hes@ecsvax.UUCP (Henry Schaffer) (12/01/86)
I tested the PKARC and PKXARC programs obtained from the self-un-packing
PKX33A12.COM program, and compared their performance to ARC version 5.10.
I used three "data" files in my test: ARC.DOC, the doc file for arc;
ARC.EXE, the arc program itself; and T, a text file I had handy.
-------ARC------------ --------PK-----------
In Out In Out
File Size Time % Size Time Time % Size Time
ARC.DOC 60168 82 59 24711 40 14 58 25513 8
ARC.EXE 32051 64 20 27519 30 9 32 21904 6
T 14808 24 53 7071 11 5 53 7102 2
All tests were run totally in ramdisk (both programs and all files) on a
vanilla IBM PC. The first "Time" is time to put the file in a new archive.
The % is the crunch factor, the "Size" is the size of the resulting
archive and the second "Time" is the time to extract the file from the
archive. All times are done with a sweep second hand, and include the
program load time (at most 1 second) and are in seconds.
The PK programs seem to run about 5x as fast as the ARC program, and to
produce compression which varies from slightly worse to considerably
better than what ARC does.
The two programs do seem to be "compatible" in that they seem able to
unpack each other's archives. (I only made a cursory test of this.)
--henry schaffer n c state univbobmon@iuvax.UUCP (Robert Montante) (12/02/86)
How about some pointers to BBS's that have this program? I've heard good things about it, but I can't find it. (I know it's available in an archive "here" somewhere [ARPAnet?] but I don't have access to that archive because I'm on the wrong net.) (p.s. If anyone is willing to email it, I'd certainly appreciate that, too...)
mdf@osu-eddie.UUCP (Mark D. Freeman) (12/05/86)
I think I have found a bug in this program. If I have a directory full of files, one of which is an archive named bar.arc, and I issue the command "pkarc a foo *.*" I should end up with an archive called foo.arc, which contains bar.arc and all the other files that are in the directory. Now try to extract bar.arc from foo.arc. It won't work, giving a phantom error which combines 'bar.arc' and the name of the first file in the archive. So, be careful about not archiving archives. -- < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > > Mark D. Freeman mdf@osu-eddie.uucp InfoSolv Corporation mdf@Ohio-State.arpa 6099 Riverside Drive ...!cbosgd!osu-eddie!mdf Dublin, OH 43017 Guest account at The Ohio State University (614) 761-8444 < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > >
madd@bucsb.bu.edu.UUCP (Jim "Jack" Frost) (12/05/86)
In article <2690@osu-eddie.UUCP> mdf@osu-eddie.UUCP (Mark D. Freeman) writes: >I think I have found a bug in this program. If I have a directory full >of files, one of which is an archive named bar.arc, and I issue the >command "pkarc a foo *.*" I should end up with an archive called >foo.arc, which contains bar.arc and all the other files that are in >the directory. > >Now try to extract bar.arc from foo.arc. It won't work, giving a >phantom error which combines 'bar.arc' and the name of the first file in >the archive. > >So, be careful about not archiving archives. > Which version of ARC are you using? When a friend and I, using ARC 5.12, I think, tried ARCing an ARC file, it just came up with another ARC file (which unpacked just fine) that was somewhat longer than the other (I would assume that this is because the compression used was optimal the first time, BTW). I suppose my I may be losing my mind (it's been months since we did this) but I am pretty sure we unpacked it. Anyone with ARC handy care to retry it? It may be a small bug (little is still bad, tho) that only shows up occasionally. Also, the ARC file we used was nearly 1Mb unpacked, about 350K packed. -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% - Jim Frost * The Madd Hacker - UUCP: ..!harvard!bu-cs!bucsb!madd | ARPANET: madd@bucsb.bu.edu CSNET: madd%bucsb@bu-cs | BITNET: cscc71c@bostonu -------------------------------+---+------------------------------------ "Oh beer, oh beer." -- Me | [=(BEER) <- Bud the Beer (cheers!)
bobmon@iuvax.UUCP (Robert Montante) (12/07/86)
First, Thanks to all who've responded to my request for a copy of PKARC
-- including Wayne Aiken, who sent me a copy from an address that I can't
send a reply to ("Bad mailer-daemon! No biscuit!"). Having tried it, I
can now add my two cents' worth to the following...
>Mark D. Freeman writes:
>>I think I have found a bug in this program. If I have a directory full
>> [...]
>>Now try to extract bar.arc from foo.arc. It won't work, giving a
>>phantom error which combines 'bar.arc' and the name of the first file in
>>the archive.
Jim "Jack" Frost writes:
>Which version of ARC are you using? When a friend and I, using ARC
>5.12, I think, tried ARCing an ARC file, it just came up with another
>ARC file (which unpacked just fine) [...]
I was also able to combine (small) .ARC files from ARC v5.12 with PKARC v1.2
Indeed, all possible permutations (PKARC archiving its own .ARC files, archiving
ARC's files, and vice versa) successfully unpacked again. I used text files.
If archiving of archives is a necessity, the speed advantage of PKARC is such
that it may be worth testing each occurrence for re-unpackability (?)
While I'm at it...I found PKARC to be much more compressive on binary files,
equivalent or a bit worse on text files. I ran across an archive some time
ago that REALLY intrigued me -- ARC v5.12 was able to un-archive it, but it
had been created by something that did Lempel-Ziv compression to all the files
in the archive. The remarkable thing was that some of these files came out
_Larger_ as a result of the L-Z compression, yet the overall .ARC file was
much _smaller_ than ARC v5.12 was able to achieve when I re-ARCed the files
to see what would happen. Yow!
*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*-=-*
RAMontante, iuvax!bobmon
Computer Science "Have you hugged ME today?"
Indiana Universitymdf@osu-eddie.UUCP (Mark D. Freeman) (12/08/86)
Summary: In <635@bucsb.bu.edu.UUCP> madd@bucsb.bu.edu.UUCP (Jim "Jack" Frost) writes: >In article <2690@osu-eddie.UUCP> mdf@osu-eddie.UUCP (Mark D. Freeman) writes: >>I think I have found a bug in this program. If I have a directory full >>of files, one of which is an archive named bar.arc, and I issue the >>command "pkarc a foo *.*" I should end up with an archive called >>foo.arc, which contains bar.arc and all the other files that are in >>the directory. >> >>Now try to extract bar.arc from foo.arc. It won't work, giving a >>phantom error which combines 'bar.arc' and the name of the first file in >>the archive. >> >Which version of ARC are you using? PKARC, as referenced in the subject line, not ARC. -- < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > > Mark D. Freeman mdf@osu-eddie.uucp InfoSolv Corporation mdf@Ohio-State.arpa 6099 Riverside Drive ...!cbosgd!osu-eddie!mdf Dublin, OH 43017 Guest account at The Ohio State University (614) 761-8444 < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > >