By: Marc Weber
Recently, a Swiss journalist contacted us at the Computer History Museum (CHM) seeking records of the start of the Austrian .AT domain in the 1980s. I asked Elizabeth “Jake” Feinler, an Internet pioneer and ISOC Hall of Famer who gave the Museum records from the NIC (Network Information Center) that she ran for nearly 20 years. Jake is a core advisor to our Internet History Program.
The NIC collection at CHM is perhaps the biggest archive of early networking history around, nearly 350 boxes spanning the gestation of the ARPANET to the expansion of the Internet in the 1980s. While it includes printouts of naming-oriented 1970s mailing lists, such as the early years of the Namedroppers Working Group, as well as many host tables, the discussions that created the .AU domain came too late—when people no longer routinely printed out emails. There are some ancient backup tapes, but it’s not clear yet if they are readable or contain the right data. The origins of .AU are quite likely gone forever.
That’s not an isolated incident. CYCLADES was an early 1970s French network that pioneered several aspects of internetworking. It is doubtful that any of the software that made it run is still around.
Years after making the famous ARPANET interface message processors (IMPs), Bolt, Beranek, and Newman, Inc. (BBN) pioneered the router business before Cisco took it over. Few records survive.
In 1980, Tim Berners-Lee wrote Enquire Within, the hypertext information system which led directly to his later invention of the Web. The only known remnant is a smudged photocopy of the source code.
Even the NIC collection itself came a hair’s breadth from oblivion—twice. When Doug Engelbart left SRI (SRI International) and again when the NIC project ended, Jake literally saved the archives from the dumpster. She kept all 350 boxes in her own garage for many years before the Computer History Museum opened nearby.
What if you held a revolution that changed all the ways we share and transport information—from books to news to personal messages—and nobody remembered quite how it started? That’s the risk we’re facing now. The digital nature of that revolution may even help erase its origins, like a snake eating its own tail.
The fact is, a growing number of institutions do permanently preserve records of networking history. But the people who have the materials often don’t know that, or how to find and approach those repositories.
At the IETF 86 in Orlando, I chaired a packed Birds of a Feather session on Networking History. The idea for a BoF came from Jake Feinler, who helped me lead the session, and we had critical assistance from networking pioneer and IETF veteran Dave Crocker. The then-IETF chair, Russ Housley, generously gave us a 90-minute slot.
Avoiding the Dustbin of History
The history of the online world can be a big topic with lots of open-ended questions from data formats, to the environments needed to emulate software, to metadata. In fact, it can grow into a whole career, as it has for me since 1995.
But for this first IETF effort we wanted to focus on what is most urgent, and also happens to be very concrete: helping save materials that will otherwise disappear. Other issues around networking history can be worried about later. This one can’t.
Our draft charter (http://www.ietf.org/proceedings/86/slides/slides-86-history-0.txt) proposes creating directories to help match at-risk net historical materials around the world with the institutions that may preserve them. It sounds simple, and it is. But many of the people holding important items—documents, software, objects, and more—have no idea that they are of historical interest, or what kinds of institutions might accept them. Some of those materials could be moldering in your basement or garage, or in the backup tapes for your company server.
Why the IETF and ISOC? As Jake Feinler said, adapting Willie Sutton’s purported explanation about why he robbed banks, “That’s where the history is.” Precisely because this is the IETF, a lot of historically important work has been done by participants or by key people and organizations they are connected with around the world. This makes it an ideal central point from which to connect with a wide variety of historic materials and local archiving institutions.
But where standards bodies and professional associations from the W3C (World Wide Web Consortium) to the IEEE (Institute of Electrical and Electronics Engineers) to the ACM (Association for Computing Machinery) and IFIP (International Federation for Information Processing) all have explicit history efforts of one sort or another, the IETF and ISOC have had none—with one important exception. The IETF process for archiving RFCs is a unique example of how an organization can make use of its own past work. Yet this process is restricted to a sharply defined subset of materials, most notably the RFCs themselves and attendance lists (blue sheets). Beyond that, preservation can be hit or miss.
The other goal in our draft charter is to try and expand the kind of real-time collection of historical material represented by the RFC process, in which things get archived as they are produced rather than years after the fact (if at all). We hope to expand such real-time archiving both within the IETF, where we have a proposal out for doing just that, as well as to help define good practices for organizations to do so in general. This is another place the unique standards and technical expertise of the IETF can help. An example is Danny Cohen’s influential RFC 1357 on archival metadata for sharing between universities, which was repackaged as RFC 1807.
* * *
We drew nearly 200 people to a session we hoped might attract 75, helped by curiosity about such an unusual topic for the IETF. Jake Feinler opened the event, and I followed with a presentation on the overall problem and various solutions. Dave Crocker then led a brainstorming session on methods to identify materials and repositories.
We had brief presentations by institutions doing networking history work in different locations, including China, Spain, University of California, Los Angeles (UCLA), and my own Internet History Program at CHM in Silicon Valley. We ended with what was supposed to be a discussion of the charter, but became an extended Q&A session. A lot of questions were simply: what can I do with my stuff?
Closing the Loop
If you’re reading this, you are very likely engaged in building things: standards, software, machines, or other parts of what makes the online world go. That takes a huge amount of time, effort, and money. But once that work has served its immediate goal, where does it go? You may be writing poetry in the sand.
Creation and preservation can be massively asynchronous. To preserve something usually takes a tiny fraction of the effort it took to make it. Yet we so often miss that small, final step; perhaps because there’s no deadline to get pressured for, or it’s dull, or we’re not sure how to go about it, or preservation is someone else’s job.
That is why we need standards for this end of the life cycle as well. Defining good practices doesn’t need to only be about creation, but can include ensuring our hard-won knowledge gets filed away for the benefit of future pioneers.
What we hope to assemble with help from participants is a directory of local institutions around the world that permanently preserve different kinds of materials, whether software, or hardware, or papers, or video, etc. So when you wonder what to do with that box of old project papers in a storage closet, or a project shuts down, or a Web site is about to be refreshed, you’ll at least have a starting point.
A handful of the repositories listed will have networking history as an explicit part of their charter, like the Internet Archive or the Internet History Program at CHM. Others may be oriented toward tech history in general, like the London Science Museum.
But the biggest group of all—and the only game in town over much of the world—will be institutions with no tie to tech history, but who are interested in specific networking materials because of their geography or the stories they are a part of. For instance, a university library might archive the materials of a networking pioneer because he or she is faculty. Other examples include government archives, local museums, and so on.
The main criteria for all the archiving institutions is that they be stable, with adequate funding and credible, long-term preservation programs. We’ve defined a starting set of metadata for collecting this kind of information.
On the other side of the equation, we plan to use the email@example.com mailing list as one starting point for gathering information about at-risk materials. These can be in the hands of individuals, small collectors, or corporations.
Some at-risk materials are those held by unstable repositories, like underfunded museums. In fact, we already encourage small collectors and museums to come up with an archiving backup plan, to make sure their materials end up somewhere stable no matter what. There are many, many cautionary tales—huge collections sold on eBay by unpaid landlords, painstakingly developed Web sites switched off by heirs, failed museums infested with rats and mold.
Why It Matters
People readily see the history of books and printing as crucial to understanding the Enlightenment, the rise of science, and much of what made our world.
But the origins of online knowledge get pigeonholed as something niche and off to the side—sometimes even by the people whose work makes it possible.
Other fields build on their pasts—from science, to literature, to mechanical engineering, to fashion. Computing, so far, has been remarkably a-historical. New computer science Ph.D.s often graduate blithely ignorant of even the greatest pioneers of their chosen profession. Yet this perpetual sense of immaculate conception may come under pressure as the field matures.
For instance, you can justly blame the current plague of patent infringement cases on the sloppiness of the U.S. Patent Office or on the patent trolls. But if engineers were better versed in what had been done before, fewer of them might be wasting effort reinventing the wheel in the first place.
For a standards-creating organization like the IETF, having absolutely clear archives of when things were decided and under what rules is good practice in terms of liability. It also helps reduce the inevitable murk around claims and counter claims for innovations.
We’re just beginning conversations about what form networking history might eventually take within the IETF and ISOC. The new mailing list (firstname.lastname@example.org) has been active since the BoF, and we’ve continued our collaboration with Brad Fidler at UCLA, who has brought in some talented students to work on ways of bringing together historical materials with the archives that might save them. One, Jacob Ferrari, has investigated tools for collecting information on both repositories and at-risk materials, and began inventorying the holdings at UCLA.
We’ve formed initial ties with the IT History Society, which maintains central lists of all kinds from historical Websites to past software projects. They’ve agreed to create a version of their existing database of IT history organizations around the world for the needs of this group.
In the West and in some Asian countries, there are a number of science and technology museums and similar institutions you might offer historical materials to as a starting point. But in much of the rest of the world, it’s not so simple. For instance in East Africa, which has become a global center for innovation in mobile phone applications, there is no obvious institution set up to preserve that history. The same is true of South America, which has done pioneering work over several generations of computing. This is where local knowledge and contacts can be essential.
One of the networking history efforts we invited to present at the end of the BoF was the Asia Internet History Projects (http://internethistory.asia/), run by Kilnam Chong. They are in the process of setting up a point person in each major country or region, to address its particular online history. BoF instigator Jake Feinler has sent a draft proposal to the mailing list for how we might form a similar network of contacts to assist in our goals of matching materials with repositories. We already have our first regional point person in Federico Novak of Argentina, who is another student working with us through UCLA.
Besides the IETF itself, both the IT History Society and the IFIP History of Computing Working Group will be helpful in identifying and reaching partners in different regions.
You can find the draft charter and the slides we presented at http://www.ietf.org/proceedings/86/history.html. The Agenda is at https://datatracker.ietf.org/meeting/86/agenda/history/.
Please join the Networking History mailing list, and invite anybody else you think would be interested (https://www.ietf.org/mailman/listinfo/nethistory/). We’ll send the proposal for expanding a formal archiving function within the IETF to the mailing list. You can find the proposal for expanding a formal archiving function within the IETF at http://bit.ly/12SABu1.
We’re seeking volunteers from the IETF and beyond to help us identify repositories and at-risk materials, as well as to potentially work on archiving materials within the IETF itself. If you can help, please write to Jake, Dave, or myself at the below addresses.
Marc Weber, chair, email@example.com
Jake Feinler, instigator, firstname.lastname@example.org
Dave Crocker, advisor, email@example.com
About the Author
Marc Weber is founder and curator of the Internet History Program (http://www.computerhistory.org/nethistory) at the Computer History Museum in Silicon Valley. He developed the Web, Networking, and Mobile galleries of the Museum’s permanent exhibition. He pioneered Web history as a topic starting in 1995 with crucial help from the Web’s main inventor, Sir Tim Berners-Lee, and early colleagues. Weber cofounded two of the first organizations in the field. He consults on the history of the online world to companies, museums, the media, universities, and patent firms.