fixing broken .Z files?

Yes; to the extent that recovery from corrupted compressed data is possible, it means that the compression is less than ideal. However, most compression programs cannot take advantage of some of the redundancy available; as a simple example, if a text about heart disease suddenly starts talking about unusual stock market trading styles, something has gone wrong - but that is unlikely to be redunancy that a compressor can compress out. Indeed, few compressors are capable of even so much as taking advantage of a text being (say) almost entirely in English; corrupting the compressed form usually results not in valid but inappropriate English words but rather in non-words. That is the sort of redundancy that makes recovery possible. And compress (.Z) compression still leaves a lot of redundancy in the compressed data. What I sketched was basically taking advantage of that redundancy to deal with corrupted compressed data. For example, I took the second paragraph above ("Yes; to...non-words.") and compressed it to .Z, then changed a single 1 bit to a 0 (which I think cannot produce an invalid code, though I'd have to think more to be certain) and uncompressed it. Here's a diff between the original and the result: 1,2c1,2 < Yes; to the extent that recovery from corrupted compressed data is < possible, it means that the compression is deficient. However, most --- 7c7 < compress out. Indeed, few compressors are capable of even so much as --- 9c9 < corrupting the compressed form usually results not in valid but --- Obviously, there was redundancy lurking in the compressed form, and that redundancy is what my suggested recovery tactics leverage. /~\ The ASCII der Mouse \ / Ribbon Campaign X Against HTML mouse at rodents.montreal.qc.ca / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B