Keeping stuff: How to preserve course papers despite technological change

Rainbows

Shortly after I arrived at Grinnell, many of us began using personal computers, products of the Digital Equipment Corporation, named Rainbows. Initially, these computers had no hard-disk storage at all. They had two drives for five-inch floppies (capacity 396 kilobytes each), and they could be used as terminals to the College's main academic computer, which did have disk storage. When I got a Rainbow to use at home, I moved the original file containing the 1983 handout from the College's computer to a floppy disk; a few years later, I copied it back from the floppy disk to the College's main academic computer (a newer one) and from there to a file on MathLAN.

I can tell that the file was on a Rainbow floppy disk at some point, because in the course of the moves the file was slightly distorted by the transfer software. The Rainbow stored files in units of 512 bytes. When it was asked to store a text file of which the length was not exactly a multiple of 512, it added characters at the end -- control-Z characters -- to pad the file to the required length. The file-transfer program that recovered the handout from the Rainbow brought the padding along with it, so now the ``original'' file has twenty-six control-Z characters at the end. If I now print the file, they show up at the end of the output.

The Rainbows had some other limitations as well. The EDT and RUNOFF programs were not available for Rainbows, for example. Fortunately, a generous DEC programmer named Anker Berg-Sonne had, in his free time, constructed an EDT work-alike called SEDT, which did run on Rainbows, and he made it available without charge. Computer Services also managed to round up a really bad knock-off of RUNOFF, which they tried to persuade people to use, though it wasn't really compatible with RUNOFF and also was extremely slow. I preferred another formatter, SROF, which I myself had adapted, mostly from source code published in a book called Software tools in Pascal, by Brian Kernighan and P. J. Plauger.

Eventually I put together a single five-inch floppy disk that contained the files that the Rainbow needed to boot, SEDT, SROF, and a spelling checker that I wrote called SPELLER. With that disk in one drive and the files to be edited on a second disk in the other drive, one could get a lot of writing done quickly and efficiently, particularly by comparison with the ponderous pace of the technology of manuscript and typewriters.