Geoffrey Brown of the Indiana University Department of Computer Science has a nice presentation available online which talks about the CIC Floppy Disk Project, and which along the way argues the case for emulation. The CIC FDP is intended to make publications deposited with federal US libraries available via FTP over the Web. In many cases this means saving not just the contents of the floppy, but also the applications needed to make the contents readable. One of his diagrams makes the point that emulation results in two separate repositories, one for documents and the other for software.

The project doesn’t appear to be strict emulation, in that some leeway is allowed. For instance, slide no. 16 bullets the software necessary for the project, one of which is Windows 98, even though “most disks were for msdos, Win 3.1”. I take that to mean that while most floppies were created on Win 3.1, they work just as well in Win 98, so let’s use Win 98 instead. Strict emulation theory probably isn’t too happy with that.

Slide 21 is the most interesting as it contains a handy summary of the problems of migration:

  • Loss of information (e.g. word edits)
  • Loss of fidelity (e.g. “WordPerfect to Word isn’t very good”). WordPerfect is one of the apps listed earlier as necessary for their emulation.
  • Loss of authenticity: users of migrated document need access to the original to verify authenticity [AA: but this depends on how you define authenticity, surely?]
  • Not always technically possible ( e.g. closed proprietary formats)
  • Not always practically feasible (e.g. costs may be too high)
  • Emulation may necessary anyway to enable migration.