cck@deneb.ucdavis.edu (Earl H. Kinmonth) (08/13/88)
In article <17362@gatech.edu> jkg@gatech.UUCP (Jim Greenlee) writes: >In reading the discussion about ARC vs. PK{ARC|PAK} vs. ZOO, a thought occurred >to me that I have never seen addressed in this newsgroup: > > What exactly is it that people have against compressed tar files? > You cannot make a selective extraction from a compressed tar file without decompressing the whole thing. With many small files, tar adds a fair amount of garbage to pad out blocks. cpio could be used for the same purpose without the garbage. cpio and tar are not that readily available for MSDOS although versions of both do exist.
brickman@cme-durer.ARPA (Jonathan E. Brickman) (08/14/88)
In article <17362@gatech.edu> jkg@gatech.UUCP (Jim Greenlee) writes: >In reading the discussion about ARC vs. PK{ARC|PAK} vs. ZOO, a thought occurred >to me that I have never seen addressed in this newsgroup: > > What exactly is it that people have against compressed tar files? > >Using compress and tar would seem to solve a lot of compatibility problems >since it is already an established "standard" in many environments (am I >pre-supposing again? :-). Does anybody have any thoughts on this, or is it >just "a dumb idea"? > > Jim Greenlee Compressed tar files are reasonably standard on Unix machines, but on no other. Sure, compression and tar programs for MS-DOS machines exist, but the examples I've seen have all been cumbersome and slow. Just in case you didn't realize it, the PKARC format was extremely compact (consistently around 50% compression), and extremely fast (it would unpack a 100,000 byte file faster than the copy command would copy it). I'll take that kind of speed and consistency, thank you very much, over any kind of tradition. ||Jonathan E. Brickman
bobmon@iuvax.cs.indiana.edu (RAMontante) (08/14/88)
Since "this machine's" ULTRIX allocates disk blocks 1K at a time, I average about 500 bytes of internal frag. loss per file, regardless of compression. If I compress a file less than 1K, there is no change (compress knows it made a smaller file, but doesn't know that the same block is allocated). With many smallish files this can be a problem; the recent Omega game posting is an example. So here's my sick solution: I compressed all the files. Then I used the UNIX-based ARC to jam them all into an archive. ARC didn't do any compression, since compress had already done a superior job, but it did flush all that internal fragmentation. Plus supplying file-by-file CRCs, allowing individual extraction or updating, maintaining an original-date directory.... BTW running ARC on the original files produce a noticeably larger archive; compress really did do a much better job, even after ARC collected everything together and added its own internal info. -- -- bob,mon (bobmon@iuvax.cs.indiana.edu) -- "Aristotle was not Belgian..." - Wanda
ruiu@dragos.UUCP (Dragos Ruiu) (08/19/88)
In article <17362@gatech.edu>, jkg@gatech.edu (Jim Greenlee) writes: > In reading the discussion about ARC vs. PK{ARC|PAK} vs. ZOO, a thought occurred > to me that I have never seen addressed in this newsgroup: > > What exactly is it that people have against compressed tar files? > > Many have been urging the adoption of ZOO because it supports hierarchical > archives - tar handles this. > Jim Greenlee - Instructor, School of ICS, Georgia Tech jkg@gatech.edu The advantage to ZOO as I see it is that it will support this on all its versions, and would be transportable, whereas compress/tar would assume you could get those utilities on MSDOS/VMS/Whatever...(Which you can, but it is more difficult than tracking down ZOO, which is conveniently enough authored by the moderator.) It would add to the hassle, and it would make it more difficult for people who are just starting out. (P.S. I agree with everyone else that ZOO sources should be posted...) -- Dragos Ruiu ruiu@dragos.UUCP "We will probably never announce a processor ...alberta!dragos!ruiu as a RISC processor."-Ken Olsen ...wanna bet?
root@mjbtn.UUCP (Mark J. Bailey) (08/19/88)
In article <2713@ucdavis.ucdavis.edu>, cck@deneb.ucdavis.edu (Earl H. Kinmonth) writes: > > > > What exactly is it that people have against compressed tar files? > > > You cannot make a selective extraction from a compressed tar file > without decompressing the whole thing. With many small files, tar adds > a fair amount of garbage to pad out blocks. I hate to disagree with you here, but at least on Unix, I use the following method for selective extractions, and it works: uncompress <file.tar.Z | tar xvf - file.in.tar As a matter of fact, by leaving off the file.in.tar above, you can uncompress and untar a file all in one sweep. It also works well with the tar options 'tvf -'. This may be a little cumbersome, but it follows unix ideology and it works. Mark. -- Mark J. Bailey "Y'all com bak naw, ya hear!" USMAIL: 511 Memorial Blvd., Murfreesboro, TN 37130 ___________________________ VOICE: +1 615 893 4450 / +1 615 896 4153 | JobSoft UUCP: ...!{ames,mit-eddie}!killer!mjbtn!root | Design & Development Co. FIDO: Mark Bailey at Net/Node 1:116/12 | Murfreesboro, TN USA