rs@mirror.TMC.COM (Rich Salz) (04/30/87)
Last week I posted an article in comp.sys.amiga saying that posting
ARC'd files increases the transmission costs to most Usenet sites. In
subsequent e-mail exchanges I have been called a fool, a jerk, and a
liar. Needless to say, this bothers me a bit.
So, can anyone explain to me the flaw in the following scenarios?
Send follow-ups to news.misc, or to me for summary.
METHOD ONE:
Create an archive:
arc a ARCFILE *.[cho]
-rw-rw-rw- 1 rs 113148 Apr 28 10:36 ARCFILE.arc
Run it through uuencode, so I can post it as a Usenet article:
uuencode ARCFILE.arc <ARCFILE.arc >UU-ARCFILE.arc-UU
-rw-rw-rw- 1 rs 155922 Apr 28 10:39 UU-ARCFILE.arc-UU
I will now be posting a 156K file to Usenet. Anyone who is curious about
this program must run ARC before they can do anything with material
herein. When my site sends this out to other sites, it will run compress
on it. This the resulting file sent via UUCP will be this size:
-rw-rw-rw- 1 rs 142769 Apr 28 10:43 ARCFILE.arc.Z
METHOD TWO:
Convert all object files to ASCII and make a shell archive out of them:
for I in *.o; do uuencode $I <$I >UU-$I-UU; done
shar *.[ch] UU-*.o-UU >SHARFILE
-rw-rw-rw- 1 rs 296960 Apr 28 10:38 SHARFILE
I will now be posting a 297K file to Usenet. Most of the files in this
article will be sources that can be directly read by anyone who receives
the article. When I send this out, my news transmission will compress the
file, resulting in something this size:
-rw-rw-rw- 1 rs 115665 Apr 28 10:46 SHARFILE.Z
This is a net savings of almost 20% that will be realized by the vast
majority of sites on the Usenet -- those that compress their articles
before sending them.
--
--
Rich $alz "Drug tests p**s me off"
Mirror Systems, Cambridge Massachusetts rs@mirror.TMC.COM
{cbosgd, cca.cca.com, harvard!wjh12, ihnp4, mit-eddie, seismo}!mirror!rscsg@pyramid.UUCP (Carl S. Gutekunst) (05/04/87)
In article <3540@mirror.TMC.COM> rs@mirror.TMC.COM (Rich Salz) writes: >Last week I posted an article in comp.sys.amiga saying that posting >ARC'd files increases the transmission costs to most Usenet sites. In >subsequent e-mail exchanges I have been called a fool, a jerk, and a >liar. Needless to say, this bothers me a bit. It should bother you, considering that you are right. Curiously, I've sent notices to many people about this; none of the few replies I got were hostile. Using ARC on posted sources makes even less sense than using tar or cpio. It wastes resources. You have to have a magic program to decode it, and working versions of ARC for UNIX seem to be rarer than hen's teeth. (OK, so maybe someone has finally posted one that works. I gave up caring a long time ago.) Putting compression in that high a transmission level is just wrong. And how many different "levels" of ARC are there floating around out there? Shar files, on the other hand, are crude but effective. All you need to decode one is a text editor. For binaries, there's always uuencode when you need it. DEATH TO ARC ON THE USENET! <csg>
emigh@ecsvax.UUCP (Ted Emigh) (05/04/87)
One thing that ARC does give you is a checksum, so you can tell if you've gotten the file correctly. Often, I get shar files and extract the correct number of character, but the file is severely corrupted. Will putting checksums (or CRC's) in SHAR make it as effecient? When will it happen? -- Ted H. Emigh, Departments of Genetics & Statistics, NCSU, Raleigh, NC uucp: mcnc!ncsuvx!ncsugn!emigh or mcnc!ecsvax!emigh internet: emigh%ncsugn.ncsu.edu or @ncsuvx.ncsu.edu:emigh@ncsugn.ncsu.edu BITNET: NEMIGH@TUCC DOMAIN: emigh%ecsvax.ncecs.edu
stever@videovax.UUCP (05/05/87)
In article <3540@mirror.TMC.COM>, Rich Salz (rs@mirror.TMC.COM) writes: > Last week I posted an article in comp.sys.amiga saying that posting > ARC'd files increases the transmission costs to most Usenet sites. In > subsequent e-mail exchanges I have been called a fool, a jerk, and a ^ ^^^^ ^^^^ ^^^^^^ ^ ^^^^ ^ ^^^^ ^^^ ^ > liar. Needless to say, this bothers me a bit. ^^^^ > [ Rich then shows an example of why ARC is a poor choice for Usenet. ] Surely, with all the brains and talent that is represented by the Usenet community, we can do better than this! If you disagree with someone, do so in a reasonable manner, not like a squalling infant! People who stoop to name-calling are simply revealing their intellectual immaturity. Rich $alz has done a great deal for Usenet -- for all of us. Thanks to him, mod.sources (now renamed) has kept the sources flowing. If anyone is in a position to know what will reduce transmission costs over the net, it is Rich! Those of you who denigrated Rich (you know who you are) should take off your diapers and grow up! Steve Rice ----------------------------------------------------------------------------- new: stever@videovax.tv.Tek.com old: {decvax | hplabs | ihnp4 | uw-beaver | cae780}!tektronix!videovax!stever