A complicating factor in transferring files from one computer to another is that you must decide whether the files you want to transfer should be treated as “text” of as “binary” . Files that contain simple text, including program source code, should be transferred in “text” mode. Compiled programs, compressed files (e.g., *.zip or *.Z files), and word processor files with embedded formatting codes are generally transferred in “binary” mode.

The reason for this binary/text confusion is that Unix, MSDOS, IBM, and other systems disagree on how to represent basic text. For example, the end of a line in a Unix text file is represented by a single character (the ^J or “line-feed” character) while MSDOS uses a pair of characters at the end of each line (a^M or “return” character followed by a ^J). Other operating systems have their own peculiarities. Most file transfer programs will, when transferring text, try to convert the transferred file into the appropriate format for the destination machine. These conversions may involve changing, adding, or deleting characters. Of course, if the file being transferred were not text but a compiled program, any such changes to individual bytes would be disastrous. Consequently, you need to be aware at all times whether the files you are working with are text or binary.

The easiest way to tell (though not foolproof) is to try listing the file on your screen using the Unix “cat” or “more” commands or the MSDOS type command. If it looks OK, its probably text. If not, or if in doubt, transfer it in binary mode.

If you do a transfer in binary mode and then discover that you have a text file with the wrong line terminators, you can correct that on the Unix side, as described in Section 6.4: Problems and Inconsistencies.