Plan a reproducible file recovery review

Download the review materials

• Synthetic files and bilingual worksheet (ZIP)
• Expected file sizes and SHA-256 hashes
• Blank review worksheet
• Method limits and reuse permission

ZIP SHA-256: d87266694f54b68334d4dd67bbfad65e9275e2fadee3e85b8ae170cf8edfd8c3

1. Define the question

Decide whether you are testing filesystem recovery, a raw file scan or preservation of a readable image. These are different tasks. The supplied UTF-8 text, JSON and HTML files suit byte comparisons and filesystem recovery; they are not a representative signature-carving corpus or proof of performance on photographs, damaged media or every failure condition.

2. Preserve a known starting point

Keep the original kit untouched on separate storage. Use a disposable test volume or test image, record its format and state, and confirm the sample hashes before making a deliberate change. Do not experiment on valuable files or the only copy of data.

3. Record the conditions

Write down the app and installed version, Windows build, storage type, filesystem, encryption, TRIM state if known, time since deletion and subsequent writes. If a condition is unknown, mark it unknown. Use the same preserved starting state when comparing tools.

4. Recover to another destination

Follow the chosen product’s documented workflow and keep output away from the test source. Record settings, failures, elapsed time and output paths. A preview or an item in a result list is not the same as a complete recovered file.

5. Compare the bytes independently

Compare each recovered file’s byte length and SHA-256 with the manifest. On Windows, the built-in Get-FileHash -Algorithm SHA256 -LiteralPath "path-to-recovered-file" command can hash a file without changing it. Record exact matches, partial files and missing files separately. Also record whether the original names, paths and metadata survived.

6. Publish the denominator and the limits

Report, for example, “five synthetic files attempted, three exact byte matches, one partial and one missing” only if that is what your own test observed. Do not turn a small synthetic run into a general success percentage. An SSD’s TRIM behavior, later writes, encryption keys and media failure can make another case entirely different.

Use and attribution

You may reuse and modify the original fixtures and worksheet for independent reviews, including critical reviews. Attribution and a link to this guide are welcome but optional. No favorable coverage is required. Do not publish private source images or customer data with your results.

ASL Recovery Center’s published scope · Recovery workflow and limits · Other reviewer resources