File collection

A file collection is a set of computer files that is deliberately maintained by someone. In today's computing, most non-personal files can be considered temporary: they have been downloaded from somewhere and can always be redownloaded if needed again. The practice of actual file collection is becoming ever more marginal, but it has a place in permacomputing for increasing resilience and reducing network dependency.

Before the ubiquity of the broadband Internet, users of personal computers usually had files of all the software they used (and usually a lot of software they never used), often on physical floppies or CDs. These collections were cared for, and even the decision to delete a rarely played PD game could be painful. While people had their private file collections, there were also public file collections, such as BBSes, which often served as repositories of commonly needed PD software and much more.

While broadband networking is probably the most important reason for the marginalization of file collection, another reason may be "obsolescence thinking" that assumes files to get bad if they are not constantly "updated". Keeping a computer offline for a long period of time often brings up large batches of automatic updates; experiences like this may contribute to the illusion of "file rot".

Today, many people have small personal servers or websites, but they are rarely used to share any files other than those directly related to the maintainer. However, they could also easily serve larger collections of files, including copies of all kinds of online resources the maintainer considers important. A lot of resources are available with licences that allow unlimited redistribution, but this right to redistribution seems to be quite underused.

In a file collection culture, nothing that is published is supposed to only stay in one place. A resource stays in one place only if it is too obscure or uninteresting to anyone else, maybe even the author. Even the kind of files that in the WWW culture could be called blog posts or social-media posts get spread to multiple places if they are considered of any interest at all. From the permacomputing point of view, a file that is only available on one server has a hard dependency on that server, and hard dependencies are supposed to be avoided.

In a permacomputing world, servers that host file collections would be just as common as public libraries. People would primarily use the servers that are geographically close to them. They would contain all the commonly used software and documentation (and their complete dependencies) along with large amounts of other freely distributable media (books, entertainment, reference databases such as Wikipedia, etc.)