13001_6035@uwovax.uwo.ca (04/10/89)
I realize that the time for voting is past, but I
still want to have my say. While it appears fairly
decided that comp.binaries.ibm.pc switch to ZOO for
archiving, there are a few more factors to consider:
(1). New users and occasional users.
These may not wish to get the large ZOO source
package to decode the listings.
(2). Cost.
The cost of extracting a UseNet posting and
transmitting it from the local system to a PC
is directly proportional to the file size.
From what I have read, ZOO does little better than
PKARC in data compression. Some of us have limited
budgets.
(3). Time.
The longer a listing is, the longer it takes to Kermit
the file to a PC.
I have tried LHARC (v1.00), PKZIP (v0.92), and PKPAK (v3.61) two test files.
File MT3.DOC is a 61489 byte text file (one issue of a mathematical magazine,
in Lotus Manuscript format). File MORIA.EXE is 350791 bytes worth of a game
(executable).
Compression type size compression time extract time
------------------------------------------------------------------------------
(file=MT3.DOC)
none 61849 -- --
LHARC (freezing) 25093 51s 26s
PKZIP -ex (reducing) 27772 36s 11s
PKZIP (shrinking) 30726 15s 13s
PKZIP -ea2 (reducing) 29748 25s 11s
PKPAK (squashing) 32057 14s 11s
(file=MORIA.EXE)
none 350791 -- --
LHARC (freezing) 169424 5:22 2:50
PKZIP -ex (reducing) 179554 2:30 0:41
PKPAK (squashing) 232565 1:10 0:51
The greatest time load is the time of data transfer from the Local server
to the PC (even with a 9600 baud groundline). Thus whilst LHARC is much
slower than any other program, it may actually allow files to be up and
running faster than anything else, by reducing the file-size. Furthermore,
LHARC can produce self-extracting COM files, by adding only 2Kb (!) to the
size of the archive. This would mean that new-users could conveniently get
files, without even having a de-archiver.
PKZIP is my next choice, as it does provide very fast decompression.
While I have not tested ZOO, I do not think (based on what I have heard about
it) that it would do much better than PKPAK 3.61, which I would say was fairly
clearly the user.
Please cut down UseNet access costs, and switch to LHARC or PKZIP.
Alex Pruss and Patrick Surry
Dept of Physics, University of Western Ontario.