Computing Applications

Changes in ACM’s Revised Copyright Policy

A helpful introduction to the new additions and alterations to this pioneering policy, particularly since the advent of digital libraries.
  1. Introduction
  2. Protection for Whom?
  3. A Tour of Version 3
  4. Conclusion
  5. References
  6. Author
  7. Footnotes
  8. Sidebar: ACM URLs

This introduction to the third revision of ACM’s copyright policy updates the policy created in 1994 [5] and revised in 1995. The outstanding feature of this version is relaxed rights for authors to circulate and post their own versions of ACM-copyrighted works. It also assures the rights of readers to access a dependable archive of high-quality material at a reasonable cost. These rights were encouraged and enabled by your generous acceptance and support of ACM’s Digital Library [3].

ACM’s Publications Board expects the environment of electronic publication to continue to evolve, although we could only have guessed the developments that occurred since Version 2. For that reason, ACM continues to call this an "interim" policy; changes continue.

Though we all understand that promulgation and archiving of scientific research will soon use networked computers without any paper at all, and although one can envision policies for that future, we don’t exactly know how to get there. A major constraint on the policy at hand is to bridge present practice with the desired future, that is, to develop a strategy that will get us there.

ACM is leading technical publications toward a future where access to significant research is fast, easy, and affordable; its policies have been studied and mimicked by other societies. The past few years have seen the launch and the acceptance of ACM’s Digital Library (DL). Several commercial publishers are also launching their own online libraries, but few have so complete a corpus in place (95% of magazines, journals, and proceedings since 1991).

Furthermore, international copyright law has recognized the potential of computing and communications. In December 1996, the World Intellectual Property Organization (WIPO) met in Geneva for this decade’s revision of the international copyright treaty, where European countries advanced controversial sui generis protections reserved to builders of databases that extended even to formerly non-copyrightable information [9]. Even while this policy was developing, ACM through its U.S. Public Policy Committee (USACM), and sister societies informed the U.S. Congress of its reservations on similar bills.1 The scholarly societies parried the more objectionable sui generis provisions, which eventually were dropped from the U.S. enabling legislation. They have already been reintroduced, however, in the new Congress.

It is important to keep an international perspective on copyright policy because copyrights are more effective than contracts in many parts of the world. (Most of you are probably accustomed to a reversed situation, where contracts are used to override local copyright law.) A civil contract can be impractical to enforce in the same venue where a copyright violation is a criminal offense. Commercial interests have elevated copyright matters to the level of international treaties, like WIPO’s draft, effective even where contractual protections are notoriously unenforceable.

ACM continues to value its archive and seeks to secure it under international copyright law. Not all scholarly societies do this; some encourage but do not demand copyright transfer [8], while others not only demand transfer, but also preclude prepublication [2]. Others argue that authors should retain all copyrights on research material [1]. ACM takes a middle road: it asks for copyright transfer in order to protect our archive, but it also allows authors liberal rights to circulate their version of their own works, and grants liberal rights to readers; for example, classroom use.

Back to Top

Protection for Whom?

The protections available from copyright are distributed to all participants in ACM publication: readers, authors, and ACM, itself.

Cynics might expect ACM’s policy to be written to protect only the institution called ACM. They would overlook the fundamental difference between ACM and commercial publishers: ACM’s mission targets professional and public interests, not shareholders.

The Publications Board debated at length policies that would allow its authors-members to retain copyright with ACM publishing under license. We agreed an appropriate contract between the author and ACM could arrange foreseeable rights anywhere along the spectrum between author and publisher, leaving residual rights with either.

But leaving residual copyright with the many, many authors weakens the interests of important third parties: the now-and-future readers. As an educational and scientific organization, ACM represents them as well. That is, ACM is not merely a publisher—organized independently of authors and readers—but it is a membership cooperative that is answerable to both.

How best to serve the interests of the publisher, authors, and readers? ACM continues its general policy of holding copyright to works it publishes in any medium. (Exceptions are provided; see Section 2.1.)

Interests of ACM as publisher. ACM incurs costs in producing its journals and magazines. In ink-on-paper technology, fixed costs include the editorial process, copyediting, formatting, and press set-up. Most of these processes still occur, though differently, under digital publishing, but production costs decrease once paper disappears.

ACM’s DL, however, has a new postpublication cost that does not end when an issue is mailed. It must be maintained on a day-to-day basis, and sustained through future changes in technology. So the commitment to the DL includes the liability of maintaining it—in a curatorial sense—forever. If the DL is to survive, ACM must be able to recover these costs of production and maintenance.

Moreover, the computing community differs from other sciences and technologies in several ways that affect ACM’s finances. Unlike physics, not all our colleagues join ACM or any other computing society. Blame this on our collectively independent spirit or a love for anarchy, but many members join ACM only for its benefits, rather than out of some sense of obligation. Unlike mathematics, the bulk of ACM’s technology does have commercial value. ACM uses the financial leverage from the latter fact to solve the former problem: to sell memberships in a hesitant market. On balance, ACM looks like many other societies.

Interests of ACM authors. Authors want their works to be accessible widely, easily, and cheaply. If there must be fees they should be reasonable, reflecting production expenses, rather than profit. Accessibility requires the work be indexed, researchable, and responsive to tools, including those yet to be invented.

A work’s appearance in an ACM publication carries an endorsement of quality. The final product reflects improvements from referees, editors, and copy editors. Usually the appearance changes to make it more appropriate for a permanent archive.

ACM journals, magazines, and proceedings publish articles of high quality, and so have a strong base of readers. They are an effective and respected outlet for young authors, building their reputations. That is, ACM offers a valuable imprimatur to the authors, as well as to the material it publishes.

Once a work appears, ACM will protect its integrity from those who would convert or distort it [4]. ACM will seek legal protection for its copyrights, even when the authors have physical or financial constraints that preclude action on their own behalf.

Interest of ACM readers. Readers also benefit from ACM’s defense of the integrity of the archive. They need important works to remain accessible in new contexts and in the new technologies as they arise, even after an author is no longer available to authorize the change.

They need a reliable scheme of unique citations that survives generations of systems and migrations of authors across the planet. If a single work were to enter the archive with multiple citations, then future researchers would be distracted by minor variants on the same theme. When a poor alias becomes a dead link, some future reader might wrongly assume the work has been lost. Even if the reader realizes that two citations refer to the same work, she must still decide which one is authoritative; if she, instead, believes them to be independent, she may confuse other readers by citing both.

Readers need to find a work promptly, easily, and cheaply, and to have confidence the work is of high quality before they pay fees for access. They need to have access to prompt permissions should their use require them. ACM readers, in particular, often need permission for reuse in classroom contexts. Many need continuing access via hard-copy interlibrary loan. Both are granted a priori.

Back to Top

A Tour of Version 3

With that background to the underlying philosophy, let’s take a quick tour of the current changes to the ACM Copyright Policy. There are also minor differences throughout.

Reader and author interests. Section 1.2 newly lists the readers’ and authors’ interests that ACM intends to protect. The interests of future readers, in particular, set ACM’s copyright policy apart from recent proposals for author-retained or public-domain copyrights. (An overview of ACM’s concerns for others is implicit in this list.)

Definitive versions and revisions. Section 2.4 changes in major ways. First, the definitions of the "definitive version" is expanded here and in the Glossary. For ink-on-paper technology, it corresponds to the images that now appear in print journals and in ACM’s DL.

Second, now that the ACM DL is operational, the interim permission for authors to post definitive versions of their own works on their local servers has been withdrawn. In its place is new and broad permission for authors to maintain their own, nondefinitive versions of their own work at any point in the work’s lifetime (probably preprints but also any revision, even after publication by ACM.) Reasonable endorsements are necessary, however, after the copyright has been transferred to ACM.

CoRR. Section 2.6 is entirely new. It outlines the Computing Research Repository (CoRR), of which ACM is founding sponsor [7]. CoRR is a lightly moderated, online repository where researchers from the whole field of computing can submit their reports, search for research results, and follow developments in selected areas.

As an experiment, ACM grants permission for authors to deposit their author versions in CoRR for the next two years. That is, authors’ versions of papers, including before submission or after copyright transfer to ACM, may be deposited in CoRR permanently, but not ACM’s definitive versions.

Of course, the definitive version, when it exists, is the one that should be universally cited in order to protect the citation trail, which is part of the integrity of our archive. Folks will, no doubt, pass CoRR coordinates from colleague to colleague, supplementing an ACM definitive citation, but the CoRR citation should not become the citation when there also is a definitive version in ACM’s DL. Publishing only a CoRR citation is a disservice to the future.

ACM is bold in granting this temporary permission. A commercial publisher might not take this risk, but the Association does it because:

  • ACM feels that appropriate use of CoRR will improve the quality of (eventual) definitive works in the ACM DL. If it improves ACM’s product, then CoRR is good.
  • ACM already delivers cost-effective, high-quality products, so its readers ought not feel much pressure to prefer CoRR over its DL.
  • ACM does not expect CoRR to offer competition in its journals’ market: its product is a permanent, complete library of polished works, with indexing and search tools that integrate into the global archive.
  • Only authors themselves can deposit into CoRR. This fact, and Section 5.2’s restrictions on tables of contents, prevent assembly of complete volumes in CoRR.
  • Such repositories will be used for preprints in some communities, whether or not ACM grants permission. ACM also represents the authors who use CoRR appropriately.
  • ACM is a founding sponsor of CoRR.

This experiment may be too risky for a profit-driven publisher that has higher costs and more income at risk. In contrast, ACM sees this experiment as another opportunity to serve its computing community. We hope that authors who benefit from CoRR review will subsequently submit their manuscripts to an ACM journal in order to strengthen ACM’s DL.

Renumbering. Section 2.7 and 2.8 were formerly numbered 2.6 and 2.7.

Author retained rights. Section 3.1 has been revised to allow authors’ postings of their own author versions as a retained right. Also listed here are the reasonable endorsements required on such a posting; they tie the archive together. This is a large concession and, probably, the highlight of this policy both for authors and contemporary readers.

In addition, the right of an employer who originally owned copyright to continue to distribute a work within its organization is here incorporated into the copyright policy.2

Distributions from non-ACM servers. Section 5.3 is changed in two ways: The allowable postings are extended to include all author versions, not just preprints and revisions, and the requirement for a server notice has been eliminated.

Back to Top


In granting liberal rights to its authors, broad permission to its readers, and accessible fees to its subscribers, ACM assumes a financial risk. It is possible that ACM’s base of subscriptions to ink-on-paper journals might simply vaporize in a cloud of Internet packets. On the other hand, the irresistible drive toward electronic publication presents every ink-on-paper publisher with an unavoidable risk, whether or not it plans for this change. Particularly in the area of computing research and development, ACM expects authors and readers to make a rapid transition. Therefore, the Association would assume an even greater risk by doing nothing.

The success of this policy depends on authors, readers, and ACM. Authors need to accept ACM’s copyright policy in preference to those of commercial publishers, and choose to submit to ACM, and to cite its definitive versions whenever they write. Readers have to support ACM electronic publication by joining ACM, by subscribing to its DL—personally or through their employers—and by participating in ACM meetings.

ACM has its own challenge to make this Copyright Policy work. It must field more products to take advantage of the economy of scale implicit in the DL. Not all of these will be traditional print media; not all of these will be in fields already identified. Finally, ACM must continue to develop value-added products on top of the DL: tools that enhance its use, or that are enabled by the DL as a foundation.

No, we are not sure that ACM’s DL and this revised copyright policy will unfold smoothly into the future. We are sure that refereed journals and collections will survive to define our field. As the Internet fills with tripe, readers will seek sources that bear ACM’s imprimatur, assuring quality. To protect that standard, however, we must revise our habits every now and then.

Back to Top

Back to Top

Back to Top

Back to Top

    1. Bachrach, S., et al. Who should own scientific papers? Science 281 (Sept. 4, 1998), 1459–1460.

    2. Bloom, F.E. The rightness of copyright. Science 281 (Sept. 4, 1998), 1451.

    3. Denning, P.J. The ACM digital library goes live. Comm. ACM 40, 7 (July 1997), 28–29;

    4. Denning, P.J. Plagiarism in the Web. Comm. ACM 38, 12 (Dec. 1995), 29;

    5. Denning, P.J., and Rous, B. The ACM electronic publishing plan. Comm. ACM 38, 4 (Apr. 1995), 97–109;

    6. Gilpin, K.N. Concerns about an aggressive publishing giant. NY Times 147, 51021 (Dec. 29, 1997), D2.

    7. Join the New Computing Research Repository. Comm. ACM 41, 11 (Nov. 1998), 35;

    8. Knapp, A.W. Copyright policies. Notices of he AMS 46, 1 (Jan. 1999), 4;

    9. Liedes, J. Copyright: Evolution, not revolution. Science 276 (Apr. 1997), 223–224.

    1Like this policy, ACM is international but it has a peculiarly U.S. context for many activities; for example; its production process.

    2This isn't really a change; the existing copyright-transfer agreement already provided that right.

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More