What have we learned so far? We're damned if we digitize and damned if we don't.
As I hope you've seen from this week's presentations, storage is not a long-term option.
Migration is possible but painful and has to be applied case-by-case.
In a blog entry summarizing the conclusions of the 2013 Preserving.exe conference, former Library of Congress staffer Leslie Johnston declared, "I was convinced this week that emulation may serve our needs better than hardware."
You'll recall from Dragan's presentation that an emulator is a computer programs that fools original code into assuming that it is still running on its original equipment. For institutions tasked with ensuring the survival of digital culture, emulation is the new kid on the block.
What made this pronouncement anachronistic, if welcome, was that the technique of emulation had already been around for decades without most curators or librarians paying it any mind.
In the early 2000s, recommending that collecting institutions preserve their work using emulation was like recommending they power their climate control systems with nuclear fusion. When they heard the technique explained by some gearhead at a conference or by their own IT department, many archivists and conservators just scratched their heads. Emulators seemed far too complicated and nerdy, compared to the apparently simpler task of mothballing the computer guts necessary to run Apple II programs or Voyager CD-ROMs.
It took till 2013 for nationally prominent archivists like Johnston to realize "we cannot all become museums of computer hardware"--that is, maintain a suite of working floppy drives, CRT monitors, and other hardware has-beens.
So what happened in those intervening years that turned this bastard stepchild into the favorite son of the cultural heritage community? In short, video games happened. Or rather they were born, then died, then born again.
While museum conservators and librarians were busily archiving digital files on CDs for the long term--"long term" in the case of self-pressed CDs meaning half a decade--their eight-year-old cousins were wielding a powerful technology that in theory could extend the life of bits into the indefinite future.
So it should not be a surprise that video games were the catalyst for museums to accept this new technique, which was both more geeky and more folksy than what institutions of high culture were accustomed to putting in their white exhibition halls.
What will be lost when emulation is the only solution for reviving out-of-date software? The 2004 Guggenheim exhibition Seeing Double: Emulation in Theory and Practice set out to find out.
The works appeared in pairs (and sometimes triples)--the first in each set running in vintage hardware and the others a migrated or emulated version of that original. The show's premise was that viewers would have to decide whether the pairs exhibited side-by-side constituted the same work or different works, based on the perceived fidelity of the re-created version.
An audience survey conducted at the exhibition concluded that in most cases emulation produced copies that were spiritually, if not physically, equivalent to the originals.
Yet once again generational age played a role in the relative acceptance of emulation. Respondents with the least digital savvy--correlated with older viewers--tended to prefer re-creations that aped not just the behavior but also the exact look and feel of the original.
For example, older viewers gave high marks to the re-creation of The Erl King, a 1985 video installation by Grahame Weinbren and Roberta Friedman, for which the Guggenheim emulated not just the source code but also the appearance of its fiberboard kiosk and bulky CRT monitor (in this case by embedding a flat screen inside a wooden box). The re-creation looked so similar to the original, in fact, that the curators had to outfit both versions with Plexiglas windows so museumgoers could see that the second had different guts--the ancient Sony SMC-70 computer and its rack of analog laserdisc players replaced with a single Linux box.
Older viewers were less approving of the re-creation of John F. Simon, Jr's 1999 Color Panel, which ran the same code but was displayed on a somewhat larger and brighter 2004 laptop in place of the original PowerBook 280c. Younger viewers who had grown up playing 8-bit games on 24-bit screens, by contrast, accepted the changes in look and feel as long as the code behaved the same. Neither the Erl King nor Color Panel is a game, yet it is tempting to ascribe younger viewers' tolerance of changes introduced by emulation to their years of experience drinking old digital vintages from new bottles. This difference matters because we are not used to looking to youth for expertise, or for that matter to popular pastimes as a model for preserving elite culture.
To neglect the opinions of older viewers, however, would be to spurn those who have actually felt an Atari joystick or Gameboy in their hands rather than wearing out their fingertips pressing the WASD keys. True to its mission, Seeing Double surveyed creators familiar with the original hardware about what was lost in translation. Joan Heemskerk of jodi.org authorized the curators to show her modded game on both the original ZX Spectrum computer from the 1980s and a Spectrum emulator running in Windows XP, but lamented how the emulated version lost many of the particularities that jodi admired about the original hardware. For example, seeing the game on a Windows PC gives the viewer no clue that Spectrum games once ran off audiotape rather than a disk drive. In her presentation for the associated symposium "Echoes of Art: Emulation as Preservation Strategy," Heemskerk also described how the crisp look and antiseptic feel of a flat screen differs from the warm buzz of a cathode ray tube:
Tempest: http://youtu.be/AMto2HJJSSA?t=20s
Now, some game fans have created software deliberately designed to mimic the look of CRT screens, or even earlier vector-based displays.
As of the mid 2010s, too few official custodians of culture have invested in emulation. This may not be surprising given the limited resources and broad mandate of today's collecting institutions. Emulating The Erl King was only possible thanks to a "perfect storm" that brought together a dedicated museum, a team of experts, an eager sponsor, and a willing artist with access to the original source material. Renewing The Erl King was a heroic effort, but heroism, as preservation expert Richard Rinehart reminds us, can't rescue everything because by definition it is unrepeatable.
Fortunately, for common platforms, emulation can be repeatable, and even better, extensible. Some emulators only emulate an operating system--masquerading as, say, the NES cartridge system while in reality running as an application in Mac OS. More sophisticated emulators, however, which simulate an actual computer chip such as the Pentium or Power PC, can be nested together to transition from one platform to another. Computer scientist and emulation champion Jeff Rothenberg is fond of demonstrating this principle by running a Windows-based emulator for the EDSAC, a forerunner of modern computers built in 1949, inside the 2005 Windows emulator VirtualPC, which runs on the Macintosh. This daisy chain of emulators, in which a Mac impersonates a Windows machine, which in turn impersonates the EDSAC, makes it possible to view a computing history of fifty years on a single device.
Fifty years from now, when the Macintosh as we know it will be long dead, a preservationist could write an emulator to simulate a 2005 Macintosh on the prevailing platform of 2065. At that point she will be able to embed Rothenberg’s daisy chain inside this new emulator, and so on into the future. By overcoming the dependence of an emulator on its parent platform, this stratagem resolves the paradox whereby tools designed to liberate artifacts from obsolescence eventually fall prey to obsolescence themselves.
How do we train digital conservators and archivists in emulation? We do this at the Digital Curation graduate program at the University of Maine by asking our students to emulate common computing platforms, starting with a simple vintage game of their choice. The extensive work experience that many of these students bring to our program doesn't necessarily help them with this; in a recent case, an archivist at a major private collection who had held a high position at the Library of Congress was struggling with this task. So I asked some of my 2nd-year undergraduates to create a screencast tutorial, and then she understood.
We may not be used to relying on 20-somethings to teach 40-somethings how to run national archives, but I believe we should be looking to amateurs for new ideas and solutions. We ignore the insights of Generation Emulation at our peril.
If this thought frightens you professionals in the audience, then I recommend you avoid my final keynote on Wednesday at the end of the day on the topic of "Unreliable Archivists."