> From: Carl Fink [mailto:[EMAIL PROTECTED] > Sent: Wednesday, December 14, 2005 7:21 PM > > > On Wed, Dec 14, 2005 at 02:34:31PM -0800, Steve Lamb wrote: > > > IIRC one of the algorithms that can be used in zip is > > lzw which is (or was) patented. > > The patent expired in 2003. > > http://www.sslug.dk/patent/lzwunisys.html
In that case, it appears that using 7-zip to produce fixed-size, split zip files followed by generating a set of PAR files would accomplish the same thing as using RAR. That seems to answer Gabriel's question that started the thread, yes? The zip files each contain CRC's, so you can tell when one is bad and the PAR files allow you to repair defects. The two approaches are both the same: a message authentication code (hash, CRC or other MAC) to detect errors, followed by forward error correction to repair detected errors. You decide how much FEC you get by how many PAR files you generate. Same old - same old, except the compression and FEC are better than older methods. I also don't know if any other program besides 7-zip can put together the split archives it produces, since 7-zip can't put together split archives from WinZip. That's not much of an objection, since the program is free and the code is open for others to build implementations around. I haven't bothered doing split 7z or tar archives, yet, but the 7z archives are similar in size to RAR and have better MAC's. That makes the main advantage of RAR only that the PAR file generation is conveniently built-in. If a customer asks for RAR, I don't argue, so the fact that some people like it is also an advantage. -- Seth Goodman -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
