38

A close friend of my mother worked at DEC in the 1970s and 1980s. She recently passed away, and in sorting through her estate, my mother discovered some reel-to-reel magnetic tape. We are curious about what might be on it. I haven't yet seen a picture of it, but Wikipedia tells me this is most likely DECtape.

Is there any chance the data on it is still good? It was not preserved with great care, but as far as we know it has also never been particularly abused. Just left in a box and moved a few times.

If the data is still valid, do we need to dig up a PDP or VAX or read it, or is there a more modern option?

6
  • 9
    Oh, maybe the friend was a corporate spy, funneling high tech secrets to the USSR, would make a good book. ;->
    – Moab
    Mar 2, 2011 at 23:38
  • 1
    I'm curious @JoeWreschnig... Were you able to recover the data? Mar 9, 2011 at 15:38
  • 4
    Guess there's no app for that...
    – Ivo Flipse
    Mar 9, 2011 at 16:09
  • We've not tried yet. I was silently hoping someone would have been crazy enough to build a DECtape -> USB reader or something similarly straightforward, but now this has entered the realm of "a project" rather than "a curiosity", and like most geeks my list of projects is already a mile long.
    – user35138
    Mar 9, 2011 at 16:25
  • We still haven't read it, but we found some paperwork that strongly suggests the contents is a Pascal compiler, probably along the lines of Sam's answer.
    – user35138
    Oct 10, 2011 at 18:34

9 Answers 9

19

If the tape is DECtape, you will definitely need to find a TU-56 drive to read it. TU-56 units are going to be pretty hard to come by, and would likely require some repair (replacement of old capacitors, decaying wiring, burned out indicators). As far as hosts go, you are way more likely to find a working vax than a working pdp-8, but you will need the right interface board (likely Qbus for the smaller vaxes). The bit density is fairly low at ~350 bpi and the signal is manchester encoded, so if you can get the data some other way you could certainly put it into an emulated pdp or vax (see simh-vax, works very well: simh.trailing-edge.com/vax.html). If you are very hard core you can build the interface card yourself: http://so-much-stuff.com/pdp8/cad/projects/boards.html . For more info on the TU-56 drive: http://www.pdp8.net/tu56/tu56.shtml . All of this is a lot of effort for ~184 kwords of data. You may want to try comp.sys.dec to see if anyone has any tu-56s laying around, and I certainly second the recommendation of contacting a computer history museum.

1
  • 12
    I don't know that I have ever seen so many archaic terms in one place. You should change your name to Time Machine! Mar 7, 2011 at 3:25
17

Dectapes are easily differentiable from standard open-reel tapes as they were 1-inch wide and the reels were maybe 4-5 inches in diameter, and the flanges were maybe a half-inch deep.

Dectape was one of the most robust media of its day, if not THE most robust. It had redundant timing and data tracks and could be read or written in either direction. The salesmen used to wow potential customers by punching the tape —multiple times!— with a hand held paper punch and showing that the tape was still readable. Told to me by one of the salesmen: After the paper-punch demo, a customer asked if it was water-proof, too. The salesman told us that he had had no idea but for the heck of it, he had dropped one into a bucket of water, and took the customer to lunch. When they got back, he said, he'd hung it on the drive and read it —successfully!— while water streamed off of it in all directions.

It didn't need any sticky markers —the timing tracks had all the position information it needed— but you will need a Dectape drive, and one or another DEC machine to run it (unless you're up for the challenge of building an interface and writing the low-level tape drivers for it).

1
  • "Gee I wonder if a Hard Drive is robust enough to drop into a bucket of water and then run right after" Mar 3, 2011 at 17:49
11

Unfortunately the tapes have probably degraded past the point of recovery. The upper bound on magnetic media lifetime is 30 years:

http://www.clir.org/pubs/reports/pub54/4life_expectancy.html

9

A computer history museum may be able to help.

1
  • The museum of computing at Bletchley Park might have the equipment needed. Also might be worth contacting the science museum.
    – user69937
    Mar 3, 2011 at 0:56
7

I work for a company that specialise in legacy systems support, and we recently had some involvement in a project to migrate data from several hundred 1/2" tapes onto modern media. The date the media was originally written to, ranged from mid 1980s to just a few years ago. We had about a 75% success rate, but I don't have any stats to show the success rate in relation to the age of the tapes or data.

You will still find many companies still using equipment like this, some financial institutions still have equipment like this, as do several companies that use industrial control systems, but the chances of finding one that is willing to help is pretty slim, as this type of equipment is very expensive to keep running due to their age.

I'd suggest Chris' answer is probably your most realistic option, but to be honest I doubt a computer history museum will have a fully working system, as equipment of this age really does require lots of TLC to maintain.

5

It should be possible to read these old tapes by using very fine iron filings and then photographing the bit patterns that become visible. Obviously this is not a speedy way to get the data off, but it is doable on old tapes because they had much lower density than today's magnetic media.

2
  • I did this fairly recently on a 1/2" tape, not to attempt to recover data, but to demonstrate that it was possible. I forget the name of the product, but essentially, it was a liquid full of manganese particles. Obviously using something like this will make it impossible to read the tape should you later find a tape device.
    – Bryan
    Mar 3, 2011 at 9:40
  • Also you need to take into account the encoding method used on the tape, as it will likely be phase encoded (en.wikipedia.org/wiki/Phase_encoding) or NRZI (en.wikipedia.org/wiki/NRZI) - in other words, don't expect to see a series of 1s and 0s that will translate into ASCII
    – Bryan
    Mar 3, 2011 at 9:41
4

Something you should consider before taking on this task is, is the data WORTH reading? Theres been several people that gave you good data on the possibility of being ABLE to read them.

But most likely its just some backup tapes of an old database or something equally mundane, and even if you can read the data, its going to be indecipherable to you.

Its not like your going to read this tape and its going to be full of crazy corporate secrets or the answer to life, the universe and everything.

1
  • I'm definitely sure it's gonna contain more than '42'!
    – fretje
    Mar 3, 2011 at 14:40
3

You'll need the hardware for the tapes. Also, the begin/end markers were adhesive stickers placed on the backside of the tape and used for optical alignment (called "end of tape"). If those stickers came off, the tape could not be read again. I don't remember if DecTapes had them, but the other (7 and 9 track tapes) had them. I had boxes of each type of tape, and threw all mine out in the mid 90s.

2
  • The markers in my experience are only used for BOT (beginning of tape), end of tape markers (AKA file marks) are usually written to the media as a specific data pattern. BOT markers are just reflective labels, as an optical sensor located by the read/write head detects a reflection from a light source. The chances of these labels dropping off is pretty slim, at least I've never known this happen.
    – Bryan
    Mar 2, 2011 at 22:23
  • No, there's an EOT reflective marker as well. It was normally placed about 20 feet before the physical EOT. During writes the drive will report "success, but we passed EOT" when it sees this go by. The writing software is then supposed to write a tape mark, an end-of-volume label (no end-of-file label unless that was coincidentally the last block of the file), and two more tape marks iirc. Reading software detects the EOV label and knows to not try to read further, instead rewinding the tape and asking for the next volume. (I may have missed a couple of details here, but the gist is correct.) Aug 11, 2015 at 4:28
2

Seems people have mentioned the age of tapes, and what is on them as factors, I'll add one more ... the size of the data chunks saved is like to be only in the kilobyte range.

In the end, if you can get to the byte data, it's probably fixed length records. Look for some common patterns to identify the headers. If it's variable length, there will be some type of common header. Beyond that, each bit will be a flag, this is back in the days when data storage - including programs - was at a super premium.

Sadly, the stacks of punch cards (sorry to clober tapes in with punch cards) I know of people keeping were usually of the assembler program itself. This was the nuclear winter plan - if all else failed - they could get the assembler back on the box and start typing away to get things going.

I hate to say it's so unlikely to contain anything valuable now - mainly because of the size. You can probably type one of those tapes full using notepad/ascii in less then 10 or 20 minutes for a real typer.

You must log in to answer this question.