Unlocking History from Magnetic Tape: Tape Ark Recovers 1970 Aboriginal Population Register

Tape Ark has successfully decoded a historically significant dataset for National Archives of Australia recovering information from magnetic tapes dating back more than fifty years.

The project focussed on the 1970 Aboriginal Population Register for the Northern Territory, a dataset of immense cultural and historical value that had been inaccessible for decades due to obsolete technology and unknown data encoding.

A Dataset Frozen in the 1970s

The register was originally stored on 7-track magnetic tape, a format widely used by large mainframe computers during the 1960s and early 1970s.

In 1997 the data had been migrated to 9-track magnetic tape, but by 2004, when the National Archives undertook a recovery program for obsolete digital carriers, specialists were only able to extract the raw bitstream—the original binary data.

While the files were recovered, they remained unusable.

The challenge was simple to state but difficult to solve:

  • The character encoding was unknown
  • The data structure could not be identified
  • Previous recovery attempts could not interpret the data

For years the dataset remained locked away, technically recovered but effectively unreadable.

A Digital Archaeology Challenge

When National Archives approached Tape Ark, they provided the original recovered files along with documentation from earlier attempts to interpret the data.

Initial clues suggested the dataset may have originated on a Control Data Corporation CDC 6400 system, using FORTRAN and potentially a 6-bit character encoding common in early mainframe environments.

Tape Ark’s first analysis quickly confirmed several important characteristics:

  • The data originated from 7-track tape
  • The dataset used 6-bit character encoding
  • The files contained structured blocks, likely representing a database backup

However, the exact encoding scheme remained unknown. Early attempts to interpret the data using known historical encodings—including SQUOZE (IBM) and RADIX-50 (DEC)—failed to reveal readable text.

Reverse Engineering the Encoding

Tape Ark’s engineers then began a deeper investigation into the structure of the raw data.

“When I first look at a file, the task of decoding the data can seem overwhelming. The first thing I do is search for patterns—whether it’s a single repeated character or groups of bytes appearing at regular intervals. Finding those patterns is often the first step toward understanding how the entire file is structured,” said Neil Richardson, Tape Ark Software Developer.

The team analyzed statistical distributions of characters and identified repeating block headers approximately every 210 bytes, suggesting a structured record format rather than a simple flat file.

By studying the bit patterns and examining how characters were distributed across the file, the team was able to determine:

  • The two most significant bits in each byte were consistently zero
  • The original 7-track parity bit had not been preserved
  • The remaining 6 bits contained the original character data

Using this information, Tape Ark developed custom decoding tools to systematically test different encoding possibilities and reconstruct the character mapping. However, a major complication remained: the original software used to write the tapes no longer exists. Without the programs that created the data, there was no reference point for how the files should be interpreted or even what the final output should look like. This meant the team could not simply validate results against known outputs or documentation. Instead, the decoding process required a form of digital detective work, analysing patterns in the raw data, testing hypotheses about possible encoding schemes, and gradually reconstructing the structure of the records without knowing exactly what the finished dataset was supposed to contain. Through iterative analysis of the data patterns and expected values within the dataset—such as Northern Territory place names—the team gradually reconstructed the correct interpretation of the encoded characters.

The Result: Data Recovered

The final decoded dataset revealed a structured database consistent with the original population register.

When the results were presented to National Archives team, the response was immediate. James Doig, Assistant Director, Digital Archives Preservation and Research commented,

“This is a really important record! The data looks great – a lot more than I had originally expected – and it matches the data structure we were expecting. Impressive detective work team Tape Ark!”

For the first time in decades, the contents of the Aboriginal Population Register could be accessed again.

NAA Job Process 1970 Aboriginal Population Register

Preserving Cultural and Digital Heritage

Projects like this highlight a growing challenge in digital preservation: data may physically survive, but without the knowledge needed to interpret it, it remains effectively lost.

Magnetic tapes, early computer systems, and proprietary data formats continue to pose barriers for archives around the world.

By combining expertise in legacy storage media with deep knowledge of historical computing systems, Tape Ark helps organizations recover information that would otherwise remain inaccessible.

For National Archives of Australia, the project represents more than a technical success, it ensures an important piece of Australia’s cultural and administrative history can be preserved and studied by future generations.