dalegass@dalcsug.UUCP (09/15/87)
I have recently been informed that posting ARChived binaries and sources actually takes more transmission time in the net than if the un-compressed files were posted. This is supposedly due to the fact that unix 'compress'es these postings before they go out, and an .ARC file actually *grows* when compressed. I find it hard to believe that the unix file transfers don't check for the 'compress' error code 2 (resulting file is larger than original). Could anybody clarify this issue? I'm not interested in statistics of compress and ARC, because I *do* realize that compressing an ARC file will make it grow--I'm not disputing that... If this turns out to be true, it certainly would make more sense for everyone to post 'shar'ed archives, which are not compressed at all. Hopefully, unix is smart enough not to use a enlarged-compression when sending, since the .ARC files get a better overall compression than 'compress'ed files... -dalegass@dalcsug.uucp
dave@sdeggo.UUCP (David L. Smith) (09/15/87)
In article <138@dalcsug.UUCP>, dalegass@dalcsug.UUCP (Dale Gass) writes: > I have recently been informed that posting ARChived binaries and sources > actually takes more transmission time in the net than if the un-compressed > files were posted. This is supposedly due to the fact that unix 'compress'es > these postings before they go out, and an .ARC file actually *grows* when > compressed. Nope. Compress will expand (marginally) an ARC'd archive, since you can only compress something so much, but once it is uuencoded, it can be compressed again, since uuencoding introduces some redundancy. Furthermore, not all sites run compressed news and for some sites (like mine) where disk space is at a premium having four or five 100K uuencoded binary files would make life quite miserable. > If this turns out to be true, it certainly would make more sense for everyone > to post 'shar'ed archives, which are not compressed at all. You can't shar binaries. Just don't work. -- David L. Smith {sdcsvax!sdamos,ihnp4!jack!man, hp-sdd!crash}!sdeggo!dave sdeggo!dave@sdamos.ucsd.edu "How can you tell when our network president is lying? His lips move."
johnk@auscso.UUCP (John Knutson) (09/16/87)
In article <138@dalcsug.UUCP> dalegass@dalcsug.UUCP (Dale Gass) writes: |I find it hard to believe that the unix file transfers don't check for the |'compress' error code 2 (resulting file is larger than original). I don't. That would mean that anything that had that error could never be sent (arc or no arc). Sounds okay to me.. (but i don't pay ld bills either - not yet at least) -- John Knutson {ihnp4,allegra,ut-sally}!ut-ngp!auscso!johnk (please do NOT mail to knutson6 anymore) Communicating, like in the good ol' days.
davidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr) (09/16/87)
A few words about the use of ARC, compress, and zoo. ARC and zoo run in
both UNIX and DOS, compress seems to run 12bit only in DOS (it could
barely run 16 bit).
Method ARC zoo comp12 comp16
real 100.16 52.72 29.98 28.14
user 92.16 49.62 24.36 24.66
output bytes 87767 82913 86122 71343
Since this is intended for information only, I draw no conclusions.
--
bill davidsen (wedu@ge-crd.arpa)
{uunet | philabs | seismo}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me