Converting Between ASCII and EBCDIC

You do need to understand a bit about the blocking factors on the foreign tape , () but once you've got that down, you can handle just about anything.

For example, to read an EBCDIC tape on tape device /dev/rmt0 and convert it to ASCII, putting the output in file was_ibm:

% dd if=/dev/rmt0 of=was_ibm ibs=800 cbs=80 conv=ascii

dd reads standard input and writes to standard output, but if you want to specify file or device names, you can use the fairly non-standard if= and of= options to specify the input file and output file, respectively.

If you wanted to convert the other way, you could use this command:

% dd if=was_unix of=/dev/rmt0 obs=800 cbs=80 conv=ebcdic

There's also a conv=ibm option, which uses a different ASCII to EBCDIC conversion table. According to the dd manual page, "The ASCII/EBCDIC conversion tables are taken from the 256 character standard in the CACM Nov, 1968. The ibm conversion, while less blessed as a standard, corresponds better to certain IBM print train conventions. There is no universal solution."

Some gotchas:

• You need to be able to read the raw device () to do the conversion, since the tape probably doesn't use standard UNIX tape block sizes.
• You need to know the blocking factor of the foreign tape, so you can tell dd about it.
• If the foreign tape has multiple files on it, you'll have to use the tape device name that allows "no rewind on close" () to read past the first file.

One last thing to mention about dd: all options that refer to sizes expect counts in bytes, unless otherwise mentioned. However, you can use keyletters to indicate various types of multiplication: k means to multiply by 1024; b to multiply by 512 (a block); and w to multiply by 4 (word). You can also show an arbitrary multiplication by separating two numbers with an x.

- TOR