Novell systems are pretty cool. They compress data in the background, and when you run a backup on a Novell server, it just sends the data off the disk to the backup software. This means that you don't need to compress the data – it is already compressed.
The downside is when you try to restore the information. You will get a lot of files that do not contain the information you thought they contained. The description of the problem is better explained in my solved problems section.
I wrote a small tool to grab the uncompressed data from a SDIF-encoded stream. It is not verbose, does not double-check data, nor is it really immune to errors in the stream. It is a quick and dirty solution to pull out whatever information you can without actually going to a Novell server.
The tool runs currently on Linux, but should be almost no problem whatsoever to port to another OS. Andrea Manzini tested it (and fixed a couple warnings) on Win32.
If the file was compressed by Novell, you will be out of luck. The compression method they use is not publically documented and it looks like they are quite against the idea of having public documentation of the decompression algorithm. So, if the file was able to be compressed and was actually compressed by Novell, this tool will not be able to decompress it and you won't be able to restore the actual contents of the file.
If the stream appears to be corrupt, the file will not be written.
Good files have the full file path in the header and will be written to the appropriate directory structure. For instance, if the file says it was backed up from SYS:PATH/TO/FILE.BIN it will be saved as SYS/PATH/TO/FILE.BIN. Case is preserved, and colons (":") are translated into slashes ("/").
Download the C source here: extract_sdif_data.c
To run it, just specify the SDIF file to pull the data from: