In a recent note, Gerhard Postpischil said: > Date: Mon, 26 Jun 2006 23:14:51 -0400 > > Paul Gilmartin wrote: > > Ah, but how do you add the flag bit without making the record at > > least one bit longer? > > You don't. But most compression programs already have a few bits to > indicate the type of compression used, so for those there would be no > additional overhead. > Ahem. Those "few bits" serve to increase the size of the file by that amount above the size of the raw file in the case (inevitable according to the Pigeonhole Principle) in which none of the available compression techniques is effective. See Jeffrey D. Smith's recent contribution to this thread for a cautionary tale concerning the unfortunate consequences of assuming otherwise.
That said, it appears that a program object (original topic) at least in UNIX FS appears always to start with the string "IEWPLMH ". So if a compressed file is guaranteed not to start with that string, it can be used as the "not compressed" flag. But that assumption reduces the generality of the technique. -- gil -- StorageTek INFORMATION made POWERFUL ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
