Obsolete Data Formats

Microsoft Word or .pdf files are not easily read today except by machines that are not commonly built to last more than a few years. Some of today’s programming languages may be supported 100 years hence but where will we find the sources for code that reads .pdf files? Where will we find the meanings of the system calls that produce pixels on the screen? One sort of commitment would be to produce upon future demand 2D arrays of 24 bit square pixels (3 colors). Perhaps the resolution would be fixed, perhaps not. Preserved secure hashes of the pixel array might add teeth to such commitments. Of course the arrays might themselves be stored or compressed by software simpler then necessary to read .pdf files.

Another approach would be to produce a .pfd reader in some simple machine language whose definition fit in 10 pages. (The IBM 701 manual was 19 pages I recall.) Perhaps some compiler such as gcc could be modified to compile into this language. This would support preservation of more general sorts of software.

Bochs might be part of the solution but I can imagine that the art of running large C++ programs might be lost before 100 years. It is plausible that bochs may survive in the spirit of antiquarian projects.