"Biological Open Source"

[Janet Hope]

On 15/11/2006, at 11:39 PM, Michael Tiemann wrote:
>
> Janet, I for one welcome your participation.  Moreover, while it is  
> not
> likely to lead to a specific license approval, it is very much on  
> topic
> for the expertise this list represents.
>

Thanks, Michael, for your positive response.  Having received some  
private as well as public encouragement and no discouragement, I'll  
go ahead and put my question to the list.  (It has been suggested  
that a new list could be started for non-software open source  
licensing issues to cut down on off-topic posts -- I think that would  
be very helpful.)

This post is long. Most of it falls under the headings "Background"  
and "Preliminary licence analysis".  Readers who are interested, but  
not that interested, are encouraged to skip those parts.

THE LICENCES

The licensor, CAMBIA, has issued two "Biological Open  
Source" ("BiOS") licences, one for Plant Enabling Technologies  
("PET") and one for Genetic Resources Indexing Technologies  
("GRIT").  Each must be read in conjunction with its own Technology  
Support Services Subscription Agreement ("PET TSSS" and "GRIT  
TSSS").  All of these instruments are currently in version 1.3 and  
are available at http://www.bios.net/daisy.

Let me say up front that I am not affiliated with CAMBIA and have not  
been involved in the process of drafting the BiOS licences.  As a  
matter of courtesy to the drafters, I am simultaneously posting my  
comments to CAMBIA's own discussion forum.  However, as Michael  
suggests, I suspect that license-discuss is where much of the  
relevant expertise is to be found.

THE QUESTION

The question is: To what extent is CAMBIA's BiOS licence scheme "open  
source" in the software sense?

In other words:  If the BiOS licences were software/copyright  
licences, would you recommend them for OSI approval?  If not, why not?

SPECIFIC QUERIES

In particular (see the rest of this post for more detail on each of  
these points):

(1) If a licence is primarily a patent licence and the patent grant  
itself purports to be open source (not just compatible with an open  
source copyright grant elsewhere in the licence), should field of use  
restrictions be permitted? (OSD #6)

(2) What should be regarded as the equivalent of "source code" in a  
biotechnology licence? Should it include materials etc in addition to  
that which would be required for disclosure under patent law? If so,  
on what terms should it be made available to users? Specifically, is  
it acceptable to charge a fee that does more than recover the costs  
for storage/retrieval/postage of samples etc (OSD #2)? Is it  
acceptable for the amount of the fee to depend on the size and type  
of the user's organisation (OSD #5) and the length of the user's  
"subscription" to ongoing "technology support services"?

(3) What is the nearest patent equivalent of the open source  
copyright licence grant to copy, modify, and distribute copies or  
derivative works (OSD #1)?  Does it depend on the type of patent (eg  
process or product)?  Is the patent grant in the BiOS licences  
effectively limited by the exclusion of "Improvements" from the  
definition of "BiOS Licensed Products"?

(4) Is a right to sublicense essential in an open source licence? Is  
it acceptable for this right to be limited or absent? If so, what  
other conditions should be met to ensure that a licensor cannot  
impose non-open source restrictions as a condition of issuing a new  
licence (eg OSD #7)? How can these conditions be met in the case of a  
technology to which the equivalent of a copyright notice cannot be so  
easily attached as it can to software code?

(5) Does the grantback in the BiOS licence go beyond a copyleft-style  
provision in restricting users' freedom to fork the innovative  
process? Is there any pressing legal or practical reason why the  
licensor in a biotechnology/patent context could not simply require  
licensees to make defined improvements available on the same terms as  
the original licence, as in the software/copyright context?

BACKGROUND

CAMBIA (formerly an acronym for Center for Application of Molecular  
Biology in International Agriculture) is a private, not-for-profit  
research institute located in Canberra, Australia.  Founded in 1994  
by its present CEO Dr Richard Jefferson, CAMBIA is a small  
organisation that -- atypically -- combines "wet lab" development of  
biotechnology research tools with intellectual property informatics  
and policy development.  CAMBIA is financed by grants from  
philanthropic organisations including the Rockefeller Foundation, by  
national and international research funding bodies, by official  
development assistance and by licence revenue from its own patented  
technologies.(See http://www.cambia.org/daisy/cambia/589.html.)

In 2005, CAMBIA launched an initiative called Biological Innovation  
for Open Society (BIOS: note the upper case "I").  The BIOS  
initiative consists of:
(1) Patent Lens: searchable databases containing EPO, US and PCT  
patent documents, together with ancillary IP-related information and  
tutorials.
(2) BioForge (styled after Sourceforge.net): a portal for protocol- 
sharing, comments on patents, and discussion tools in both public and  
secure environments; intended to develop into a collaborative  
technology development platform.
(3) Biological Open Source (BiOS with a lower case "i") is the aspect  
of the BIOS initiative that is "intended to extend the metaphor and  
concepts of Open Source to biotechnology and other forms of  
innovation in biology". (See CAMBIA BiOS License for Plant Enabling  
Technology Version 1.3, Recitals, first paragraph.)

A description of the materials and methods that researchers can  
obtain from CAMBIA, including an indication of which materials and  
methods are available under which BiOS licence, is at http:// 
www.cambia.org/daisy/cambia/materials.html.  These technologies are,  
of course, quite different from those licensed under OSI-certified  
software licences.  Plant enabling technologies and genetic resource  
indexing technologies consist not of software code (though software  
may be one component), but of heterogeneous methods and materials,  
including living biological materials.  Typically, such technologies  
are legally protected by a mixture of intellectual and personal  
property rights, including patents.  By contrast, I understand that  
most open source software licences are primarily copyright licences,  
though some also contain a limited patent grant.

As I mentioned in my first post, the fact that the BiOS licences  
apply to plant enabling technologies and genetic resources indexing  
technologies means that they cannot be certified by the Open Source  
Initiative as being "open source" licences because the terms of the  
official Open Source Definition are not broad enough to accommodate  
the legal and technical differences just described. In consequence,  
there is no certification signal to help potential users determine  
whether or not signing up to a BiOS licence will actually give them  
the nearest equivalent to the "software freedom" guaranteed by the  
principles of open source software licensing.  Clearly, the task of  
translating open source licensing principles from one technology  
setting to another is one that creates room for both legitimate  
adaptations and distortions of the kind that would lead the OSI to  
refuse to certify the relevant licence if the technology in question  
were, in fact, software.  By my preliminary analysis, the BiOS  
licences contain a mixture of the two.  However, I (and many others)  
would like to be able to supplement that analysis with the expert  
views of people on this list.

Note that I am in no way impugning CAMBIA's good faith or suggesting  
that the BiOS licences are "bad".  Certainly, CAMBIA has made a  
serious effort at a very difficult task.  The licences may well be  
good licences, and good for innovation.  However, the question here  
is whether it is reasonable to describe them as "open source".

PRELIMINARY LICENCE ANALYSIS (WARNING: MAY CONTAIN ERRORS! TINLA!!)

For simplicity, I refer here only to the terms of the BiOS Plant  
Enabling Technology (PET) agreement. The PET scheme closely resembles  
the GRIT scheme. The main difference is that whereas the PET scheme  
contains a limited sublicensing right, the GRIT scheme does not  
contain any right to sublicense.

1. Overall scheme of the licence

The BiOS licensing scheme encapsulated in version 1.3 of the PET  
Licence and corresponding TSSS Agreement contemplates an initial  
contribution of IP and technology by CAMBIA that may act as a seed  
for ongoing technology development. ("IP and Technology" is defined  
in Licence cl. 1.8.)  Thus, the subject of the initial licence grant  
is a collection of CAMBIA's own patents, materials and technology  
data.(See definitions of "CAMBIA Patents", "CAMBIA Material" and  
"Technology Data" in Licence cll.1.9, 1.10 and 1.13 respectively.)   
Downstream, this initial subject-matter may be supplemented by  
"Improvements", "Improvement Patents" and "Improvement  
Materials" (Licence cll 1.6, 1.7 and 1.11 respectively) developed by  
the licensee,  non-exclusively granted back to CAMBIA under Licence  
cl. 3 and then redistributed by CAMBIA to all BiOS licensees as part  
of the grant of "IP and Technology".

2. Parties

Although the licence is described on the BIOS forum as a "template",  
one party (CAMBIA) is explicitly named throughout the licence text as  
the licensor.  I understand that this is discouraged in the OSI  
approval process because it contributes to licence proliferation;  
however, this is a relatively minor point.

The sense in which the BiOS licences are templates is that they  
contemplate a number of licensees all receiving CAMBIA's permission  
to use the IP and Technology on "substantially similar" terms. (See  
Licence Recitals, paragraph 3. Why are the terms "substantially  
similar" and not identical?  Two reasons: (1) different licensees may  
agree to different numbered versions of the licence (see Licence cl.  
7.5 and TSSS Agreement cl. 7.6). (2) The terms on which licensees can  
access any materials needed to practise the licensed inventions vary  
according to the type and size of the licensee organisation (see TSSS  
Agreement cl. 3 and Annex D).)

3. Patent grant

The licence grant is contained in clause 2 of the PET Licence, the  
key part of which reads:
"2.1 CAMBIA hereby grants to BiOS LICENSEE under Licensed Patents in  
the Field of Agriculture a worldwide, non-exclusive, royalty-free  
right and license to make and use the IP & Technology for the purpose  
of developing, making, using, and commercializing BiOS Licensed  
Products without obligation to CAMBIA...."

The nature of the grant is crucial to the question whether the BiOS  
licences can be justifiably described as "open source". Note that  
unlike open source software licences, the BiOS licence is primarily a  
patent licence.  While patent grants in open source software licences  
must be compatible with an open source copyright grant, they are not  
the primary means by which these licences seek to protect users'  
rights to access and use the technology.  The question here is not  
merely whether the BiOS patent licence grant is \textit{compatible}  
with open source principles, but whether it is itself "open  
source" (whatever that is taken to mean in the biology context).

3a. Grant is restricted to a particular field of use

The BiOS patent licence grant is limited to a particular field of  
use, namely Agriculture. ("Agriculture" is defined quite broadly in  
Licence cl. 1.3.)  There is a question whether this conflicts with  
OSD # 6.

While a strict formal interpretation of open source principles would  
presumably prohibit field of use restrictions in open source  
biotechnology licences, the difficulty with such an interpretation is  
that in the patent context, field of use restrictions, together with  
territorial restrictions, may be a useful tool for making technology  
that would otherwise be bound up in exclusive licences more readily  
available for public interest and broader commercial use.  The  
licensing policy of Public Intellectual Property Resource for  
Agriculture (PIPRA), a collective intellectual property management  
regime for agricultural biotechnology recently established by a group  
of land grant universities in the United States, is a case in  
point.   One of PIPRA's primary aims is to overcome the fragmentation  
of public sector intellectual property ownership by identifying  
residual rights retained by members who have assigned unnecessarily  
broad exclusive rights in important technologies to major commercial  
firms. One of the architects of the initiative explains the  
importance of field of use and territorial restrictions in this  
context: "Best practice includes partitioning of patents: if you go  
and license something like your agrobacterium technique, license it  
just for cotton; or better, for cotton in the US; or even better...  
define which varieties, or... constrain it to varieties owned by the  
licensee company in the US.  The more you can constrain the space of  
the technology grant, the more is still left over [that you may  
choose] to put into the commons." (Greg Graff, personal  
communication.) Thus, the value of such field of use restrictions in  
terms of achieving wider access to key biotechnologies may outweigh  
the value of keeping to a strict analogy with open source software  
licensing.

What do people think about this? I understand that with respect to  
patented code, the F/LOSS community objects to field of use  
restrictions in software patent licences because such restrictions  
are perceived as capable of insidiously undermining freedoms granted  
in relation to the same code under an open source copyright licence.   
However, the approach of open source community leaders to this  
situation appears to have been to campaign for maximum breadth of  
patent licences, while remaining willing to compromise for the sake  
of workability.  Thus, many open source licences, including the  
Apple, IBM and Mozilla licences, have field of use restrictions in  
their patent grants. The W3C Patent Policy also represents a  
compromise.  How far should this willingness to compromise go in a  
case where the patent grant is the main or even the only grant in the  
licence?

3b. Scope of grant

The grant is a "right and licence to make and use the IP and  
technology for the purpose of developing, making, using and  
commercialising BiOS Licensed Products".  Note that the grant does  
not explicitly include a right to sell the IP and Technology itself.   
Nor does it permit the licensee to make or use the IP and technology  
for purposes other than developing, making, using or commercialising  
BiOS Licensed Products.  "BiOS Licensed Product" is defined in  
Licence cl.1.4.  That definition is discussed further below, but  
significantly, it is restricted to assets that are substantially  
distinct in some respect from the licensed IP and technology and are  
"intended for commercialisation".  "Commercialisation" is not defined  
in the BiOS licence.  However, in light of recent US case law  
concerning patent infringement by universities (eg Madey v Duke), it  
may be quite broad.

The OS mantra is "anyone, anywhere, for ANY PURPOSE, must be free to  
copy, modify AND DISTRIBUTE the software, for free or for a fee...".   
At first glance, the BiOS patent grant seems considerably narrower  
than this.  However, patent law grants a patent owner a different set  
of rights to those of a copyright owner.  It is therefore not a  
straightforward question what the equivalent of the broad open source  
copyright licence grant should be in the patent context.  It is  
arguable that an open source copyright grant covers all of the  
exclusive rights of the copyright owner that are relevant to the  
licensed technology, and that an open source patent grant should do  
the same.

IAAL, but not a patent lawyer.  However, I understand that in  
Australia at least, the precise form of a broad patent licence grant  
that would cover all of the owner's rights-to-exclude would depend on  
the nature of the patent claims.  Section 15 of the Australian  
Patents Act 1990 gives the patent holder the right to exploit the  
invention. The definition of "exploit" in Schedule 1 of the Act  
distinguishes between products and processes. For products, it  
includes the right to make, hire, sell or otherwise dispose of the  
product. For processes, the definition includes the right to make,  
hire, sell or otherwise dispose of products resulting from the  
process. Without looking at the claims in CAMBIA's patents, it  
appears that they are process patents (ie methods of doing plant  
enabling technology and genetic resource indexing technology). The  
licences grant the right to use the methods. Any products are  
supplied separately under the support agreement. The licences  
actually grant the right to sell products using the methods under  
licence. Thus, the licence grant may be broad enough to satisfy open  
source principles with respect to the initial IP and Technology.

What do you all think about this?

3c. Limited right to sublicense

A final point to note regarding the BiOS licence grant is that it  
includes a limited right to sublicense, as follows:
"2.1.1 granting limited sublicenses to third parties... to conduct  
research and/or development activities...for BiOS LICENSEE, provided  
that the product and/or other results (including all intellectual  
property rights) resulting from said limited sublicense are owned  
exclusively by BiOS LICENSEE, said limited sublicenses to terminate  
when such activities cease or such ownership terminates, and
"2.1.2 granting limited sublicenses to third parties... for the sole  
purpose of commercializing BiOS Licensed Products that embody the IP  
& Technology or are generated by use of the IP & Technology, said  
limited sublicenses to terminate when said commercialization ceases.
"BiOS LICENSEE shall be responsible to ensure... that any  
Improvements produced by sublicensees are considered to be  
Improvements hereunder.... BiOS LICENSEE shall provide a list of  
sublicensees to CAMBIA in writing at least once a year....
"BiOS LICENSEE shall further be responsible to ensure that...third  
parties to whom a sublicense has been granted are notified... that  
said third parties have no right to sublicense absent the execution  
of a BiOS License Agreement with CAMBIA, and that CAMBIA may be  
approached for a BiOS License Agreement.
"Other than stipulated under this Article 2.1, no further right to  
sublicense is granted to BiOS LICENSEE hereunder."

I am unclear as to whether a right to sublicense is essential for OSI  
certification of new licences.  My understanding from Larry Rosen's  
book is that the right to sublicense is considered desirable, but is  
not a feature of all OSI-approved software licences.  I imagine that  
in the absence of a sublicensing right, OSD #7 becomes very important  
because unless people are sure they are entitled to a licence no  
matter what, the need to go back to the initial licensor for  
permission to use the technology would not be a mere formality and  
could become too restrictive.  However, it is not clear exactly how  
OSD #7 would operate in relation to the exchange of biological  
materials, methods etc, as distinct from code or other content to  
which a copyright notice can be easily attached.

Can anyone clarify the OSI requirements re sublicensing?  (Sorry if  
this has been dealt with in previous posts.)

4. Source code

The freedom to copy and modify open source software programs and to  
distribute copies and modifications is only one important aspect of  
open source licensing.  Another is the practical enablement of that  
freedom through unrestricted access to software source code.  In the  
case of patented biotechnologies, it may be argued that the  
equivalent of source code is automatically publicly available because  
of the disclosure requirement under patent law, and that an open  
source-style licence grant is all that is needed to render a patented  
technology "open source".  On the other hand, biotechnology licensors  
typically go beyond this level of disclosure when attempting a  
genuine transfer of technology.  This practice suggests that patent  
disclosure many not be enough to give full practical effect to the  
legal freedom to use and improve a technology that is intended to be  
conferred by an open source licence grant.

In the BiOS PET licensing scheme, separate provision is in fact made  
for access to materials needed to practise the licensed patents.   
This access is governed by the terms of the TSSS Agreement (see  
Licence cl. 4) and appears to be conditional upon (1) a substantial  
annual payment for at least three years, depending on the licensee's  
organisation type and number of employees and (2) significant  
reporting requirements regarding new materials, technology data and  
project ideas.  If access to materials needed to practise the  
licensed payments is regarded as equivalent to access to source code,  
these restrictions represent a significant departure from established  
open source principles of non-discrimination among licensees, access  
to source code for free or at the minimum charge required to recover  
the provider's costs, and freedom from ongoing obligations to the  
licensor, including obligations that take the form of reporting  
requirements.

5. Grantback
The next phase in the scheme established by the BiOS licences and  
TSSS agreements is triggered when a licensee, in the course of  
exercising the licence grant, makes or discovers something new.

The BiOS PET licensing scheme distinguishes between "BiOS Licensed  
Products", which may be developed, made, used and commercialised  
"without obligation to CAMBIA", and "Improvements", which are subject  
to a grantback obligation  (see below) that also covers Improvement  
Patents and Improvement Materials (Licence cll. 1.7 and 1.11,  
respectively).  The definitions of these different categories of  
potential follow-on innovations are complicated and tinged with  
uncertainty.  However, the fundamental intention appears to be to  
allow the licensee to appropriate any new asset that has been  
generated through use of the IP and technology and is intended for  
commercialisation.  Such an asset is a "BiOS Licensed  
Product" (Licence cl. 1.4), provided it does not fall within the  
definition of "Improvement" (Licence cl. 1.6).

The definition of "Improvement" is thus a critical element of the  
BiOS scheme. This definition is potentially extremely broad.  Note in  
particular that even something as minor as a repeatable observation  
can constitute an Improvement, and that an Improvement need not be a  
Plant Enabling Technology (itself very broadly defined in Licence cl.  
1.12) because of the catch-all "but for the terms of this License  
Agreement...".

5a. Does the definition of "Improvement" limit the licence grant itself?

A preliminary question is whether the broad definition of  
"Improvement" in Licence cl. 1.6 effectively limits the scope of the  
main licence grant in Licence cl. 2.  As noted earlier, the purpose  
of the grant is limited to "developing, making, using and  
commercialising BiOS Licensed Products".  "BiOS Licensed Products" is  
defined in cl. 1.4 to mean "any tangible or intangible asset of BiOS  
LICENSEE (including without limitation any material or method, but  
excluding Improvements)".  This suggests that anything that falls  
within the definition of Improvement automatically falls outside the  
definition of BiOS Licensed Product and therefore is not covered by  
the grant even for purposes other than commercialisation.

Similarly, any new development that is neither a BiOS Licensed  
Product nor an Improvement does not appear to be covered by the terms  
of the licence.  In either case, further permissions may be required  
in order to avoid infringing the licensor's proprietary rights.   
CAMBIA appears to be free to make such permissions conditional upon  
payment of royalties or other conventional licensing terms -- or to  
refuse them altogether.

5b. Grantback: copyleft equivalent or licensing "club"?

 From an open source perspective, the breadth of the definition of  
"Improvements" is not necessarily a problem, although it could be if  
it extended the range of follow-on innovations that would be caught  
by the grantback well beyond those that would be caught by a copyleft- 
style provision relating to "derivative works".  As I understand it,  
the question of what constitutes an appropriate incentive for  
licensees to innovate in a copyleft-style licence is a separate issue  
from compatibility with open source principles.  The broadest  
examples of copyleft "hooks" in open source software licences are  
tailored to catch all derivative works that are distributed outside  
the boundaries of the licensee's own organisation; other open source  
licences permit varying degrees of freedom to operate with derivative  
works. The BiOS grantback appears to fall somewhere in the middle of  
this spectrum.  In a 2004 article on open source patent licensing,  
Sara Boettiger and Dan Burk point out that copyright confers  
exclusive rights only against unauthorised copying or other  
violations of the specifically enumerated rights of the copyright  
owner arising out of contact with the copyrighted work itself.  By  
contrast, independent creation is not a defense to claims of patent  
infringement. The BiOS licence appears designed to replicate the  
copyright situation by excluding from the definition of Improvement  
improvements that are "developed without the use of the IP and  
Technology".  Similarly, the BiOS arrangement appears intended to  
mirror open source software licences by permitting in-house use of  
improvements that are actively protected as trade secrets: these are  
not caught by the BiOS grantback. (Note, however, that this escape  
clause may be somewhat narrower than the positive requirement in open  
source software licensing for a derivative work to be externally  
deployed before it triggers any copyleft obligation: clearly, there  
is a difference between active dissemination of a technology and  
merely failing to  adequately protect a trade secret.  Further, note  
that in-house use of an improvement under conditions of trade secrecy  
may not be permitted at all under the BiOS scheme unless interpreted  
as a form of "commercialisation": see Licence Grant in cl. 2.2 and  
definition of BiOS Licensed Product in cl. 1.4.)

The point at which the BiOS licence provisions relating to follow-on  
innovations appear to me to depart from open source principles is not  
in the breadth of what may be caught by the grantback, but in the  
fact of the grantback itself.  While the copyleft or "reciprocal"  
obligation to be found in some open source licences has been  
characterised in academic literature as a type of grantback, this  
does not appear to me to be an accurate description.  A copyleft  
licence does not establish a tit-for-tat relationship between the  
licensor and licensee.  Rather, a copyleft licensor says to the  
licensee: "Do as you have been done by".  The copyleft obligation is  
not to the licensor per se -- though he or she may be responsible for  
enforcing it -- but to the whole community of potential users of the  
follow-on innovation.  The licensor may or may not be a member of  
this community, depending on the scope of his or her activities and  
interests.  For this reason, I prefer to describe the copyleft  
obligation as a "grantforward" (as in "passing it forward") rather  
than a "grantback".

The BiOS grantback, by contrast, appears to be just that: whether or  
not the intention is noble, it still constitutes a type of privilege  
granted by the licensee to the licensor in partial consideration of  
the licence grant.  Under cll. 3 and 4 of the Licence and relevant  
provisions of the TSSS Agreement, a BiOS licensee is obliged to grant  
to CAMBIA a worldwide, non-exclusive, royalty-free licence (with the  
right to sublicense to other BiOS licensees) to Improvement Patents,  
Improvements, Technology Data, and any  Improvement Material provided  
by the licensee to CAMBIA and necessary to practise Improvements.  
(Definitions are in Licence cll. 1.7, 1.6, 1.13 and 1.11  
respectively.)  Thus, the form of the BiOS licence differs from a  
copyleft open source licence: instead of requiring licensees to  
license a subset of follow-on innovations on the same terms as those  
of the original licence, it builds the follow-on licence into the  
original licence and places the initial IP owner, CAMBIA, at the  
centre of the network as a kind of gatekeeper. While CAMBIA does have  
an obligation to make these follow-on innovations available on the  
same terms as the initial licence, that obligation extends only to  
existing BiOS licensees.  In essence, the arrangement is a licensing  
"club".  Given the restrictions on sublicensing by BiOS licensees  
(Licence cl. 2), new members can join this club only by executing a  
new licence with CAMBIA; CAMBIA is under no obligation to grant such  
a licence.

Club-type licensing arrangements are actually quite familiar in the  
biotechnology context.  Depending on the circumstances, such an  
arrangement may be good for innovation, but it seems to me that it is  
not open source. Why?  Because an arrangement in which the initial  
licensor retains a central position linked to his or her ownership of  
the seed IP and Technology is inconsistent with the freedom to fork  
the innovative process.

The freedom to create a ‘code fork’ is regarded by some as a defining  
characteristic of open source -- recall Eric Raymond's open letter to  
Sun a little while back. According to Steve Weber (author of The  
Success of Open Source), under the terms of an open source licence,  
anyone who is dissatisfied with the conduct of a project leader -- on  
technical, administrative, political or even purely personal grounds  
-- is free to take the collaborative effort in a new direction. Even  
though in practice, forking is rare, the ever-present possibility  
makes project leaders responsible to their co-developers and ensures  
that no individual or group unduly dominates the process of  
technology development. At the same time, it ensures that a promising  
technology need not be left on the shelf because of waning interest  
or incapacity on the part of an initial innovator.  This is a concern  
often expressed on this list in the form of: "What happens to the  
licence if So-and-so, the licensor, goes bankrupt or disappears in  
ten years' time?"

Of course, all else being equal, it is natural for an initial  
innovator to remain in charge of ongoing development and to act as a  
champion of the technology. However, it seems key to the open source  
approach that the initial innovator not use his or her ownership of  
the intellectual property in the initial seed technology to retain  
control over its ongoing development.

It appears to me that this freedom to fork the innovative process  
with or without the approval of the initial licensor is lacking in  
the BiOS scheme.  Why?
(a) The grantback incorporates stringent reporting requirements to  
CAMBIA (Licence cll. 3.2 and 3.4) and obliges the licensee to provide  
CAMBIA with any materials necessary to practise Improvements (Licence  
cl. 3.2).
(b) Discussions of technology data between BiOS licensees must be  
channeled through CAMBIA (TSSS Agreement, cl. 2.3).
(c) CAMBIA retains the right to impose an obligation on licensees to  
include CAMBIA’s trademarks in downstream technologies (Licence cl.  
7.9).  (This provision closely resembles one that I thought was  
removed from an early version of the open source BSD software licence  
because it imposed too much of a restriction on the freedom of  
downstream developers -- is this correct?)
(d) Under Licence cl. 7.5 and a similar provision in the TSSS  
Agreement, CAMBIA "may publish revised and/or new versions of the  
BiOS License for Plant Enabling Technologies from time to time. Such  
new versions will be similar in spirit to the present version, but  
may differ in detail to address new or newly identified issues. Each  
version will be given a distinguishing version number. ... Where  
there is... any controversy between the parties respecting the  
interpretation or application of the terms of this Agreement, the  
latest... version of the Agreement published on the BIOS website  
shall be controlling."
(e) Perhaps the most extreme instance of CAMBIA's apparent attempt to  
keep its hand on the reins of future development is in the TSSS  
Agreement, cl. 6, which reads in part:
"From time to time BiOS LICENSEE may have interest in exploring  
solutions to certain problems in its field of business. In such case  
BiOS LICENSEE shall [emphasis added] send a written project proposal  
to CAMBIA. To the extent that such project proposal is consistent  
with the terms and intent of the BIOS License, which consistency  
shall be determined by CAMBIA in its sole discretion, CAMBIA shall  
use its best efforts to... bring the proposal to the attention of  
other parties within the BIOS Initiative...".
In other words, the licensee is obliged to send a written proposal to  
CAMBIA, but CAMBIA is not obliged to do anything with it.  The effect  
(whether or not intended) is that CAMBIA gets a private preview of  
any new project.

Thank you for your input.

Janet







-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20061116/b7540433/attachment.html>