Wednesday, March 12, 2008

ISCB Member Feedback Sought on Revised Software Sharing Policy Statement

International Society for Computational Biology Revised Software Sharing Policy Statement

Draft approved by the ISCB board of Directors on February 14, 2008

Open for comment from the ISCB membership and bioinformatics community

Comment period closes April 15, 2008

I. Introduction
Bioinformatics software availability is extremely important to the field of bioinformatics. The International Society for Computational Biology (ISCB) is committed to the advancement of the understanding of living systems through computation. In support of that mission, we believe that research results should be shared with the scientific community so that they can be reproduced and built upon. Scientific research may include the development of software and algorithms. Therefore, ISCB is disseminating this statement to make recommendations on software availability policies for funders of bioinformatics research, for scientific journals that publish bioinformatics research, for bioinformatics researchers, and for their employers.

This statement has been revised from the original 2002 statement, incorporating feedback from the ISCB membership.

II. Recommendations

  1. Publishers, granting organizations, employers and researchers have a responsibility to uphold the core principle of sharing methods and results. If a researcher's software is necessary to understand, reproduce and build on scientific results, then the software should be made available. This principle is imperative for peer-reviewed scientific publications, recommended policy for granting agencies, and encouraged practice wherever individuals and organizations are committed to advancing science. ISCB supports the recommendations of the National Academies of Sciences report, "Sharing Publication-Related Data and Materials: Responsibilities of Authorship in the Life Sciences."
  2. Grantors and publishers should require statements of software availability in grant proposals and research reports. These statements should clearly describe how to obtain the software, and terms of use. The statements should be specific about cost, source code availability, redistribution rights (including for derived works), user support, and any discrimination among user types.
  3. No single licensing or distribution model is appropriate for all research projects, and therefore should not be mandated by either publishers or grantors.

III. Implementation when software sharing is warranted

  1. In most cases, it is preferable to make source code available. We recommend executable versions of the software should be made available for research use to individuals at academic institutions.
  2. Open source licenses are one effective way to share software. For more information, see the definition of open source, and example licenses, at www.opensource.org.

For more information, see the previous posting which includes information about the original 2002 policy statement, member discussion, and useful links.
We invite the computational biology community provide comments on this blog, or to send email to policy@iscb.org.

[Notes added 3/14/2008:
(1) An old version of the second sentence in section III-1 was erroneously included when this was first posted on 3/13/2008; it has been fixed now.
(2) The Board is releasing this proposed language for discussion by the ISCB membership; it will not become final until after the 1-month open discussion period, whereupon the Board may revise it further in response to ISCB member feedback.]

11 comments:

Barb said...

Some additional information about the generation of this revised policy statement:

The ISCB Public Affairs Commmittee gathered feedback on the original software sharing statement from the community at ISMB 2007, and via a blog and email. The committee drafted a revised statement along with additional recommendations, and the ISCB Board made further changes to obtain the statement posted here. Additional changes may result from the feedback ISCB receives during this open comment period.

In presenting the proposed new language to the Board, the Committee also stated:

It is not sufficient to put out a policy statement. In support of the goal to increase the level of sharing in support of scientific advances, ISCB should take the following additional actions:

* Author an article educating our members.
* Contact journals that publish computational biology research.
* Contact major funders of computational biology research.
* Conduct a survey of our members (primarily to raise awareness).
* Promote the creation of standard software licenses useful in sharing of scientific research.


Some explanation about the changes in the revised statement:
* The original statement contained a definition of software levels that do not appear to have been used much (unless this use is not visible on the internet). These levels did not capture all possibilities because they included a number of different axes such as price, source code availability, and distinctions between types of users. Therefore, the levels have been removed. However, the need for clear description of licensing terms remains. We also retain a statement of the minimum level of sharing.
* The new statement emphasizes the core principle of sharing results in a way that they can be verified and extended, and relates that to ISCB's mission.
* The new statement retains the recommendation against mandating any particular license.
* The new statement has more favorable language about open source and refers the reader to the open source web site for the definition of open source and examples of licenses.
* The new statement does not mandate source code distribution, but encourages it as one means of achieving the underlying goal. Our wording is consistent with the tone in the NAS report and the HHMI software policy: we allow the possibility of executable distribution but imply that it is preferable to share source code.
* The new statement directly refers to the NAS report, significant work which was not available when the original statement was written.
* The original statement addressed only granting organizations. This statement expands the audience, in response to feedback ISCB received.

Sean Eddy said...

This is a much improved statement. I'm particularly happy to see ISCB's support now for the NAS Cech report on sharing of data, materials, and software upon publication.

You might make it more clear that executable-ony distribution is a minimal expectation, not a recommendation. In reviewing papers, I will generally downweight the importance of a paper that reports a new piece of software but only makes it available in executable form.

I think the notion of only being open with "academic researchers" should be removed, at least from the standpoint of published results. In the Cech report, George Poste had a wonderful line, that the attempted distinction between academic and corporate researchers "must be rejected as a false taxonomy". If we're going to ask that a company like Celera release its data into Genbank when it publishes in Science, it is rather hypocritical for computational biologists to refuse to share their published results to scientists who happen to work at a company. The NAS report concluded, and I agree, that there is a single scientific community as far as the open scientific literature is concerned, where everyone plays by the same rules.

Anonymous said...

I am pleased that the ISCB has taken the steps to develop a revised statement, and this version is a tremendous improvement in my opinion. I am particularly happy that this statement clearly supports sharing as a basic principle and supports the NAS statement on sharing.

There are a few points where additional clarity would be useful, as follows:

* I think it is important for II.2 to indicate the purpose of providing statements about availability. It could conclude with a sentence like, “The nature of software distribution is an appropriate criterion for consideration in the review of manuscripts and grant proposals, to assess the significance and impact the work is likely to have.”

* The text about publishers/grantors in II.3 should be made more precise, "No single licensing or distribution model is appropriate for all research projects, and therefore no single model should not be universally mandated by either publishers and grantors. However, in specific instances, it may be appropriate to mandate a specific license."

[Example: The goal of making redistributable interoperating code, e.g., for GMOD, only works effectively if there is a single license.]

My understanding from Peter Karp is that the text conveys what the original ISCB statement text was meant to indicate.

* I think the statement III.1. should delete "In most cases" and just simply say "It is preferable to make source code available." I cannot think of any case where not providing source code (under any license) is preferable for the scientific community.

Anonymous said...

Sometimes scientists publish that their software will be available, but they then break this promise in various ways. The most glaring way is never to provide the software at any time or in any form.

I'd appreciate a recommendation from ISCB about the accountability of authors after publishing statements of availability. Are authors accountable to the journals, or just to their granting agencies and their employers? For example, should readers be invited to contact journals if they encounter problems, and would journals be justified in threatening to retract articles?

Anonymous said...

I support the policy statement and I think it is a step in the right direction for the bioinformatics community.

One detail which I would like to point out is the issue of retraction of publicly available software. I couldn't see anything in the policy (maybe I missed it) which recommends that software, once available, has to be in a form that guarantees (legally) its future ability. Note that this applies equally to open source software or otherwise. Authors should not be able to later rescind the rights of a software user, once the software is made publicly available. This has many downstream implications, for example, author A makes his software available in 2001. Author B builds on author A's software and also publishes it in a high impact journal in 2003. In 2005, Author A retracts his license, effectively making all uses illegal. What does Author B do? What does the high impact journal do with the article?

Anonymous said...

I agree with the spirit of the statement and with the comments posted so far. However it's not written very well -- one of several awkwardnesses is the double negative "No model should not be mandated". Didn't participate in drafting it but I'd be happy to volunteer some edits for a version that looks cleaner.

Anonymous said...

This revision is a major improvement over the first version (which had really upset me at the time). I think, however, that it remains too timid when to comes to open-source.

As several previous comments also noted, the current wording makes it sound as if making executable code available is the recommended mode of operation for researchers, whereas it should be considered a bare minimum. (In many cases, it will be insufficient: high-performance applications, database interfaces, different platforms, all get in the way.)

It also makes it appear that the ISCB is seeking "shelter" behind the NAS and the federal funding agencies; however prestigious the NAS is, its concerns are very broad and it should be the job of the ISCB to take the lead when to comes to computational biology.

So, I would recommend that the ISCB Board take this promising revision one step further and assert clear leadership with a forthright statement: that making source code available to other researchers is the best, and sometimes the only, way to disseminate knowledge and encourage further work in computational biology.

(The first statement had appeared to consider software as just tools, that one might buy just like bench equipment; it is that, of course, but it is also, in itself, an object of research and development within the community. It is this aspect that requires source code to be available -- not biological discovery per se. But it goes without saying that further biological discoveries are enabled by new tools, including new algorithms and software.)

Finally, one minor comment about licenses: as it currently stands, the statement about licenses is a no-op -- it might as well not be made at all. ISCB could make a stronger statement, without recommending specific licenses, by stating that individuals, institutions, and businesses engaged in research in computational biology should recognize the primacy of dissemination and further research and choose licenses accordingly. This would reinforce the overall message and provide a platform for lobbying journals, funding agencies, and major businesses in favor of the most open terms possible.

AhsanShankar said...

Nice article and thanks for sharing here

Anonymous said...

Can anyone recommend the well-priced Software Deployment program for a small IT service company like mine? Does anyone use Kaseya.com or GFI.com? How do they compare to these guys I found recently: N-able N-central remote control
? What is your best take in cost vs performance among those three? I need a good advice please... Thanks in advance!

Anonymous said...

This blog is great! Thanks for your hard work on it.

Anonymous said...

Thank for you for your hard work on this siteI really appreciate it.