20 Jul
2014
20 Jul
'14
4:19 p.m.
> PC-DOS/MS-DOS 3.10 (three point ten) had a limit
of 32M for a regular hard
> disk.
On 07/18/2014 10:51 PM, Fred Cisin wrote:
I think the disk could be larger, could it not (although practical maximum
for the consumer drive technology of the era was probably around 120MB)? It
was the initial FAT16 implementation that topped out at 32MB, so a drive
would have to be partitioned (and I don't recall what the limit there would
be - although I think the CHS addressing scheme topped out at 8GB,
presumably limits on the number of partitions possible would come into play
first).
Good catch; thank you.
Yes, the limit for a logical DOS Drive in 3.10 was 32M, but the physical
disk drive could be substantially larger. It was not at all uncommon to
use a 40M drive, split into 2 (or more) DOS drives.
I don't think there was anything in 3.10 which
stopped it from addressing a
partition beyond the 32M boundary, but maybe I'm wrong. I'd assume that the
drive test program would work at the drive level, not at the partition level.
You are absolutely right.
FAT16b came in with DOS 3.31, and I think its
individual partition size was
then capped at 2GB (at least assuming 512-byte sector sizes).
and almost all of the 2G limits should have been 4G, since they were a
limit of 32 bits. But, they ended up at 2G instead of 4G due to
programmers who never expected to come near such enormous limits and used
a signed, instead of an unsigned number. Because of that, most of DOS
was willing to accept the existence of a NEGATIVE 2G file or drive!
(Note: copying a negative sized file to an almost full drive fails to
increase the free space.)