Random Access Tape?

Nico de Jong wrote:

Truely random is not possible on a tape, as you (as you said) still have
to
read the blocks inbetween.

Or not. 3480's have a locate funstion in the control unit that finds a
block by block ID. The CPU doesn't read the data, The CU probably does.

That you dont _use_ the data for anything, is
beside the point.

Some kind of semi-random is possible, but it is a stretch of the
imagination.

I have come across a program running on a (IIRC) Tandberg streamerdrive.
This drive makes multiple passes when writing the data. When the tape
changes direction, the program (I believe it was a backup program)
remembers
the blocknumber of the first block on the track, together with the name of
the file being written. When the program finishes, all this data is saved
somewhere.
When you then want to restore a specific file, the program checks this
table, and sees on which track it is located. It will then go directly to
the correct track and find the program. However, you can only "jump" whole
tracks. You still have to read past all irrelevant blocks.

I believe the Multics backup worked this way. Multics tapes had
software-assigned block numbers, and the backup recorded the starting
block number for each file.

Joe Pfeiffer wrote:

writes:

[snip...] [snip...] [snip...]

DECtapes had random access.

Not as the term is normally used. For that matter, disks don't really
have random access (which would be uniform access time to the words of
data regardless of the current location of the head; seek time and
latency keep the disk from being truly random access).

While you could request any block on a DECtape, the time to get that
block was determined by how far away it was from where you currently
were on the tape.


Yes, and dynamic RAM memory units are *not* random acess by this criteron.
While the contents of any memory location can be delivered within a specified
maximum time (memory cycle time), the memory units farther from the output
actually take longer to acess, because even electrons traveling at the speed
of
light *do* take time based on the distance traveled.

For me, if I can access a "black box" device to acquire data by giving a file
name,
and I do *not* have to care how the device retrieves that file...then it's
random
access. DECtapes and Stringy Floppies will automatically position the heads
relative
to the tape, and read the requested data. A hard disk or floppy disk will
position the
heads and read the requested data. I don't have to care. I issue the
command...I get
my data.

In article ,
Carl Pearson wrote:

Please don't toppost. I've fixed it for you...once.

wrote:
In article ,
Carl Pearson wrote:
Howdy, Group,

Been having a conversation with this guy regarding tape vs disc.

He asked if a hard or floppy disk was more like a tape recorder, or a
record player.

Neither. A tape recorder and a record player have the same
geometry; Both were one long string of info packets
packaged in a spiral.

I'm siding with record player, due to tape's inability to have random
access.

I know, record players are WORM drives, and they don't record with metal
oxide, but the random access feature seems so important that it
outweighs tape's next-bit-in-line way of reading data.

(Or, it's at least *as* important. But in my view, if tape could really
have random access, we wouldn't be using the current much more expensive
disc drive technology. On another front, imagine the wait time involved
in getting to the last sector of a 50 terrabyte file with tape!)

DECtapes had random access.

Good point about the similarity in geometry, I had mentioned that in the
original post.

Not quite. There is the physical geometry and then there is the
paths the heads have to travel before getting to the location
where the data is stored. I think you also are confusing the
two "geometries". I can't recall the term used for the actions
needed to get at a location.

I think the old DEC drives are what's hanging this guy up, though given
that they held less than a meg of data, I really don't see how they in
any way correspond to a modern disc drive.

It's not clear to me what your friend is trying to teach himself.
I'll hazard a guess that he's trying to distinguish a "one-way"
device with an "any-way" device. With most magtapes to get at
any set of bits, you started at the beginning of the tape and
kept reading until you hit what you wanted. If the next
set of bits were located "earlier" on the magtape, one had
to rewind the tape to beginning of the tape and start all over
even if the bits you wanted were at the end of the tape.

With a disk or floppy, there are round circle of "tape"; each
circle has a diminishing diameter. So getting at a location
is "faster" because the disks can start at the point where
the last access was done. There are two "ways" these devices
can physically travel to get from here to there; it's rather
like walking around a city on an island whose roads intersect.

I can't tell you a thing about how CDs work; those critters
are magic spinning saucers to me. :-)

Note to others: something just exploded outside; sounded about
a block away and the lights flicked. I may not be online
later.

/BAH

In article ,
Andrew Swallow wrote:
wrote:

In article ,
Carl Pearson wrote:

Howdy, Group,

Been having a conversation with this guy regarding tape vs disc.

He asked if a hard or floppy disk was more like a tape recorder, or a
record player.


Neither. A tape recorder and a record player have the same
geometry; Both were one long string of info packets
packaged in a spiral.

A better analogy is a blackboard. You can write anywhere
you want in any order that you want.

Yours is better but I think the one I just posted is better
than yours. It takes lots of tries to get a good
description so that our unwashed ;-) can begin to understand.

/BAH

In article ,
Joe Pfeiffer wrote:
writes:

In article ,
Carl Pearson wrote:
Howdy, Group,

Been having a conversation with this guy regarding tape vs disc.

He asked if a hard or floppy disk was more like a tape recorder, or a
record player.

Neither. A tape recorder and a record player have the same
geometry; Both were one long string of info packets
packaged in a spiral.

But a hard or floppy disk is "more like" a record player, since you
can pick up the tone arm and drop it someplace else on the disk
relatively quickly.

For reads, sure. But I didn't think a floppy was a single stream
of bits. A record player position (if it's not started at
the beginning) is a hit and miss type of seek. You can reduce
the time of the seek by starting in the middle but you still have
read everything after that to get the data you want.

..With a tape, you have to fast-forward or rewind
past all the data you're not interested in.

Those were added "features" to smooth operator feathers :-).
But you still had to start somewhere before the spot you
wanted and then sequentially walk down the track.


I'm siding with record player, due to tape's inability to have random
access.

I know, record players are WORM drives, and they don't record with metal
oxide, but the random access feature seems so important that it
outweighs tape's next-bit-in-line way of reading data.

(Or, it's at least *as* important. But in my view, if tape could really
have random access, we wouldn't be using the current much more expensive
disc drive technology. On another front, imagine the wait time involved
in getting to the last sector of a 50 terrabyte file with tape!)

DECtapes had random access.

Not as the term is normally used. For that matter, disks don't really
have random access (which would be uniform access time to the words of
data regardless of the current location of the head; seek time and
latency keep the disk from being truly random access).

Yea, I thought about that term when I typed it but left it in.
I've probably drifted the thread by being this careless.

While you could request any block on a DECtape, the time to get that
block was determined by how far away it was from where you currently
were on the tape.

But you could backup and stop and then rocker arm the tape
until you find the location. Magtape drives didn't seem
to have this ability. I never understood why that function
didn't get into them.

/BAH

In article ,
Steve O'Hara-Smith wrote:
On Fri, 21 Oct 2005 12:03:09 GMT
Carl Pearson wrote:

He keeps claiming that tape *can* have random access. Were this the
case, of course discs would be more like a tape recorder, as the random
vs sequential access argument would be thrown out of the discussion.

Well I have heard of a UNIX file system being mounted directly
from a DEC tape drive - apparently the superblock accesses became very
visible

I recall a case where a -10 was run from a DECtape to make room
for a program and data that couldn't fit. It took all night
and stand alone to do the computing job.

Strictly speaking random access should imply that the time taken
to access any data item is independent of the location of that data item.
A tape is clearly not random access nor is a disc, but a disc is a closer
approximation to random access than a tape.

We're not talking about any old tape, but a DECtape.

/BAH

On Sat, 22 Oct 05 10:18:56 GMT
wrote:

In article ,
Steve O'Hara-Smith wrote:

Strictly speaking random access should imply that the time taken
to access any data item is independent of the location of that data item.
A tape is clearly not random access nor is a disc, but a disc is a closer
approximation to random access than a tape.

We're not talking about any old tape, but a DECtape.

Not even a DECtape achieved anything like constant time to find
a word.

--
C:WIN | Directable Mirror Arrays
The computer obeys and wins. | A better way to focus the sun
You lose and Bill collects. | licences available see
| http://www.sohara.org/

Peter Flass wrote:
I believe the Multics backup worked this way. Multics tapes had
software-assigned block numbers, and the backup recorded
the starting block number for each file.

Not to my knowledge. Yes, Multics standard tape format had
a block number in each block, and the tape DIM noticed an error if
the number was not monotone increasing. It did not provide
this block number to its caller, and the hierarchy backup
software did not record the block number in the dump map,
and the retriever/reloader software did not try fancy positioning
tricks to find a particular file.

The tape format described in
http://www.multicians.org/mspm-bb-3-01.html
had some cool attributes though. If an error occurred when writing
a tape block, the tape DIM just wrote the block again, with the
same block number. To read a block, it read until past the desired
block. This meant that while a tape was reading, an operator could
put the drive in standby, dismount the tape, mount it on another
drive, switch unit numbers, and ready the new drive: and the
tape DIM would notice the block number mismatch, resynch, and
continue reading without indicating an error.

Multics tape format was also specified to have a single tape mark
every 128 blocks, so that spacing forward to a given record number
would be fast. We never used this feature and I think after a while
eliminated it: it just made it difficult to copy a Multics tape on a
360.

Carl Pearson writes:
Howdy, Group,

Been having a conversation with this guy regarding tape vs disc.

He asked if a hard or floppy disk was more like a tape recorder, or a
record player.

I'm siding with record player, due to tape's inability to have random
access.

One word DECtape.

scott

Of course the LGP-30 had main memory AND the accumulator on DRUM, and
addresses were NOT sequentially assigned. A good programmer located
his working storage so that it could be most often fetched in the same
revolution as the instruction, then fetch the next instruction wihout
wasting a whole revolution of the drum. There was a chart showing what
memory locations could be accessed from an instruction at any location,
without losing a revolution. It had 4096 words of memory, IIRC.

Don't modern tape drives have multiple tracks, going to the end and
back multiple times, so if you want to get to something 95% through the
complete linear image, you only have to scan down one track part way to
get it, not through 95% of the data?