català | castellano | english home   sitemap   avís legal   crèdits   contacte  
home home

R. Joseph Anderson

The A of R&D: Archiving the Papers of Industrial Physicists

(This paper is based on the “Report of the Project to Document the History of Physicists in Industry.”  For a copy of the full report, which includes citations, contact janderso@aip.org.)

America is a uniquely business-friendly culture—it's materialistic, hard-working, and success oriented.  Calvin Coolidge, who was U.S. president in the 1920s, summed it laconically by saying that "the business of America is business."  So it seems surprising to learn that American archives and historians alike haven't done much to document and describe the history of business.  In fact, Helen Samuels, the former head of the MIT Archives, has said that most of the surviving documentation of American business is in the form of court records from antitrust suits and other legal actions.  She added that relying on these to understand business history is like using divorce records to understand the history of marriage. The final report of the 1997 symposium on “The Records of American Business” summed it up by saying that comparatively little is known about American industrial institutions because "there is very little documentation with which to work."   And if archives haven't done much to identify and preserve business records in general, this is more than doubly true for records that document industrial research and development.

American business from at least the time of Thomas Edison and Alexander Graham Bell forward has been justly famous for inventing new technologies and turning them into profitable products. In 2004 American investment in research and development peaked at over $300 billion, and two-thirds of this investment came from the private sector. Economist and biographer Sylvia Nassar has said that what's extraordinary about American industrial research and development is not the inventions that they've produced.  Every advanced culture, from the ancient Chinese forward, have inventions that they can boast about.  Instead, what's been unique about American business is sustaining product innovation over more than a hundred years.

Moving now to why I'm talking about industrial development and industrial records: about 40% of PhD physicists in the U.S. work in the private sector, and the percentage is growing among younger physicists who have entered the job market since the late 1990s.  The AIP Center for History of Physics and the Niels Bohr Library & Archives are, together, a discipline history center, and our joint mission is "To Preserve and Make Known the History of Modern Physics and Allied Sciences."  As you may know, this doesn't mean that we try to acquire the records and papers of physicists and physics programs.  In fact, we accession only the records of the American Institute of Physics and its ten Member Societies, along with materials like oral histories and photos and a few orphan collections that are especially important and that we can't place elsewhere. Instead we encourage and help archivists—mostly at universities and government repositories—to appraise and take in the historically valuable papers of their physicists and physics programs. Because physics  is a generally difficult area for archivists to document—especially for the archival generalists at most American repositories--we provide a variety of support services ranging from advice on what records and what physicists and physics programs are likely to be important, to the International Catalog of Sources or ICOS, which contains records of 9,000 archival collections at about 900 repositories worldwide, grants to help process records and papers, and other services.

And in the 1980s my predecessor, Joan Warnow Blewett, and others  began developing a new approach—called documentation research and documentation strategies—to focus on the most difficult records that science produces.  Under Joan's leadership, AIP worked first with the Department of Energy national laboratories, conducting site visits and interviews and identifying extant records.  Most important, they developed new records schedules that would allow the U.S. National Archives to document for the first time the actual work done by the bench physicists at the national laboratories.  Next, beginning around 1990 Joan and other staff conducted a ten-year study of the archives of multi-institutional collaborations.  Throughout this work, which produced detailed strategies for documenting the history of “big science,” the importance of industrial physics kept asserting itself, and we were reminded time and again that we knew even less about the records that they produced than for physicists in any other sector. As a result in the late 1990s as the Collaborations study was winding down, we began preparing a project to study industrial R&D.  Our initial investigations confirmed that archivists and historians alike know less about the records of industrial physics than any other area in our subject field. We didn't have a clear sense of who the industrial physicists were, what the nature of their work was, what records they created and used, and what archives—if any—were available to preserve the records.

As we developed the proposal for what became this study, we relied in part on the methodology that we had developed in the Collaborations study and sought advice and input from both industrial physicists and historians in the field. By the time we sent out proposals for funding, we’d created the framework for the study, which consisted of a few main activities:  we'd do site visits and interviews with senior physicists, R&D managers, and information professionals at the central laboratories of 15 of the 27 major high-technology companies that employ the majority of physicists and allied scientists in industry.  We’d use standard question sets for the interviews, which would generally take about two hours.   The interviews would be recorded and transcribed, and we’d use a relatively new qualitative software, Nvivo, to code and analyze the transcripts.  We anticipated that we’d find less documentation for industrial physicists than for their academic and government counterparts, so unlike for our earlier studies, we added a component for career length oral history interviews with especially important corporate physicists.  In other words, instead of only exploring what documentation was being created by the scientists who work in the corporate world, we'd also be creating some new documentation in the form of oral history interviews.

We used a qualitative approach—interviewing a smaller sample intensively—instead of the better known quantitative method.  Because our study was the first investigation of an unexplored field, the lengthy question-set interviews allowed us to raise research questions that are open and exploratory.  And it allowed the interviewees to describe complex situations that can’t be represented in survey questionnaires. The interviews show how scientists and companies’ record-keeping practices are influenced by personal backgrounds, company culture, management and organizational trends, and technology.

The project has had two primary staff, myself as director and part-time project archivist and a full-time project historian, plus a number of consultants and other support staff. We selected a judgment sample of 15 labs based on size, representation of different industry sectors, a variety of product mixes, and ownership.  The industrial sectors include computer hardware, aerospace/defense, energy, transportation, telecommunications/internet, advanced materials, and photonics.   All but one of the companies that we originally selected agreed to participate, and we replaced it with an alternate from among the 27 companies. The laboratories included in the study are those at IBM, Xerox, Lucent Technologies Bell Laboratories, Lockheed Martin, Exxon Mobil, Honeywell, Eastman Kodak, Corning, General Atomics, 3M, General Electric, Texas Instruments, Agilent, Raytheon and Ford.

We got underway at the beginning of 2003 and completed the study at the end of 2007, five years later.  During the course of the project we did one or more site visits at all of the 15 laboratories and conducted a total of 129 intensive question-set interviews at the labs with physicists, R&D managers, and information professionals.  This latter category of information professionals includes technical librarians, records managers, and archivists.  We have also visited and conducted interviews with staff at academic and public archives in the U.S. that collect the records of business, and visited business archives in Scotland and Germany.  We have done extensive background research throughout the study, and we've completed longer career-length interview with 16 major figures in corporate research.  And we've now completed analyzing the interviews using Nvivo software.

Our primary goals for the study  were twofold.  First we wanted to learn what documentation is created and how it's used and preserved in corporate R&D.  Second we wanted to create the foundation for developing a national documentation strategy in the U.S. for identifying, appraising and preserving historically valuable R&D records. And in order to provide a basis for understanding how to preserve R&D records, we also needed to learn why physicists choose to go into industrial research; what their career paths are like, and how corporate R&D is organized, managed, and funded. We haven't had time to address the important issue of access to records in this study, but we've developed a good sense of what records exist, how they're created, and where they're likely to end up.

I want to turn now to some of our findings, and I'll begin by saying that the results of our study do not provide easy answers to the difficult questions about how to identify records of corporate research and development that are likely to be of enduring historical value and how to preserve them permanently.  However, they do provide guidelines for understanding and documenting the work of the physicists and R&D managers employed in the 15 companies in the study and, by extension, in other large high-technology companies. They also allow us to recognize dominant themes in the changing structure of corporate research and the creation and preservation of relevant records.

The study confirms that the organization and management of industrial R&D is volatile, changing frequently in response to economic cycles, new managers and management philosophies, spin-offs and other factors.  All of the companies are trying to facilitate transfer of tacit information more effectively, and they have taken an increasingly direct role in promoting innovation (the development and introduction of new or improved products or production processes) over invention (the creation of new instruments or ideas).  These concerns have had an important impact on the work of nearly all the scientists in our study.  It means that there is strong emphasis on development instead of research and on shorter versus longer-term projects.  It also means that most of the physicists that we talked with no longer work primarily with other physicists.  Instead they work mostly with engineers, marketers, customers and others in the innovation chain. All of these factors have progressively reduced the freedom of industrial physicists and R&D managers over the past 50 years in deciding what projects to work on and how long to spend on them.  At the same time, however, physicists in industry appear to have considerable autonomy within the constraints set by their organizations, and some physicists, especially those who entered the workforce after 1990, said that working in industry is preferable to academic jobs because of easier access to project funding and the strong competition for tenure in academic institutions.

The study also confirms, first, that records keeping in industrial R&D has changed significantly over the past quarter century in response to computerization, physicists’ preferred means of sharing information, and other factors. And second, industrial R&D records are at risk and are frequently lost or destroyed. The biggest single change that we found concerns laboratory notebooks.  Only about half of the researchers we interviewed maintain lab notebooks, which were once ubiquitous in corporate laboratories and all other research venues.  They appear to have been replaced as primary documents supporting intellectual property claims in some of the laboratories by notices of invention.  And physicists report that old lab notebooks, either their own or those of colleagues, aren’t an important source of information for them.  Instead, they’re more likely to review the published literature and consult with other physicists, especially old-timers in their laboratory, to learn about earlier research that might shine light on new projects.  We realized at the beginning that most physics experiments are conducted by computer today, and we expected that paper notebooks were being replaced by electronic equivalents better able to present and store digital data.  However, only one interviewee reported that he uses an electronic notebook.  The interviewees who still maintain lab notebooks cope with the limitations of paper by pasting in screen shots of data, keeping bound copies of printouts as an appendix, etc.

The advent of computers has contributed not only to the decline in the use of notebooks.  It has speeded up the work that physicists do and created new forms of communication and reporting, like email and PowerPoint.  These new formats change the content of the information that they contain, for better or worse.  For example, several interviewees said that they believed that PowerPoint reports depended more on graphics than data, “dumbing down” and reducing the information content of the report.  A major obstacle to their long-term preservation is that unlike paper files, electronic records require prior planning and ongoing commitment in order to survive, and most of the physicists whom we interviewed do not make any effort to preserve electronic files, while those that do limit it to saving the files for a few years for personal use.  The failure to develop satisfactory solutions for preserving the electronic information that we produce every day is common to all segments, but in the companies that we’ve studied it appears that electronic media are actually reducing the amount of documentary material like laboratory notebooks and technical reports that is being produced.

There are generally no state and federal regulations regarding R&D records as there are for financial and administrative records. The library director at one of the companies said:

“what I’ve discovered is there are federal . . . and state government rules for any kind of information that has to do with money. . . . There are rules about how long you preserve this information, how long you archive, and when you destroy it. And anything to do with personnel issues. . . . But there are no federal government rules governing intellectual capital. If you want to throw it away, you may. Internally you make your own rules.”[i]

So companies are on their own in deciding what R&D records to keep and what to destroy.  Their responses vary widely.  A few, for example, continue to mandate the use of traditional laboratory notebooks and maintain effective methods for tracking them from the time that they’re issued to the individual scientist until they’re eventually filled and returned, typically to the technical library.  For these companies, however, what happens next varies.  A few preserve the records permanently, others keep them for a period of time and then destroy them, and two follow a middle course of trying to preserve the notebooks and other papers of their most distinguished researchers, who are designated as fellows.  Other companies make little or no effort to track or preserve notebooks.

The typical function and responsibilities of business archives are different from their academic and government counterparts in a number of ways.  Business archives have been traditionally less stable than technical libraries.  They may exist largely because of the interest of an individual CEO or other top manger, and they’re one of the operations that are among the most likely to be cut back or even closed during hard times.  In the companies that had both archives programs and technical libraries and that preserved some or all of the laboratory notebooks, it varied as to which took responsibility and custody of laboratory notebooks and lab reports.  On the other hand, the corporate archives that we visited typically collected product catalogs and other gray literature and administrative records, including the records of the top R&D managers. None of the libraries that we visited collected these materials.  One of the archivists said that records like notebooks and lab reports are probably safer in technical libraries than archives over the long run, because the libraries tend to more stable parts of the organization.

These findings and others described in the study's final report represent real problems and may suggest that there’s little chance of preserving a record of the work of physicists and others who are employed in corporate research and development.  However, other findings are more encouraging and, if action is taken, offer the promise of preserving some of their history.

First, a number of the industrial labs that we visited have a large core of extant science records in the form of notebooks and technical reports. This is certainly better news than what we found at some of the Department of Energy labs where we conducted our first documentation study in the 1980s and often better than what exists for academic physicists.  The bad news is that there’s no structure or tradition for making these available to researchers.  With the exception of when historians are doing an authorized history or are able to charm and negotiate their way into the technical library’s resources, these collections are closed to outsiders.  This is often, although not always, true even for materials that are many years old and have lost their intellectual property value. As I mentioned earlier, this study does not treat access to R&D records, which is an issue that must be addressed by individual companies.  However, as many of our interviewees have pointed out, intellectual property in industry has a short shelf life whether it takes the form of U.S. utility patents, which are issued for a period of 20 years and cannot be renewed, or trade secrets.

Another positive finding is that high-tech companies do find value in their history, at least for branding purposes.  Typically, major anniversaries are celebrated by extensive and sometimes relatively costly commemorations and events.  They sometimes, although not often enough, even provide an opportunity for professional histories.

And we have been surprised by the number of companies that make some kind of provision for archival programs.  Six of the 15 companies in our study have active archives that range in staff size from one to five employees.  The companies are IBM, Corning, Ford, Lucent, Agilent, and Xerox   These programs vary widely in terms of funding, mission, and the materials that they collect, but any archives program significantly increases the chances that general company information, often in the form of grey literature, will be preserved, and they provide the most likely opportunity to preserve higher level documentation of research and development programs, including records of senior managers and CEOs.  And a couple of the in-house archives collected selected lab notebooks. In addition to the six companies with archives programs, five companies--GE, 3M, Kodak, Exxon Mobil, and Texas Instruments--have transferred some business records to public or university archives.  The most recent examples, Exxon Mobil and Texas Instruments, have also made relatively generous financial contributions to help support their collections, which can be critical for archives.  Exxon Mobil provided $300,000 and Texas Instruments $200,000.  The downside of these archival transfers is that they can be hit or miss, and in the past donations to outside archives have been static, one-time efforts to unload records.  In addition, in nearly all cases they have not included R&D records.  Education and advice from professional archivists and historians may represent one way of strengthening existing archives and history programs.

I will mention just one more problem and then finish by mentioning the products that are available from this study and citing a couple of our primary recommendations.  The problem is the lack of interest and awareness by most scientists in preserving their own history.  Corporate physicists enjoy more autonomy than we expected, and senior researchers and managers have special clout within the organization.    In some cases they are able, as the IBM archivist told us ruefully, to take their records with them when they leave the company and donate them to archives of their choice.  In other cases senior staff and managers can encourage their companies to develop archives and history programs.

Now, to finish, I want to mention the primary products of our study, and they consist of three things.  First, is the draft "Final Report of the Project to Document the History of Physicists in Industry."  The 70-some page report describes the study and our recommendations in detail, and we are currently distributing the draft to all the participants in the study and to others for review and comment, and we will publish and distribute the final version this summer.   Let me add that I'll be glad to send anyone who' would like to see it a copy of the draft—just give me your email address.  And I'll be glad to receive your comments.  The second product are catalog records in our International Catalog of  Sources or ICOS and on OCLC World Cat that describe the extant collections we identified during the course of the project.  And the third project consists of the collection of question-set interviews, longer career length oral history interviews with major industrial physicists, and the project records.  We're adding these to the archives now and making them available to researchers.  In addition to providing hitherto unavailable information on individuals and companies, we hope that these records may provide the foundation for follow-up studies of industrial R&D, perhaps quantitative in nature, by others.  We are currently deciding where AIP is going from here in documenting corporate physicists.  We plan to continue to concentrate on doing oral history interviews with important corporate physicists, and we have extended the position of project historian, which was grant funded through 2007, at least through 2008 so that he can investigate a new study, perhaps of physicists as entrepreneurs.

The final report describes the best practices that we identified and contains a list of 12 recommendations.  The recommendations are intended to provide modest and realistic guidelines for preserving the historically valuable work of physicists and other scientists and engineers who work in industrial research and development.  None of them involve elaborate programs or large expenditures.  Instead they rely on initial planning before records become scattered and disorganized, standardizing record keeping practices, and taking advantage of existing programs.  And in closing, I'll mention 6 of the 12 recommendations.

Recommendations:

1.       All large for-profit organizations have formal records management programs in order to comply with federal and state regulations, and they should include and enforce R&D records as part of the program.  A small number of records that document the R&D process and policies (listed below) should be saved permanently and preserved either by the company or in cooperation with external archives.

2.       Laboratory notebooks are a traditional and important source of documenting the work that physicists and other scientists and engineers do and were once widely used in industry.  In so far as they are still used, companies should preserve the originals or duplicates permanently as part of existing records management, library, or archives programs.

3.       Like their counterparts in pharmaceuticals, high-technology companies need to develop electronic laboratory notebooks or the equivalent digital formats to replace the paper notebook.

4.       The preservation of essential R&D records depends on cooperation between a company’s scientists and engineers and its information professionals.  Physicists and science managers need to accept an active role in helping to identify and working to preserve the history of their programs.

5.       U.S. utility patents are granted for 20 years and cannot be renewed.  High-technology companies will benefit by allowing outside researchers access to intellectual property holdings whose time limits have expired.

6.       A number of companies are establishing successful partnerships with public or academic archives.  With financial support, existing archives may provide the expertise and facilities to identify and preserve historically valuable records, including the records of research and development.  The AIP Center for History is available to help arrange and facilitate such partnerships.

 Finally, and I will end here, I believe that historians need to focus on developing research projects that help explore American R&D broadly, not just concentrate on a few elite companies where sources have traditionally been available. Oral history plays a very important role, and while most of our records from the project will be available, including the longer career length oral histories that we will continue to do, they barely scratch the surface. 




bottom