walton@tybalt.caltech.edu.UUCP (04/12/87)
In article <3124@jade.BERKELEY.EDU> spencer@eris.BERKELEY.EDU (Randy Spencer) writes: >I haven't read the Ckermit manuals for Unix and the Amiga in a while, but >do they include compression, and is compressing the file before transmitting >sort of redundant and repeatitive? > The compression code in C Kermit is in the machine independent code shared by the Amiga and UNIX versions, so the answer to your first question is yes. Kermit's compression is a simple replacement of repeated runs of a character with a three-character sequence meaning "start of repetition"-"repeat this many times"-"this char". The repeat count is limited to a maximum of 94 since Kermit maps a subset of the printable ASCII characters into integers. This is much simpler than what compress does, so kermit'ing compress'd files OVER AN EIGHT BIT LINE is still a big win. Over a seven bit line, you'd best send the clear text, since in that case Kermit is forced to "escape" the eighth bit by turning each character with the high bit on into a special escape character followed by the desired character with the eighth bit off. This won't hurt a text file, but will about double the communications of a binary file like the output of compress. Steve Walton guest on:walton@tybalt.caltech.edu Hypercube Programmer