On Wed, March 14, 2012 00:26, Nataraj wrote:
I think the reality is that nothing lasts forever. Optical media is probably much more likely to survive ICBM's, but then you may not have a drive to read them...
Nataraj
About five years ago I was asked to recover data from a 2400' reel mag-tape of unknown provenance in an unknown format. The possessor was a university professor and the tape dated from the late 1970s. It contained data from social science research projects he had conducted at that time.
I eventually managed to read the tape at 1600 bpi in raw block format and from the headers determined that the encoding was EBCDIC and that the tape had been created on a CDC machine. However, the entire tape after the headers was blank. Not corrupt, not zeroed, just blank. Apparently the operator had mounted, initialized and labelled an evidently new tape but never actually put any data on it.
So the professor had been carefully storing a scratch tape for decades.
Always restore from your backups on a regular basis just to ensure there is something there to restore from.
James B. Byrne wrote:
On Wed, March 14, 2012 00:26, Nataraj wrote:
I think the reality is that nothing lasts forever. Optical media is
probably much more
likely to survive ICBM's, but then you may not have a drive to read
them...
About five years ago I was asked to recover data from a 2400' reel mag-tape of unknown provenance in an unknown format. The possessor was a university professor and the tape dated from the late 1970s. It contained data from social science research projects he had conducted at that time.
*chortle*
a) the data was vastly obsolete b) should have been copied over c) had you asked me, and only telling me the above, I would have told you, with 95% (at least) confidence, that it was EBCDIC. <snip>
a CDC machine. However, the entire tape after the headers was blank. Not corrupt, not zeroed, just blank. Apparently the operator had mounted, initialized and labelled an evidently new tape but never actually put any data on it.
<snip> <chuckle>
I should talk to my friend, who's a history professor up in Canada. At least I think all his stuff is on CDs or DVDs.
mark
On 03/15/12 6:31 AM, James B. Byrne wrote:
I eventually managed to read the tape at 1600 bpi in raw block format and from the headers determined that the encoding was EBCDIC and that the tape had been created on a CDC machine.
if it was from a 70s' vintage CDC system, I'm sort of surprised it wasn't BCD, not EBCDIC... BCD was an earlier 7 bit character code. lucky that was 9-track, the CDC stuff I remember used 7-track tape, even MORE unobtanium.
On 03/15/12 6:59 AM, m.roth@5-cent.us wrote:
c) had you asked me, and only telling me the above, I would have told you, with 95% (at least) confidence, that it was EBCDIC.
actually, I dealt quite a bit with 9-track ASCII tape from that era... DEC, Data General, HP, etc systems were all ASCII based. I'd only expect EBCDIC if it was from an IBM system, or one of the clones like RCA.
fun bit of personal trivia... my wife's first job out of college was at DEC as a new Technical Writer, and her first assignment was to sort out the Mag Tape subsystem for VAX/VMS 1.x and write first the internal design specs, then users documentation. VAX/VMS had excellent mag tape support, just as 9-track mag tape was fading into oblivion.
fun bit of trivia #2: my very good friend is a hobby-machinist, builds scale working live steam trains when he's not doing hardware/software consulting jobs.... he has a vintage bridgeport mill which he grafted a 100VDC servo motor out of a high speed Ampex 9-track take drive as the Bridgeport came with a 3-phase motor and its rather hard to get 3-phase AC in a residential garage workshop :) This DC motor is smooth, powerful and quiet, and lets him run the mill at variable speeds without messing with the V-belts.
On Thu, Mar 15, 2012 at 12:20:15PM -0700, John R Pierce wrote:
On 03/15/12 6:31 AM, James B. Byrne wrote:
I eventually managed to read the tape at 1600 bpi in raw block format and from the headers determined that the encoding was EBCDIC and that the tape had been created on a CDC machine.
if it was from a 70s' vintage CDC system, I'm sort of surprised it wasn't BCD, not EBCDIC... BCD was an earlier 7 bit character code. lucky that was 9-track, the CDC stuff I remember used 7-track tape, even MORE unobtanium.
On 03/15/12 6:59 AM, m.roth@5-cent.us wrote:
c) had you asked me, and only telling me the above, I would have told you, with 95% (at least) confidence, that it was EBCDIC.
actually, I dealt quite a bit with 9-track ASCII tape from that era... DEC, Data General, HP, etc systems were all ASCII based. I'd only expect EBCDIC if it was from an IBM system, or one of the clones like RCA.
On our CDC system back in the previous century, it is true that BCD was common, but EBCDIC could be selected as could ASCII. I think ASCII with parity was default on 9 track.
But, on the 7 track, it used a 6 bit character set that excluded lower case, eg was only upper case alpha, decimal digits and special characters. I don't remember how binary was encoded on them, probably pairs of 6 bit 'bytes' since the PPs were 12 bit 'machine'.
////jerry
fun bit of personal trivia... my wife's first job out of college was at DEC as a new Technical Writer, and her first assignment was to sort out the Mag Tape subsystem for VAX/VMS 1.x and write first the internal design specs, then users documentation. VAX/VMS had excellent mag tape support, just as 9-track mag tape was fading into oblivion.
fun bit of trivia #2: my very good friend is a hobby-machinist, builds scale working live steam trains when he's not doing hardware/software consulting jobs.... he has a vintage bridgeport mill which he grafted a 100VDC servo motor out of a high speed Ampex 9-track take drive as the Bridgeport came with a 3-phase motor and its rather hard to get 3-phase AC in a residential garage workshop :) This DC motor is smooth, powerful and quiet, and lets him run the mill at variable speeds without messing with the V-belts.
-- john r pierce N 37, W 122 santa cruz ca mid-left coast
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On 03/15/12 1:20 PM, Jerry McAllister wrote:
But, on the 7 track, it used a 6 bit character set that excluded lower case, eg was only upper case alpha, decimal digits and special characters. I don't remember how binary was encoded on them, probably pairs of 6 bit 'bytes' since the PPs were 12 bit 'machine'.
now that I think about it, the BCD I remember *was* a 6 bit code, with just the core 64 characters. and the CDC machines I remember were 30/60 bit word sizes and the PPs were 12 bit... machines like the CDC 6600. so the 6 bit character code fed nicely into 7 track tape (6 bits plus parity). yes, this was 1960s hardware, still in use in the early/mid 70s, IIRC, that 6600 was largely used to support a 7600.
anyways. ancient history. i'd have said 'been there, done that, got the t-shirt', except I wore those t-shirts out 30 years ago.
John R Pierce wrote:
On 03/15/12 1:20 PM, Jerry McAllister wrote:
But, on the 7 track, it used a 6 bit character set that excluded lower case, eg was only upper case alpha, decimal digits and special characters. I don't remember how binary was encoded on them, probably pairs of 6 bit 'bytes' since the PPs were 12 bit 'machine'.
now that I think about it, the BCD I remember *was* a 6 bit code, with just the core 64 characters. and the CDC machines I remember were 30/60 bit word sizes and the PPs were 12 bit... machines like the CDC 6600. so the 6 bit character code fed nicely into 7 track tape (6 bits plus parity). yes, this was 1960s hardware, still in use in the early/mid 70s, IIRC, that 6600 was largely used to support a 7600.
ACK! I was on one of those 6600's in the mid-eighties, having finished my AA and working on the BS (part time, of course), with that weird, um, I think it was 63 chars, and one bit for somethingorother. I fondly (NOT!!!) remember the NOS (which I have referred to ever since as the Noxious Operating System, esp. since the term before I'd been on a PDP-11).
But even when I started programming in '80, BCD was considered very old, and not used. Of course, I was all IBM m/f.
anyways. ancient history. i'd have said 'been there, done that, got the t-shirt', except I wore those t-shirts out 30 years ago.
Wonder if I still have a few punch cards laying around....
mark
On 3/15/2012 3:20 PM, John R Pierce wrote:
On 03/15/12 6:31 AM, James B. Byrne wrote:
I eventually managed to read the tape at 1600 bpi in raw block format and from the headers determined that the encoding was EBCDIC and that the tape had been created on a CDC machine.
if it was from a 70s' vintage CDC system, I'm sort of surprised it wasn't BCD, not EBCDIC... BCD was an earlier 7 bit character code. lucky that was 9-track, the CDC stuff I remember used 7-track tape, even MORE unobtanium.
Sounds like it might have been a system 360 plug compatable machine manufactured by CDC. I worked on several of those back in the 70's. These machines were the "omega" series and ran MVS or VM. We had both an OMEGA II and an OMEGA III at good ole UGA.
Harold
-
Harold Pritchett -
James B. Byrne -
Jerry McAllister -
John R Pierce -
m.roth＠5-cent.us