Read what a captured file actually contains
A saved message, HAR or ZIP can contain a different layer than its filename suggests. Follow a handoff through payload recovery and meaningful comparisons.
Start with a concrete handoff
Suppose a support handoff includes saved emails, a HAR capture and two versions of a ZIP export. The receiver needs the attachments, one captured response and the documents that changed. Comparing the entire ZIP byte stream or downloading every HAR URL does not answer that task. Recover the stored layer first, then compare its contents.
| Input | Question | Evidence to inspect |
|---|---|---|
| EML | Which attachments were saved? | Decoded payload ZIP and MIME filename map |
| HAR | Was the response body captured? | Recovered body or explicit omission |
| Two ZIP files | Which expanded documents changed? | Full paths and actual-content hashes |
| Two binary files | Which same offsets differ? | Both hex values and missing tails |
| SRT / WebVTT | Which visible cues exceed the rules? | Cue identifiers, durations and issue codes |
Keep recovery provenance beside the output
Email filenames may contain Unicode or unsafe local path characters. A safe download name is convenient but can lose the visible original spelling. Preserve the map from original message and MIME part to recovered filename. A zero-byte attachment should remain a zero-byte file, not disappear because it looks uninteresting.
For HAR, distinguish captured Base64 bytes from a Unicode text field serialized as UTF-8. Record entries with missing bodies rather than treating them as successfully recovered files. Content that is absent from the capture must come from a separate, deliberate source.
Ask which difference matters
For an archive delivery, compare expanded paths and bytes. Repacking the same documents can change timestamps, entry order and compression without changing the documents. For an opaque binary file, same-offset differences locate changed bytes but do not explain an inserted record or a file-format structure. The comparison report should make that alignment explicit.
Visible subtitle text needs a counting rule
Formatting tags can inflate a raw character count. Subtitle review should state whether tags, spaces, line breaks and Unicode code points contribute to the chosen rule. Review a flagged cue in playback with its duration and overlap context; a clean numeric report cannot establish that the translation reads naturally.
Verify the recipient’s next step
- Open one recovered attachment and verify the filename map.
- Check that a needed response body was actually recorded.
- Inspect the exact changed archive paths.
- Read a flagged subtitle cue in playback.
- Retain originals when the selected format or scope is unsupported.
Stored context makes a recovered record usable
Presentation notes belong to slides, review comments belong to selected body text, and footnote text belongs to a visible reference. Exporting the words without their relation can make a handoff ambiguous. Preserve these relationships and state logical coordinates or supported numbering rather than inventing rendered positions.
An embedded source map presents a similar distinction: a source path does not guarantee source content. Keep unavailable records beside recovered UTF-8 files so the recipient knows what the capture actually supplied. The archive map connects safe download names with source indexes; it cannot fill in absent material.
References
- Subtitle Edit reading-speed request
A specific user task source; current tool descriptions define the supported local scope.
- EML attachment extraction question
A specific user task source; current tool descriptions define the supported local scope.
- Recorded browser responses question
A specific user task source; current tool descriptions define the supported local scope.
- ZIP content comparison question
A specific user task source; current tool descriptions define the supported local scope.
- Binary comparison question
A specific user task source; current tool descriptions define the supported local scope.
- PPTX speaker notes task
Specific first-person task; current tool scope defines support.
- Word comments context task
Specific first-person task; current tool scope defines support.
- Footnote reference task
Specific first-person task; current tool scope defines support.
- Embedded source recovery task
Specific first-person task; current tool scope defines support.