ppy/OOR-IPR-03_chat-transcript_20100930a_unedited.txt PeterYim: . Welcome to the Joint OOR-Ontolog-NCBO-CC-IAOA-OASIS "OpenOntologyRepository_IPR Policy and Issues" Panel Discussion (session-3) - Thu 30-Sep-2010 * Topic: "OOR-IPR session 3: discussion and consensus on licensing arrangements for the OOR Initiative, and positions we might take on related IPR issues" * Chair: Dr. LeoObrst (Co-convener, OOR; MITRE) * Panelists: (2HJA) o Mr. PeterYim (Co-convener, OOR; Ontolog; IAOA) - presenting the "Strawman" o Mr. MikeDean (Co-convener, OOR; Raytheon-BBN) - Commentary, as team lead for the OOR system development o Mr. BrucePerens (original author of the "Open Source Definition") - Commentary o Mr. JamesBryceClark - (General Counsel, OASIS) - Commentary o Mr. JohnWilbanks (VP of Science, Creative Commons) - Commentary (invited) Please refer to details on the session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2010_09_30 . anonymous morphed into YuriyMilov anonymous morphed into JamieClark JamieClark: would a condition that your theme music be played, before you speak, violate a requirement that your contribution be unconditional? BrucePerens: We need to have a discussion of "license combinatorics". JamieClark: amen JamieClark: "star 3 to un-mute, star 4 to re-mute." No need to re-test, you heard me already. BrucePerens: Default to BSD is OK. Having a large combination of other licenses in the mix is _not_ OK for reasons I will explain. anonymous1 morphed into Matt Liebenson1 anonymous1 morphed into ElizabethF BrucePerens: I say stuff like this all the time. Unfortunately I can't get either the Supreme Court or Congress to listen, nor even the Patent office. BrucePerens: No. Well, there are some limited possibilities. BrucePerens: Nor whether Ontology is a copyrightable nor patentable art. alex: should all ontologies adopt a licence? I mean as soon as a reposotory is part of the federation should those ontologies being hosted by those repositories part of the federation adopt one licecing schema? JamieClark: I actually think ontologies are useful as an instance of proof that the existing IP legal regime has some holes and uncertainties. So a facile statement that ontologies are, or or not, covered, seems of little value. JamieClark: ... Simple-minded lawyer's version of "license combinatorics": Tragically, not everything works with everything else. No matter how much you wish it so. See, for example, chart: http://www.gnu.org/licenses/quick-guide-gplv3-compatibility.png alex: not all the ontologies being hosted by swoogle have the same licencing schema JamieClark: Bruce sez (if I understand correctly): FreeBSD implicitly includes a patent license due to the patent exhausting doctrine, but doesn't hit the hot button of an exp[licit patent grant. BrucePerens: Right. It's the doctrine of patent exhaustion, and the fact that the license does not explicitly say it only grants copyright rights. BrucePerens: The problem, Jamie, is that Ontology is executed. alex: sourceforge.net makes iteasy for people to use their repository by provieding several licences for the prjs they host BrucePerens: But not all of the stuff on SourceForge is meant to be combined. We will be combining ontologies. Open Source projects that are meant to be combining pieces arrive at a common set of compatible licenses. PeterYim: @Jamie and All (ref. Jamie's post above about mute/unmute) - its "*3" on your phone to unmute and "*2" on put your phone on mute again (i.e. not "*4") alex: the networking nature of ontos is a problem alex: how do people licence libraries of software. when using ontologies as modules they are similar to libraries BrucePerens: Library licenses allow for more combination than non-library licenses. LGPL is an example. GPL terms but combine it with anything. BrucePerens: Non-assert is not as safe as a license. JamieClark: Bruce and I aren't disagreeing here: any contributor who thinks their stuff IS executable (or bears similar use characteristics) ought to have the ability to involve a license that (like BSD) gives them as-is-whereis-waiver protection. BrucePerens: Jamie, is the contributor the one who needs to determine that, or the person who might have to process the ontology? JamieClark: On "Nonassertion covenants", here is a writeup: http://www.iphandbook.org/handbook/ch07/p06/ AdamPease: Does current thinking support allowing LGPL code as part of the OOR? JamieClark: And here is an early (2005) nonassertion instance: http://www.oasis-open.org/committees/office/ipr.php JamieClark: @Bruce: well, if we let people who contribute have a choice, the decision's made by each of them at design-time. Then at run-time, each user decides whether the aggregated rights on offer are good enough for their planned use. (Or they let some filtering function like an IPR rule substitute for their own judgment.) BrucePerens: That's why I'd rather grant the right by default. Why make things harder for the user? alex: which is run-time for ontologies? is it run-time for the software using the ontos? Stuart Turner: Are the licenses mentioned so far those of interest or desired or are they in fact the existing licenses of all software, including libraries, in the current OOR technology stack? BrucePerens: We can't hear you Adam. *3 more than once. BrucePerens: Usual $12,000 charge for me to explain this to the customer's lawyer, and then the lawyer charges that much too. Stuart Turner: We had Palamida perform a IP analysis on a set of proprietary code in order for us to remove IP restrictions and replace components with compatible open source libraries. These services were gifted, but we had the impression the costs were at least what Bruce mentioned. BrucePerens: That's right. Include a GPL component in software and you have to treat the entire work as GPL until you've removed that component. Which I generally like for my software but it might not work for you. BrucePerens: It's a question of "contribute" vs. "interoperate". JamieClark: raising my hand again BrucePerens: Viral is pejorative. PeterYim: Sorry, Bruce ... "reciprocal" BrucePerens: Or "share and share alike". PeterYim: got it! BrucePerens: I have nightmares that MPAA will start getting kindergartens to teach kids not to share their toys. JamieClark: @Bruce: Maybe the pejorative itself is fixed in time and culture. I pointed out to a child that a certain thing might have a virus recently, and her reaction was 'cooooooool'.) JamieClark: (and she set about debugging it.) BrucePerens: And not enough about the license applied to ontologies. BrucePerens: We need to establish two things: 1. that we can make a tool together without license issues, and 2. that we can actually have ontologies interoperate when we're done. JamieClark: Mike's here, let's ask him. BrucePerens: That was my statement of approval. PeterYim: @MikeDean - good point about LGPL PeterYim: Bruce: LGPL is designed to combine with other licenses BrucePerens: LGPL can be combined with other licenses. There are additional requirements on distribution. You must be able to replace the LGPL component in situ, so we usually distribute it as a DLL. BrucePerens: It's possible from a legal perspective. BrucePerens: You must distribute source with the LGPL component or respond to requests for it for a period of 3 years. BrucePerens: BSD is in there because we know it doesn't have combinatorial problems. BrucePerens: BSD is more used than MIT and is functionally equivalent. BrucePerens: 3-clause BSD or 2-clause is probably immaterial. BrucePerens: They need an IPR committee. JamieClark: I agree BSD is optimally useful. But you are already talking about going out with a set of tools that's pregnant with Mozilla and Eclipse. So of what relevance is a generalized desire for BSD terms, if getting there would require rework of all the software packages you are now assembling? JamieClark: (*** Aspirational statements and license terms are distinct phenomena. ***) MikeDean: There probably should be a list of allowable library/dependency licenses as well, e.g. Apache and LGPL. MikeDean: There also needs to be some escape clause for evaluating other licenses on a case-by-case basis and/or re-evaluating the list periodically. JamieClark: Ick! lawyers! At run-time! JamieClark: If your system induces each contributor to ask a lawyer, every time, isn't that a fail? BrucePerens: They will anyway, if they work for an organization bigger than 50 people. BrucePerens: You can accept a contribution on a defined list of licenses along with a statement from the contributor that they have the right to contribute the work. BrucePerens: Don't say this in public before you have tenure. Stuart Turner: We are exposing to the community (healthcare/NIH) information about relevant extant ontologies. Our objective is more a registry than a repository. Despite clearly wishing for universal adoption of open source and open content licensing, we have a considerable number of ontologies (also terminologies/vocabularies) that have proprietary licensing that is not compatible with this vision. Excluding these ontologies is not consistent with our goals of informing the community, unfortunately, so we need to remain agnostic regarding the chosen license (for content). We also wish to identify a canonical and current source (IRI/URI/OID) for a given license as well. PeterYim: @Jamie - good point about substantiating our position with "why" we don't want to see ontologies be patentable PeterYim: maybe we should start a wikipage to build out the substantiation to our position JamieClark: i think we could contribute substantially to that debate, yes JamieClark: Among other things, John Wilbanks (who;s absent today) told us 3 weeks ago that his group is going to publish further on this point. BrucePerens: Software patents are the elephant in the living room of Open Source. BrucePerens: Tim Berners-Lee and W3C would probably join you in calling for there to be no patents on ontologies. BrucePerens: Software Freedom Resource Center. MikeDean: Thanks everyone PeterYim: Bruce: ERRATA (above) ... it's Software Freedom law Center JamieClark: Thanks all, very interesting. PeterYim: Action items: (i) modify the strawman based on today's conversation, essentially in removing "compatible" gift licenses as much as we can, and adding a bullet on libraries (Apache and LGPL) .... and float that to the mailing lists for further discussion and consensus building as planned; (ii) starting a wikipage to develop and substantiate our position on "no ontology patents;" (iii) check if there is serious objection (from the co-organizing communities) in our posting the two "no patents" positions ... if there is no serious objection to do it, we'll start a wikipage to collect endorsements to those positions; (iv) make contact and "recruit" legal professionals who will be willing to work with us (on a pro bono basis, at least initially, possibly long haul) so we do have proper legal advice and opinion when the team needs to deal with IPR issues in the future PeterYim: Great session ... thanks everyone! PeterYim: -- session ended: 12:33pm PDT -- Sent transcript to: peter.yim@cim3.com