as of the end of the funding in a way that lets others learn from what has been done... (By the way, we give away ISIS, in source form, free. And, Rick gives away the Mach kernel, also in source form, free. So, even if we aren't compelled to do so, both groups are certainly making their software accessible. However, the ISIS group has started to release the system under a copyright notice to protect our rights in the event that we might elect to develop a commercial version of the system later.) Ken Birman