My colleague Leslie Johnston blogged last week about computer hardware preservation and declared a change of opinion on the subject. Her motivation came as a result of discussions at a recent Library of Congress invitational meeting, Preserving.exe: Toward a National Strategy for Preserving Software.

From Nick Montfort’s presentation, “The Trope Tank” given at the recent NDIIPP “Preserving.exe” meeting,

I attended the same meeting and also changed my opinion–but in the opposite direction. (It’s a good meeting when the ideas presented shake things up a bit).

Leslie said she now favors emulation over hardware preservation as the means for providing the computing environment to access application-dependent content. Maintaining the many, many hardware configurations just to run all the different generations of video games, for example, is not practical for most institutions. Once built an emulator is much easier to manage, as it runs on modern commodity equipment. And, since many games and other applications already have successful emulators, there is every reason to expect that approach to work broadly going forward.

All this is undeniably true. I embrace emulation as the most efficient and sensible approach to reanimating older code. Apparently Bruce Sterling, citing Leslie in Wired, does as well.

But let’s hold on a minute. Even if emulation serves the vast majority of access purposes for collecting organizations, there is undeniably more to be learned from older applications than just bringing the code to life. Original hardware allows for a much fuller recreation of the original application experience. The browser emulation of K.C. Munchkin! on textfiles.com is excellent, for example, but it is not the same as orginally played in 1981 on the Magnavox Odyssey game system, with its clunky controllers hooked up to a big console television set in the living room. That difference may be insignificant for most purposes–and it is obviously wonderful to have the emulation–but anyone interested in a richer material context for it or any other obsolete software will surely want to know how it was first integrated into people’s lives.

Yes, I know that libraries and archives should not be computer museums. I was declaring that in public 25 years ago and I still say it today–but with a little less enthusiasm.

Consider the irony if future uses skip libraries and archives in favor of computer museums to research old software.

The presentations I saw at Preserving.exe about what museums and media labs are doing with old hardware–including documenting elements beyond the computer, such as modems, controllers, connectors, custom keyboards, external drives, low-resolution CRTs and placement in homes–convinced me that preserving the code is only part of the story. Nick Montfort, director of the MIT Trope Tank, summed it up best: let’s have both emulation and the original hardware, where possible.

The “where possible” part is tricky, but I now believe that research libraries and archives should consider the selective acquisition of some small holdings of older equipment, perhaps examples of some of the more common platforms, for specialty use. That equipment is still fairly easy to get in workable condition, but that won’t always be the case; my guess is that it will get increasingly collectable, expensive and rare over time. Tacit knowledge about how to set-up and use older equipment is also perishable. Even if only a few institutions undertake some small efforts, there will a base for preserving a fuller range of the essential characteristics of software than from emulation alone.

5/31/2013: fixed typos