Why Live In The Past
In his article "Why Change What Works?" (PC World, Sep 89), Hal DuPrie shows a very biased comparison of ARC and PKZIP. SEA did sue Phil Katz, which DuPrie fails to mention was a major blow to the shareware world. Regardless of the fact that many people stopped using ARC because of that suit, ZIP was an instant success in the bulletin board world.
Comparing ZIP to ARC, DuPrie states that ZIP has only slightly better (by 5 to 10 percent) compression. I have found that the Beta version of PKZIP v1.0 usually beats ARC by at least 15 percent. It's inefficient to maintain files on a bulletin board in multiple formats. So why not convert all the files to ZIP files (as many bulletin boards are doing). The question of keeping track of multiple file formats is easily solved by a number of shell programs that detect which compression format was used and what program will uncompress the file; no problem for novices.
I called PKWare and was informed that both an Amiga and VAX VMS version are under development and should be released by the end of the year. However, ZIP lacks the machine support that ARC has. I would score ARC a point over ZIP in that respect. ZIP 0, ARC 1.
The source code for ARC was released only YEARS after the program had been on the market, and the ARC format still belongs to SEA. However, Phil Katz has released the algorithms for ZIP compression and generously released the ZIP format into the public domain. Several programmers have released source code in both Pascal and C for creating/extracting ZIP files. That code can be ported to other machine formats as easily as the source to ARC. Shareware and public domain authors need not worry about infringement on any utilities they write for ZIP. The ZIP format's public domain status scores ZIP 1, ARC 1.
ZIP is reliable. Phil Katz enjoyed a healthy success with PKARC, so his reputation carries on in ZIP. The code is no more bound to be bug-ridden than ARC. The ARC 6.02 manual (p.28) reads "Version 6.0...is a substantial rewrite..." Why would a 'substantial rewrite' of ARC be any different than changing PKARC into PKZIP? In fact, ARC 6.02 failed to handle shared files at all (program aborted) where PKZIP and PKARC both correctly handled the situation. ARC 6.02 also does a poor job of housecleaning if the program aborts. A very large temporary file and a zero byte file remained in the directory after using Ctrl-Break. I won't count the rewrite against the new version of ARC. But it should be of concern to heavy users of ARC. However, the incorrect handling of shared files is of concern to network users. ZIP 2, ARC 1.
The remark about keeping both ARC and ZIP on a system for compatibility is rather meaningless. The savings ZIP offers definitely earn its keep. Whether having both programs on a system would "...probably take up more space than the savings..." is beside the point and is relative only to the price of eggs. In fact, for those users that do not require compatibility (for instance users that do not log onto bulletin boards) ZIP is the clear winner because of its better compression. ZIP 3, ARC 1.
ARC is no more compatible with ZIP than ZIP is compatible with ARC. ARC is not compatible with anything but ARC. If it had not been for SEA suing PKWare, a newer PKARC would probably still be compatible with PKARC. However, that is a moot issue that arose entirely because of the law suit. Somehow ARC is now compatible with the so called 'Deviant' Squashing method used by PKARC. Is it coincidental that this happened after SEA gained the rights to the source code for PKARC? For that I score against SEA, and software consumers should be aware that SEA still has the option to sue any of their other competitors. ZIP 4, ARC 1.
Moreover, I find ARC no easier to use than ZIP. A command line interface can hardly be considered 'comfortable.' Any of the popular shells in the BBS community can make either utility equally friendly. That's honest. Neither program scores a point for their 'interface'.
Finally, I see Phil Katz as the Steve Jobs of compression programs. He came up with some innovative ideas for PKARC and came through again with PKZIP. And he even sells ZIP for less than what SEA sells ARC. The expandability of the ZIP format shows it will be around without change longer than ARC was without change. ARC now offers better flexibility and marginally better compression than previous versions, at the expense of compatibility. Darn the luck. ZIP will become the de facto standard and replace ARC, as major bulletin boards (such as Exec-PC) and on-line services (such as CompuServe) support ZIP. ARC will continue to follow in the footsteps of ZIP.
5,518,337 bytes, consisting of over 2 megs of Pascal source code and over 2 megs of binary .EXE files resulted in the following: Program Compressed Ratio ---------- ---------- ----- ARC 6.02 3,215,042 41.7% PKARC 3.61 3,092,035 43.9% PKZIP 1.0á 2,419,261 56.1%
Matthew Palcic is a Computer Engineering student in Dayton, OH and is the author of several shareware programs including DynaBoot.