Early computers and A D Booth

[chris_leyson]

The Booth multiplier is quite well known and still used today although probably with a few improvements, but I didn't know was that he very nearly invented the floppy disk. http://www.bbc.co.uk/news/technology-35924858. Floppy disks weren't stiff enough so he developed drum memory.

Another nice read, https://www.dcs.bbk.ac.uk/site/assets/files/1029/fourth-man.pdf

Lyons LEO computers were also being developed at the same time and I just remembered my mum telling me that she worked at Lyons in Hammersmith as a punchcard operator, must have been early to mid 1950's, she told me that every now and then an engineer would appear and check the neon lights on a panel, damn, I wished I had taken more interest.

[Gyro]

If only he'd tried a stiffer nickel coated disc before switching to the drum! I wonder what else intervened to persuade him that the drum was a 'better' direction. It can't have been ease of machining or construction. Maybe he was worried about 'bit density' changing with radial head position and opted for constant speed. 

Edit: He'd have been at multiple platters in no time flat!

Edit, Edit: Come to think, I've a feeling that the early drums were 1 head per track. Maybe he didn't envisage head movement at that stage.

[Brumby]

Edit, Edit: Come to think, I've a feeling that the early drums were 1 head per track.

As I recall, that's correct.

[coppice]

Its usually Kathleen Booth who is credited with the idea of the floppy disk, although both Kathleen and Andrew Booth did a lot of early work on magnetic storage.

[SimonR]

The Booth Multiplier was first implemented in the HEC1 computer by Dr Bird. He went to Andrews booths workshop to ask for permission to use it.
Dr Bird is a remarkable man and bright as a button, I was lucky enough to meet him a couple of weeks ago.

[Cerebus]

If only he'd tried a stiffer nickel coated disc before switching to the drum! I wonder what else intervened to persuade him that the drum was a 'better' direction. It can't have been ease of machining or construction. Maybe he was worried about 'bit density' changing with radial head position and opted for constant speed. 

Edit: He'd have been at multiple platters in no time flat!

Edit, Edit: Come to think, I've a feeling that the early drums were 1 head per track. Maybe he didn't envisage head movement at that stage.

I don't think I've ever encountered a drum store that wasn't head-per-track. I know some movable head drums existed,  but they were comparatively rare in practice.

When I first worked on computers drum stores were something you would actually encounter. Drum stores usually had bandwidth and latency advantages over disk stores. This advantage eroded over time and drum stores disappeared in the late 70's to early 80's. They were often used as the primary swap memory on virtual memory machines. Drum access times were measured in milliseconds, disk access times in tens of milliseconds.

The drums could be run physically faster than any disk. This had bandwidth advantages as magnetic storage was constrained by the minimum physical size of magnetic domain that could be stored on the medium. Some were so big and rotated so fast that they were regarded as a real hazard to property, life and limb should any mechanical failure take place. I've stood next to a drum cabinet that was over 5 feet tall and 3 feet square - from memory the drum was 30 inches in diameter but that memory is rather fuzzy.

As disk technology improved drums gave way to fixed head disks, then to fixed (moving head) disks, then to removable disk packs (typically 14 inch diameter aluminium disks).