pifs
Data-free FUSE filesystem backed by π
TLDR
SYNOPSIS
pifs -o mdd=metadata-directory [mountpoint]
DESCRIPTION
pifs (πfs) is a FUSE userspace filesystem that stores file data by locating each byte's position within the digits of π rather than on disk. Because π is conjectured to be a normal number, every finite byte sequence is theoretically present somewhere in its expansion; πfs records only the index and length needed to retrieve each byte.Files are split into individual bytes, each looked up in π using the Bailey–Borwein–Plouffe formula. Metadata (paths and offsets) is written to the directory given by mdd=. The filesystem is largely a proof-of-concept: storing even small files is extremely slow because locating arbitrary digit sequences in π is computationally expensive.Requires libfuse and a C build toolchain. Install build dependencies on Debian-based systems with autotools-dev, automake, and libfuse-dev.
PARAMETERS
-o mdd=path
Directory where πfs stores metadata (filenames and byte offsets in π).mountpoint
Directory where the filesystem is mounted.
CAVEATS
Lookup performance is impractical for real workloads; the project README notes that storing a 400-line text file can take several minutes. If metadata is lost, file locations cannot be recovered even though the data still exists in π. For a newer related project, see inferencefs.
HISTORY
πfs was created by Philip L. as a humorous exploration of the idea that all possible data already exists within π, inspired by a 2001 observation that a normal π would contain every finite file. The repository has been widely shared since its initial release.
SEE ALSO
fusermount(1), mount(8)
