Origin of "partition" in storage devices
jim stephens
jwsmail at jwsss.com
Wed Feb 2 03:15:17 CST 2022
On 1/31/2022 11:01 AM, Tom Gardner via cctalk wrote:
> <snip>
>
> It seems clear it was used in memory well before HDDs but when it got
> started there is unclear.
> * IBM PC DOS v2 was an early user in 1983 with FDISK and its first PC
> support of HDDs
> * UNIX, Apple OS's and IBM mainframe all seem to come later.
>
We had a local grown OS which was really a way to file paper tapes on a
disk with file pointers. We had an allocation method, all the file
system things, a VTOC (copied a lot of ibm terms). Eventually added
half inch tape.
This was on a Microdata 1621 minicomputer, and was named MPS, or
Microdata Programming System. We could take paper tape distributed
software from Microdata, read them in and patch the terminal read /
write and in the case of compilers, the paper tape read / write
instructions and point them easily at our disk.
Anyway our files had a name and disk pointer which essentially was the
cylinder and sector of the file. All was on a 2.5 mb platter using both
surfaces of our the removable pack on the 5mb Western Dynex drive.
I added the capability to select surfaces which occurred when you booted
the system and it would remember a surface system variable used by the
disk routines that the programs didn't have to be concerned with. We
could run from either platter this way with a copy of the system.
Used the other platter for storing an image and doing a backup. Before
that time, we had not written any way to back up, so I built that up.
I later implemented a system to save out to tape, and rebuild and
install to the disk.
Also could later when I worked at Microdata use multiple drive and do
the copying as well as double sized disks.
There was a lot of odd data stored outside the disk space (allocated and
free).
I implemented a small rotating log of when the disks were booted, a name
for the disk system running you could query, and other information.
I don't know how this might fit into partitioning disks, but it did
divide the two platter system into two separate file systems. I never
moved it to any hardware requiring more complication, but would have
just added more indexing as to where the disk was mounted and what the
volume was.
You'll note that our 2.5mb file system isn't much larger than two 1.4mb
3 1/2" floppies, but we ran all the system software that we needed,
system generation, sources and user files and easily had 20 to 25% of
the space left over. Much smaller systems back then.
Single user similar to a large high wattage PC I suppose.
The partitioning was implemented in 1972 (or dual platter backup).
BTW we had an entertaining file recovery and sort that ran on either
platter. We'd erase the "other" platter to no files, and then copy all
the files in on the other volume to the alternate. When you finished
all of your files were sequential and not wasted space.
Our file system used linked sectors, and the free space was allocated by
a linked list of sectors, not a bitmap. The beginning of files were
luckily odd enough that they could be recognized and another utility
could recover all the files and rebuild the entire VTOC with nothing but
the files present. IMplemented over a couple of horrible days when the
VTOC got erased.
thanks
Jim
More information about the cctalk
mailing list