3.5 inch floppy disk
This format was designed to be smaller than the 5.25" floppy drive and its media and fixed some of the 5.25" design problems. The disks are housed in a stiff casing, opposed to the "floppy" plastic the 5.25" disk was housed in. Despite the more rugged design, 3.5 inch floppy drives tend to to be less reliable due to their compact nature.
3.5" floppies came in three different densities (easily distinguished by a second hole on the side) and while only the earliest disks were single sided, almost all 3.5" floppies can be written to on both sides.
The 3 densities are:
- Double density (DD, 720 kB)
- High density (HD, 1.44 MB)
- Extended density (ED, 2.88 MB)
Like earlier floppies, there were many different ways to format a disk, often depending on what type of computer the disk was formatted with. On PC's, DD disks were usually (DOS-)formatted to 720 KB (though in practice it had about 714 KB of usable space), HD disks formatted to 1.44 MB (in practice it was around 1.38 MB) and the uncommon 2.88 MB disks (primarily used in IBM PS/2s) formatted to around 2.76MB. On older Macs the DD disks formatted to around 820 KB, and on Amiga 880 KB.
In the 1990s the demand for higher capacities increased. Software utilities appeared that could format disks to higher capacities like 1.72 MB. By loading a small TSR-driver such formatted disks could be used like the normal 1.44 MB formatted disks in DOS. Well known tools are fdformat (fdread), vgacopy (vgaread). Even Microsoft made use of this by introducing the DMF format (1.68 MB) for distributing software on disks. Current archiving tools like Winimage also support disks with increased capacity.
Most over-formatting utilities don't handle ED disks, as the format never gained much popularity.
The overall transfer speed when reading from a 1.44 MB disk is about 30 kB/s. This could be tuned by formatting disks with an optimal track interleave that reduces access times after a track change on sequential reads significantly. Tools like vgacopy offered this feature.