Chapter 2 — Key Takeaways
The big idea
Deleting a file removes the pointer, not the data. When you delete a file or quick-format a volume, the operating system invalidates the file's directory/MFT entry and marks its clusters "free" — but the actual bits, the magnetized regions or charged cells, are left physically in place. The file is orphaned, not erased. Data is genuinely destroyed only when something overwrites it: ordinary reuse of the free space, a deliberate secure wipe, or — on an SSD — TRIM plus garbage collection physically erasing the flash. The gap between marked free and actually overwritten is the window every recovery and most investigations live inside. On magnetic media that window can be months; on a TRIM-enabled SSD it can be seconds. This one idea is why both of your disciplines exist.
The translation stack
You must move fluidly up and down this ladder; each layer thinks in a different unit.
BIT -> BYTE -> SECTOR -> CLUSTER -> FILE
1 8 bits 512 or e.g. 8 one or more
=1 byte 4096 B sectors clusters
(256 (drive's (file (what the
values) unit) system's user sees)
unit)
- The file system thinks in clusters; the drive thinks in sectors; the hardware thinks in magnetic domains or flash pages. You translate between all of them.
- The bridge calculation, committed to memory:
byte offset = sector × sector size(and its inverse,sector = byte offset ÷ sector size).
Numbers and signatures worth knowing cold
| Fact | Value |
|---|---|
| Bits in a byte / values in a byte | 8 / 256 (0–255) |
| One hex digit | one nibble = 4 bits |
The byte 0x4D |
77 decimal · M ASCII |
| Printable ASCII range | 0x20 (space) – 0x7E (~) |
| JPEG signature / footer | FF D8 FF … FF D9 |
| ZIP / Office signature | 50 4B 03 04 (PK..) |
| PDF signature | 25 50 44 46 (%PDF) |
| Windows executable | 4D 5A (MZ) |
| Common sector sizes | 512n · 4Kn · 512e (4 KB physical, 512 logical) |
| Default NTFS cluster | 4,096 bytes (eight 512-byte sectors) |
| LBA 2048 × 512 | 1,048,576 = 1 MiB (why partitions start there) |
Physical persistence: HDD vs. SSD
| HDD (magnetic) | SSD (NAND flash) | |
|---|---|---|
| A bit is… | magnetic orientation of a platter region | charge trapped in a cell |
| Reading disturbs it? | no | no |
| Overwrite in place? | yes (re-magnetize) | no — must erase the whole block first |
| Deleted data after delete | persists until reused (months) | TRIM + GC can physically erase it (seconds–minutes) |
| Recovery instinct "we have time" | usually true | dangerously false |
A byte has no intrinsic meaning — 0x4D is 77, or M, or a flag, depending on interpretation — and multi-byte values carry endianness, which is a property of the format (NTSF/FAT little-endian; JPEG and network order big-endian), not just the machine. Slack space — the unzeroed gap between a file's end and the end of its last cluster — quietly preserves fragments of older, deleted files, making it a productive hunting ground.
You can now…
- ☐ Read a hex dump's three columns and convert any byte among binary, hex, decimal, and ASCII.
- ☐ Recognize a file from its signature regardless of its extension, and account for endianness in multi-byte fields.
- ☐ Explain physically how an HDD and an SSD each store a bit, and why that difference governs recoverability.
- ☐ Compute byte offsets from sectors and back, accounting for 512 vs. 4Kn sizing, and calculate RAM and drive slack.
- ☐ State, defend, and bound "deletion removes the pointer, not the data," including the cases (overwrite, TRIM, encryption, physical destruction) where the data truly is gone.
Looking ahead
Chapter 3 — Storage Technology. You learned how a single bit is stored; next you build out the whole machines — the mechanical anatomy of hard drives and their failure modes, the internal architecture of SSDs and how flash wears out, and the multi-disk worlds of RAID, NAS, and SAN — and what each failure means for getting the data back.
One sentence to carry forward: The format burns the map, not the territory — and on magnetic media, the territory waits.