OpenOntologyRepository Architecture

OpenOntologyRepository_IPR Policy and Issues Discussion

Background

As far as the OpenOntologyRepository (OOR) Initiative is concerned, we left off (at the end of the OntologySummit2008) with a vision and consensus that:

• OOR will be an open global collaboration (like the Apache Project),
• the OOR community will collaborate to develop a federated infrastructure of persistent open registries, repositories and associated services, for ontologies,
• contributions of the OOR community will include tools (technology) and content
  • tools and technology so contributed shall carry an open source technology license,
  • content (essentially ontologies) contributed shall be published under some open content license
• the OOR team will stand up at least one instance of the OOR, and
• while the OOR technology and content will be open, proprietary ontologies and ontology repositories should also be able to connect to and interoperate with it.

We also agreed, as a community, that we will defer discussion on specifics of the IPR issues (what licenses to standardize on, what does that mean, etc.) until we are further down the road with the work. ...(that was end April 2008)

We recognize that this is time to revisit the OOR-IPR issues and has identified that one of the critical tasks at hand is for the OOR team is to clarify IPR issues related to Ontology & OOR and adopt a consensus IPR policy for OOR contributions.

To this end, an OOR-IPR mini-series of virtual events are being organized:

• 2010_09_09 - Thursday: Joint OOR-Ontolog-NCBO-CC-IAOA-OASIS Panel Discussion - "IPR issues in Ontology and the OOR" session-1: an exposition on relevant IPR regimes - Keynote speaker: George Strawn - Chair: Peter P. Yim - Panelists: JamieClark, John Wilbanks, Bruce Perens - ConferenceCall_2010_09_09   

• 2010_09_16 - Thursday: Joint OOR-Ontolog-NCBO-CC-IAOA-OASIS Panel Discussion - "IPR issues in Ontology and the OOR" session-2: what are the IPR issues relating to open ontology repositories (and ontologies in general)? - Chair: Mark Musen - Panelists: Cameron Ross, Alan Rector, John F. Sowa, Bruce Perens, John Wilbanks, Peter P. Yim - ConferenceCall_2010_09_16   

• 2010_09_30 - Thursday: Joint OOR-Ontolog-NCBO-CC-IAOA-OASIS Panel Discussion - "IPR issues in Ontology and the OOR" session-3: discussion and consensus on licensing arrangements for the OOR Initiative, and positions we might take on related IPR issues - chair: Leo Obrst - Panelists: Peter P. Yim, Mike Dean, Bruce Perens, JamieClark - ConferenceCall_2010_09_30   

OOR-IPR mini-series objective

This "OOR-IPR mini-series" will, hopefully, start a dialog among the global ontology community, to specifically address IPR issues relating to the "open ontology repository (OOR)" initiative. The discussion will, invariably, touch upon IPR issues pertaining to ontology in general as well.

This mini-series is jointly organized by the OOR-team, the Ontolog-community, NCBO (US National Center for Biomedical Ontology), CC (Creative Commons), IAOA (the International Association for Ontology and its Applications) and OASIS (Organization for the Advancement of Structured Information Standards).

Given the complexity of the issues involved, one can look as this mini-series to merely be the beginning of a quest, by the collaborating parties and their communities, to fully understand the issues, and to get themselves into a position to address them.

In the opening session, we have invited some of the top experts in open IPR to give us an exposition on the IPR landscape and the state of relevant IPR regimes.

As a community, I trust we will, over the course of this mini-series and subsequent actions and events, address a very important set of issues, that really have to be cleared, before one can stand up effective open ontology repositories, and the world can get the most out of the potentials in science & technology of ontology and semantics.

---

(draft) consensus IPR Policy for the OOR Initiative

Our intent is to make the OpenOntologyRepository (OOR) software and system, as well as the content hosted in the open public instance of the OOR (repository) that this OOR-team will operate and maintain, to be as free, open, and unencumbered by IPR restrictions as we possibly can, so as to allow users of the software, system, and ontology content, the maximum freedom and flexibility. As such, we shall stipulate that ...

• software codes contributed to the OOR effort has to carry the "Simplified (two-clause) BSD License" (FreeBSD License).
  • the above will be the default position, unless it is otherwise mutually agreed in writing, between an authorized representative of the OOR team and the contributor. Such exception, even if agreed upon, needs to bind the contributing software codes to a compatible, non-reciprocal, "gift," open-source license, such as (included, but not limited to) the MIT License, the Apache License version 2.0 or the Mozilla Public License 1.1 (MPL);
  • it is, of course, also acceptable if the software contribution in question is in the public domain
  • Software libraries, under either "gift" or "reciprocal" software licenses, may also be used in the OOR software or system.

• Any content contribution to OOR should specify an IPR license, as part of the OOR "gatekeeping" requirements, which will be maintained in the content contribution metadata. Each repository using the OOR software can specify its own set of acceptable IPR content licenses.

• That content (essentially ontologies or other "knowledge organizational structures (KOS's)") contributed to the open public instance of OOR (which the OOR team will operate and maintain) will need to be licensed under either one of the following:
  • the "Simplified (two-clause) BSD License" (FreeBSD License), or
  • the "attribution only" Creative Commons Attribution License (CC BY 3.0 or its successor), or
  • the content being contributed is in the public domain.

(draft) Additional Community Action

As a joint community co-organizing this effort, we seek support and (optional) endorsements by individual members of the communities involved on the following:

• We hypothesize that ontology are fundamental to enabling interoperability and should be freely available for its full potential to be realized for the good of all. We, therefore, support the position that ontology should not be patentable. This joint-community will provide the platform for the thorough analysis and debate to validate that hypothesis, build out a strong case to support our position, which will then enable us to adequately inform the process that will be making the official ruling in this matter.

• We will encourage those members of our community who believe that software patents ( ref. ) will discourage, rather than encourage, innovation, to actively participate in the public debate and in collective efforts that advocate the position that software should not be patentable.

---

This was our position towards the end of the three Sep-2010 OOR-IPR mini-series sessions. It has since been overtaken by the above "Draft Consensus OOR IPR Policy" which will be presented for adopted by the community. (The positions below are maintained for archival and reference purposes.)

(Now obsolete) Candidate IPR Policy for the OOR Initiative

• That software contributions to the OOR effort has to carry a compatible, non-reciprocal, "gift," open-source license. If one is not specified by the contributor, the software license will default to the "Simplified BSD License" (FreeBSD License).
  • compatible open-source licenses include (but not limited to) the MIT License, Apache License version 2.0, Mozilla Public License 1.1 (MPL), and Eclipse Public License version 1.0 (EPL);
  • it is also acceptable if the software contribution in question is in the public domain

• Any content contribution to OOR should specify an IPR license, as part of the OOR "gatekeeping" requirements, which will be maintained in the content contribution metadata. Each repository using the OOR software can specify its own set of acceptable IPR content licenses.

• That content (essentially ontologies or other "knowledge organizational structures (KOS's)") contributed to the open public instance of OOR (which the OOR team will operate) will either one of the following:
  • a compatible, non-reciprocal, "gift," open-source license. If one is not specified by the contributor, the software license will default to the "Simplified BSD License" (FreeBSD License).
    • compatible open-source licenses include (but not limited to) the MIT License, Apache License version 2.0, Mozilla Public License 1.1 (MPL), and Eclipse Public License version 1.0 (EPL), OR
  • a Creative Commons Attribution only License (CC BY 3.0 or its successor).
  • it is also acceptable if the content contribution in question is in the public domain

(Now obsolete) Candidate Position Statements that we will make

(as a joint community co-organizing this effort) which we will seek (optional) endorsements by individual members of the communities involved.

• we believe ontologies are fundamental to enabling interoperability, and therefore support the position that ontology should not be patentable.

• we also believe that patents covering software will discourage, rather than encourage, innovation, and therefore support the position that software should not be patentable.

---

Input and Discussion

• 2010.08.26 - from Peter P. Yim

Given the fundamental nature of ontologies (Upper Ontologies or Foundational Ontologies), I trust we should consider the very notion of whether or not ontologies (or, at least, certain ontologies) should be given IPR protection ... along the same lines of thinking as:

o the fact that no one can copyright the alphabets or book titles

o the fact that no one is allowed to patent algorithms, and the ongoing debate about software patents or whether genes can be patented

• 2010.08.25 - IAOA confirms its participation as a co-organizer of the OOR-IPR mini-series - see: http://ontolog.cim3.net/forum/iaoa-general/2010-08/msg00000.html

IAOA has decided to join the OOR-team, the Ontolog community, NCBO (US National Center for Biomedical Ontology) and Creative Commons (CC), and participate as co-organizer for the mini-series on "Intellectual properties Rights (IPR) issues relating to the OOR Initiative". The mini-series will start in September 2010.

• 2010.08.17 - From John Wilbanks (after an telecon with Peter P. Yim on the preparation for the OOR-IPR mini-series)

Peter - we are confirmed that we can list CC as part of the conveners. Looking forward to working with you.

• 2010.05.20 - from Nigam Shah to John Wilbanks:

Hope your grant flurry has passed. I wanted to pick up on our thread again .. esp the item 1 below. BioPortal is being adopted by several communities (e.g. the Open Ontology Repository community) and there is an active debate on how to handle IP for the contents of a particular installation. The discussions have taken the form of a "Joint OOR-Ontolog-NCBO IPR mini-series", comprising 3 sessions. We would love to get your input while we try an figure out the IP issues around ontology content. ... Peter Yim, cc'ed, is the lead organizer for the mini-series and can answer further questions about it.

• 2010.05.20 - from Nigam Shah (NCBO) to John Wilbanks:

(follow up from Sage congress discussion) Things we'd like your advice on:

1 - how to handle the intricacies of open licensing on "BioPortal", which contains code we write, code from public libraries, ontology content that others submit as well as content we license from NLM.

2 - how best to support the shared names effort led by Science Commons.

• 2010.05.06 - From Peter P. Yim to Cameron Ross ref. the OOR-IPR writeup

... rather than dismissing Technology IPR as being something

we won't be talking able, I'd suggest we need that conversation, partly to educate the community on the options, and partly to facilitate the OOR community to make a conscious decision on the selection of (a) technology license(s) for the OOR initiative.

• 2010.05.06 - from Cameron Ross (initially based on a telecon with Dave Rubenson (NCBO) and Peter P. Yim with modifications based on subsequent discussions) - updated 2010.08.31

1) We are going under the assumption that, similar to software, an

ontology, or ontology module, may be owned, and that this ownership may be protected under copyright law. We are also assuming that the rights of ownership include the ability to distribute some rendering of an ontology and to have the privileges associated with using the ontology specified within a license agreement. This is analogous to the copyright licensing model widely used for traditional software-related artifacts. One of the tenets of the OOR initiative is to promote and develop the mechanisms required to support the OPEN distribution of ontologies through federated repositories. Therefore, Ontology artifacts contained within such repositories will need to be associated with an OPEN license to realize this objective. This model of open dissemination is analogous to the one employed by Open Source software community. However, we believe that there are specific differences between traditional software artifacts and ontologies that render existing Open Source licenses ineffective for this purpose. Licenses promoted by the Creative Commons may be better suited to for licensing ontology-centric artifacts. However, how the current Creative Commons licenses map to the requirements of the OOR requires further investigation. Therefore, one of the objectives of the IPR panel session is to evaluate existing open licenses and to make specific recommendations regarding ontology licensing. If the panel should find that an appropriate license does not exist, then the panel will work with one of the existing open licensing groups to develop a license that is tailored to the needs of an the OOR objectives. Note that the potential concern over the proliferation of open licenses for ontologies will be considered and, if at all possible, strategies will be developed to help minimize this concern.

2) The OOR initiative is tasked with the OPEN distribution of

ontologies through federated repositories. However, the term OPEN needs to be defined. There are specific use cases that call for controlled access to an open ontology repository. For example, there may be specific policies that prevent ontology dissemination to specific jurisdictions. There have also been use cases presented on the OOR discussion forum that require controlled access to proprietary extensions of the OOR. These use cases suggest that IPR within an OOR instance, or federation of instances, may vary according to a specific ontology. The panel session will consider these differences and make recommendations on how to manage the variability of rights. This will include, but is not limited to, the concerns associated with integrating ontologies that are released according to different licenses. Issues with license compatibility will be specifically addressed.

3) The federation of OOR instances is expected to be populated with

content from a variety of sources originating from a variety of jurisdictions. This raises concern over the provenance of ontology content being contributed. A similar concern is currently echoed within the Open Source software community. The legal integrity of the OOR could be undermined if the ownership of contributed content were to be brought into questions. This is similar to the concern that arose Santa Cruz Operations (SCO) claimed that specific contributions made by IBM to the Linux kernel were actual owned by SCO. SCO subsequently approached several Linux end users demanding compensation. The OOR IPR panel session will consider this concern and will recommend mechanisms to help vet the provenance of both software and content contributions made to the OOR initiative, thus protecting the legal integrity of the OOR federation.

4) Ontological engineering is maturing rapidly and best-practices are

emerging to support the construction of ontologies as a collection of engineered artifacts. One specific practice is constructing modular ontologies. Although this is a practice that the OOR should support, there are certain IPR concerns that arise. If we assume that the OOR will allow ontologies to be contributed according to a multitude of licenses, how does this impact the licensing of an aggregating ontology, otherwise known as a derivative work? Specific issues include the compatibility and viral nature of certain licenses. Note that similar concerns exist within the Open Source software community and dealing with this issue has become a major challenge. The OOR IPR panel session has a unique opportunity to learn from these lessons for the benefit of the ontology community.

5) There are efforts underway to convince standards bodies such as

ISO, IEC, ANSI etc. to adopt ontologies as a means of formally specifying standards. The OOR IPR panel session will consider this possibility and will attempt to structure their recommendations to facilitate this endeavor.

6) Given the fundamental nature of ontologies (Upper Ontologies or

Foundational Ontologies), the IPR panel will consider the very notion of whether or not ontologies (or, at least, certain ontologies) should be given IPR protection ... along the same lines of thinking as:

o the fact that no one can copyright the alphabets or book titles

o the fact that no one is allowed to patent algorithms, and the

ongoing debate about software patents or whether genes can be patented

Well, that's all I have for now. Perhaps we can bounce it back and

forth to flesh the concepts out a bit. Please let me know what you think.

• 2010.04.28 - Bruce Perens confirmed his participation in the OOR-IPR mini-series as a panelists and advisor (after a extended telecon with PeterYim)

• 2010.04.08 - from Peter P. Yim to John F. Sowa:

Thanks, John. I concur, in principle. I think practically all the key

players who are involved in OOR or SIO now are aligned on this. If we look at the community and their contributions now ...

Contributing platform & code: NCBO-Stanford, BBN, NEU, UToronto, CIM3, ...

Contributing content: the usual Upper Ontology suspects, federated

content from BioPortal, ... and just getting geared up to enlist other to contribute domain ontologies and other KOS's.

This is a very important conversation, that has to be taken with the

above community (to gain consensus and buy-in) to get to the next level of detail. Some may already have IPR policies (NCBO?) ... Some may have constraints and we will have to mitigate.

I think the upcoming IPR session and the discussion revolving around

that would provide the right opportunity for us to get things tied down more concretely. Maybe we do not have a huge urgency to define it much further now (especially in view of the content contributors, that are still emerging.)

We definitely should talk about it more, at the OOR team meetings,

even before we put the IPR session together.

• 2010.04.08 - from John F. Sowa to Cameron Ross and Peter P. Yim:

Cameron and Peter, ... I believe that Eclipse and Creative Commons are both extremely

important developments, and I suggest that we support both.

There are so many different licenses around that I wouldn't

even try to summarize the issues. But following are some considerations:

1. The content contained in the OOR should be contributed for

free use by anyone for any purpose. I believe that the Creative Commons license allows such use, provided that the users cite the original copyright holder.

2. The tools should be available under a license that allows

free use and modification of any software for any purpose, private or commercial. There are many different licenses that support such use, and I'd leave it to the lawyers to decide which one is best.

3. The overwhelming majority of software on Source Forge is

effectively dead (or in IBM's euphemism, functionally stabilized). A few people other than the original contributor may have downloaded the software, but most of it languishes for years without any updates by anyone.

4. There are a few highly successful open source projects

(e.g., Linux, Apache, Mozilla, Open Office), but most of the development on those projects is limited to a small core of insiders, which gets most of its support from businesses and other organizations that actively use that technology.

5. Eclipse is one of the few highly successful open source

projects that is widely used, developed, and extended by a wide range of developers other than the original sponsors. It is also widely accepted by the IT industry as a mainstream development platform.

6. Endorsement by the Eclipse Foundation would ensure that

the SIO project is taken seriously by the IT industry.

For these reasons, I believe that we should definitely make

the SIO software compatible with Eclipse. I think that we should also have a stand-alone version that does not require Eclipse, since many users wouldn't want to and shouldn't be required to download the entire Eclipse platform just to use the OOR and supporting tools.

I am not familiar with the policies of the Eclipse Foundation,

but if Cameron or others have contacts with them, I would encourage inquiries about some kind of collaboration.

• 2010.04.07 - Peter P. Yim to Cameron Ross:

Cameron, ... Thank you for the earlier IPR-related post and the suggestions.

1. I think this should be an OOR rather than a SIO conversation. (May

I, therefore, suggest using the [oor-forum] list.)

2. A lot of discussion has been had (as Mike Dean put it, during the

Ontology Summit 2008 presentations: "everyone had to have a slide about "open").

3. The general direction has been gravitating (similar to your

suggestions) toward an Apache type ("academic", rather than "reciprocal") open source license for the tools and platform, and for a Creative Commons license for the content (i.e. ontologies, vocabularies, etc.) ... and nothing has been locked in concrete ... which needs to be done, eventually, when we get into the next level of granularity in terms of licensing.

4. We also need to address whether to register them, point to to

them, and how to federate with non-open content, etc.

5. As mentioned in response to your earlier message today, the OOR

team has an IPR session planned. ...

• 2010.04.07 - Cameron Ross:

Last December I explored the possibility of having the Eclipse Foundation steward a project for an Ontology Development Environment(ODE). There are pros and cons with doing this, but in the end I think that the pros may out way the cons in the context of the SIO project. I'm sending this email off-list as I don't want to distract the focus of the SIO discussion unnecessarily, but please feel free to bring up the subject on the list if you think it's appropriate. I can contact Donald Smith, Director of Ecosystem Development for the Eclipse Foundation, to see what's involved if there is support for the idea.

---

References and Resources

• OpenOntologyRepository initiative - homepage

• About the OOR effort
  • 2010_04_29 - OntologySummit2008_Communique
  • 2010_04_29 - Presentation on OOR at the Ontology Summit 2008
  • 2010_06_24 - Presentation on OOR at the SemTech2010 Conference

• "Open Source Definition" by Bruce Perens - http://www.opensource.org/osd.html

• Creative Commons Licenses (3.0) - http://creativecommons.org/about/licenses

• "What is Science Commons?" by John Wilbanks - http://creativecommons.org/weblog/entry/5695

• Open Source Licensing: Academic v. Reciprocal - ref. http://www.ibm.com/developerworks/linux/library/os-license2/

• IPR Policy of the Eclipse Project - http://www.eclipse.org/org/documents/Eclipse_IP_Policy.pdf
  • EPL (Eclipse Public License) and EDL (Eclipse Distribution License) - ref. http://www.eclipse.org/legal/

• Open Source Licenses - ref. http://www.opensource.org/licenses/category - ref. pertinent licenses:
  • Apache License (v2) - http://www.oss-watch.ac.uk/resources/apache2.xml
  • MPL (Mozilla Public License) - http://www.mozilla.org/MPL/
  • BSD (Berkeley Software Distribution) - ref. http://en.wikipedia.org/wiki/BSD_licenses
  • GPLv2 (GNU General Public License, version 2) - see: http://www.gnu.org/licenses/gpl-2.0.html
  • GPLv3 (GNU General Public License, version 3 - 29 June 2007 (current)) http://www.gnu.org/licenses/gpl.html
  • LGPL (GNU Library or "Lesser" General Public License) - ref. http://www.opensource.org/licenses/lgpl-license.php
  • AGPL (GNU Affero General Public License) - ref. http://www.affero.org/oagpl.html
  • Examples:
    • Apache Software: IPR Registry - http://archive.gria.org/docs/ogsadai/1.0/docs/IPR-registry.html

• OASIS IPR Policy - http://www.oasis-open.org/who/intellectualproperty.php

• on Software Patents, ... etc.
  • http://en.wikipedia.org/wiki/Software_patent
    • http://en.wikipedia.org/wiki/Software_patent_debate
  • "Are Software Patents Evil?" by Paul Graham / Mar. 2006
  • "Abolishing software patents would unleash more innovation" by Vivek Wadhwa / Aug 9, 2010.

• First-to-invent v. first-to-file patent regime - ref. http://en.wikipedia.org/wiki/First_to_file_and_first_to_invent

• (US) Patent Reform Act of 2010: An Overview - http://www.patentlyo.com/patent/2010/03/patent-reform-act-of-2010-an-overview.html