home *** CD-ROM | disk | FTP | other *** search
-
-
-
-
-
-
- ╔═══════════════════════════════════════════════════╗
- ║ ║
- ║ BATTLE OF THE COMPRESSIONS ! ║
- ║ ║
- ║ Arj, Pkzip, Lha and Arc go head-to-head ║
- ║ in an intense, heated battle of compression ║
- ║ ║
- ║ ║
- ╚═══════════════════════════════════════════════════╝
-
-
-
-
-
-
-
-
-
-
-
- Since the release of ARJ and the new version of LHA, there has not
- been an objective comparison of all the current compressions. Obviously,
- ARJ has its own comparison within the file ARJ.DOC, but once reading it,
- found them to be rather subjective and laxed in explanation and true
- experimentation.
-
- Because of this, I ran several extensive tests with all four of the
- major compression programs:
-
- ARC 6.02
-
- PKZIP 1.1
-
- LHA 2.12
-
- ARJ 2.2
-
-
-
- Other methods, such as PAK, ZOO, LHARC, and the remainder umteenth
- compressions were not included because they are out-dated, old and not
- widely utilized.
-
- ┌─ ─┐
- │ARC performed bad enough, I didn't want│
- │to horridly embarrass ZOO, PAK, etc ...│
- └─ ─┘
-
-
- The study is split into four groups:
-
- 1) Single file, regular compression methods.
-
- 2) Single file, self-extracting compression methods.
-
- 3) Multiple file, regular compression methods.
-
- 4) Multiple file, self-extracting compression methods.
-
- NOTE: LHA and ARJ are the only SFX files included, considering the
- rather large size of both ZIP2EXE and MKSARC self-extractors.
-
- In the self-extracting mode, the following commands were used:
-
- a) LHA S FILENAME.EXT
-
- b) ARJ F -JE1 FILENAME.EXT
-
-
- Also, speed wasn't considered in my experimentation because
- all methods rated around .043 second maximum difference.
- I did not consider that ratio large enough to justify a
- "speed" comparison.
-
-
- In the comparisons, the boxes are separated into six sections, and
- four sections, respectively.
-
- FILE refers to the type of file being compressed. A large document
- file would be something like 'ARJ.DOC' or the like.
- Large executable file would be 'TC.EXE' from Turbo C++.
-
- SIZE is the original size of the file(s) being compressed.
-
- LHA,ARJ,ARC,ZIP are the compression methods used.
-
-
-
- ------------------------------------------
- THIS ARCHIVE WAS COMPRESSED WITH LHA 2.12
- ------------------------------------------
-
-
- =============================================================================
-
- Our first trial involves six different file types of various sizes.
-
-
-
-
- Regular Compression
- """""""""""""""""""
-
- │ File │ Size ║ LHA │ ARJ │ ARC │ ZIP │
- ├────────────┼────────╫────────┼────────┼────────┼───────┤
- │ │ ║ │ │ │ │
- │ LARGE DOC │ 87289 ║ 30553 │ 29294 │ 43062 │ 30945 │
- │ │ ║ │ │ │ │
- │ MEDIUM DOC │ 31399 ║ 11394 │ 11267 │ 15253 │ 11673 │
- │ │ ║ │ │ │ │
- │ SMALL DOC │ 13723 ║ 5554 │ 5520 │ 7227 │ 5675 │
- │ │ ║ │ │ │ │
- │ LARGE EXE │ 876480 ║ 438072 │ 430285 │ 616481 │ 462487│
- │ │ ║ │ │ │ │
- │ MEDIUM EXE │ 211168 ║ 109133 │ 107306 │ 147837 │ 113582│
- │ │ ║ │ │ │ │
- │ SMALL EXE │ 88834 ║ 48235 │ 47827 │ 65570 │ 51009 │
- └────────────┴────────╨────────┴────────┴────────┴───────┘
-
- 1) In all single-file compressions, both document and executable were
- compressed the greatest by ARJ 2.2.
-
- 2) ARJ compressed:
- a) Large Doc the best (66.4 %)
- b) Large Exe the worst (50.1 %)
-
-
-
-
-
- Now, once the above files are converted to self-extractors:
-
-
-
-
- Self-Extracting File
- """"""""""""""""""""
-
- │ File │ Size ║ LHA │ ARJ │
- ├────────────┼────────╫────────┼────────┼
- │ │ ║ │ │
- │ LARGE DOC │ 87289 ║ 32186 │ 34639 │
- │ │ ║ │ │
- │ MEDIUM DOC │ 31399 ║ 13027 │ 16612 │
- │ │ ║ │ │
- │ SMALL DOC │ 13723 ║ 7225 │ 10981 │
- │ │ ║ │ │
- │ LARGE EXE │ 876480 ║ 439705 │ 435743 │
- │ │ ║ │ │
- │ MEDIUM EXE │ 211168 ║ 110803 │ 112765 │
- │ │ ║ │ │
- │ SMALL EXE │ 88834 ║ 49904 │ 53286 │
- └────────────┴────────╨────────┴────────┴
-
- 1) In self-extracting files, LHA 2.12 clearly outperformed ARJ 2.2,
- especially in SMALL DOC compression. The docs were 34% tighter
- in the LHA files than the ARJ.
-
- 2) In the MEDIUM EXE files, both compressions rated about the same.
-
- 3) LHA compressed:
- a) Large Doc the best (63.1 %)
- b) Small Exe the worst (43.8 %)
-
-
-
-
-
-
-
- ===========================================================================
-
- Our second trial involves multiple files being compressed using the
- regular compression method. Note that the multiple document files
- involves 190 text files, and multiple executable files involves
- 19 binary files.
-
-
-
-
-
-
- Regular Compression
- """""""""""""""""""
-
- │ File │ Size ║ LHA │ ARJ │ ARC │ ZIP │
- ├────────────┼────────╫────────┼────────┼────────┼───────┤
- │ │ ║ │ │ │ │
- │ MULTI DOC │ 1344961║ 476168 │ 474106 │ 715585 │ 511631│ 190 TEXT FILES
- │ │ ║ │ │ │ │
- │ MULTI EXE │ 552145 ║ 344284 │ 343849 │ 418390 │ 356818│ 19 BINARY FILES
- │ │ ║ │ │ │ │
- └────────────┴────────╨────────┴────────┴────────┴───────┘
-
- 1) Again, ARJ beats out the others with:
- a) Multi Doc 64.8 %
- b) Multi Exe 37.7 %
-
-
-
-
-
- Now, once the above files are converted to self-extractors:
-
-
-
- Self-Extracting File
- """"""""""""""""""""
-
- │ File │ Size ║ LHA │ ARJ │
- ├────────────┼────────╫────────┼────────┼
- │ │ ║ │ │
- │ MULTI DOC │ 1344961║ 477224 │ 474106 │
- │ │ ║ │ │
- │ MULTI EXE │ 552145 ║ 345863 │ 349194 │
- │ │ ║ │ │
- └────────────┴────────╨────────┴────────┴
-
- 1) And again, LHA defeats ARJ with executable files,
- whereas ARJ does best with multiple documents.
-
-
-
-
-
- =============================================================================
-
- The conclusion of this test is obvious:
-
-
- ARJ works best when compressing files with the regular .ARJ extension.
- However, if you want to create a self-extracting file, LHA clearly
- outperforms ARJ 2.2 in all the trials.
-
-
-
- ╒══════════════════════════════════════════════════════════════╕
- │ Using the conclusion: IS ARJ WORTH THE EXTRA 3% DIFFERENCE ? │
- ╘══════════════════════════════════════════════════════════════╛
-
- Noticing that ARJ 2.2 is superior to LHA in its compression, we
- come to the question of "Is it worth it to change ?".
-
- And to answer that, we could take several viewpoints:
-
- 1) Any difference is better ! If ARJ gives me 5% better compression,
- and the next compression method gives me another 5%, then all that
- adds up, especially if I use it to save on disk space.
-
- 2) Changing compressions is a hassle, who cares about the 5% difference.
-
- 3) I'll change my compression when I get better than a 5% difference,
- maybe a 10% or 20% better compression.
-
-
-
- and finally, other observations:
-
-
- 1) ARJ 2.2 jams up once compressed by an executable file compressor,
- whereas LHA 2.12 does not.
-
- 2) ARJ 2.2 is radically larger than LHA 2.12.
-
- 3) ARJ 2.2 has many more features than LHA, like sorting files inside,
- searching for a text string within the archive, etc.
-
- 4) In the .H files, it has M. ADLER and R. JUNG as the creators of
- ARJ. Is that a joke ? I'm a psychologist, and think it's kinda
- cute, but...
-
-
-
-
- Comments or suggestions ??
-
-
- Write me:
- DAVE SMITH
-
- Compuserve ID 71441,2723
-