Category : File Managers
Archive   : NARC40.ZIP
Filename : NARC.DOC

NARC (tm) zip - A STAND-ALONE DE-ZIP UTILITY
(no other files required)


NETWORK COMPATIBLE



Documentation for NARC.EXE


Written by Gary Conway and Shawn Merrick


Infinity Design Concepts, Inc.


Louisville, Kentucky


Copyright (c) 1987-90, All rights reserved


Version 4.0

























NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.










NARC.EXE is a shareware program. NARC.EXE is and will remain
the property of the author. This program may not be used in
any connection with commercial ventures, nor as a sales aid,
without the expressed written consent of the author. All
rights are reserved.


Infinity Design Concepts, Inc.
1052 Parkway Drive
Louisville, Kentucky 40217-2333
Voice: (502) 636-1234
BBS: (502) 635-5471
CIS: 72657,3006

Member IEEE AOPA
PCCL WINGS
KKUG
NSPE


A very special thanks to Curt Edwards, for running the
SoftStone BBS system for many years. Curt has decided to
retire from SYSOPing after a long time of helping others, we
at IDC would like to thank Curt for his endurance, his
patience, his help and most of all, for the Stone..



NARC is a trademark of Infinity Design Concepts Inc.





















NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.











REGISTRATION


If you find yourself using NARC, please take the time to do
the right thing and that is register your copy. You have been
provided the opportunity to freely test the program before
even thinking about registering. This is only fair, so, in
fairness, you should reciprocate and register your copy, if
you continue using the program.

Why register ?

1) You get the NARCCFG.EXE program for customizing NARC.
2) You get notification of updates to all IDC software.
3) You get patch table information.
4) You get online phone support.

Customers residing outside of the United States of America
should send a check or money order drawn on a U.S. bank or
in U.S. funds or use a credit card.

You will find the registration form in the archive with this
document under the name REGISTER.FRM. Please use this form
for registration.


Infinity Design Concepts, Inc.
1052 Parkway Drive
Louisville, Kentucky 40217-2333
(502)-636-1234

The user is granted the right to make unlimited copies of
the SHAREWARE versions of the program (except NARCCFG.EXE
and NARCCFG.DOC) and to distribute these copies as desired
without profit, EXCEPT that Infinity Design Concepts, Inc.
reserves the SOLE right to distribute the program(s) for
profit.

NOTE: The word ZIP is used throughout this document and
almost all functions that apply to ZIP files, can
be applied to ARC and ARK files as well.


NOTE: NARCCFG.EXE is NOT a shareware program and may not
be released to the general public in any form
under any circumstances whatsoever.



NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

.............................................................................
........................ ............................
........................ TABLE OF CONTENTS ............................
.............................................................................

Page
FEATURES UNIQUE TO NARC..............................1
WHAT IS SHAREWARE.................................. 2
NARC ? WHAT IS IT ANYWAY........................... 3
ACKNOWLEDGEMENTS................................... 3
ABOUT ZIP FILES.................................... 4
COMPATIBILITY...................................... 5
Archive Storage Methods Supported By NARC...... 5
Packing
Squeezing
Crunching
Squashing
ZIP Storage Methods Supported By NARC.......... 5
Storing
Reducing
Shrinking
Imploding
OVERVIEW........................................... 6
Video Modes ................................... 6
Mouse Support ................................. 6
Sound ......................................... 7
Environment Variables.......................... 7
HELP............................................... 9
COMMANDS........................................... 9
Extract Command................................ 9
View Command................................... 10
Print Command.................................. 11
ZIP-wind Command............................... 12
DRV-wind Command............................... 12
DirTree Command............................... 12
Quit Command................................... 12
ALTERNATE COMMANDS................................. 13
Function keys.................................. 13
Find Command................................... 14
Kill File Command.............................. 14
Page UP, DOWN, HOME,END........................ 14
Using extensions besides ZIP and ARC........... 14
Calling IDCshell.EXE........................... 14
Command Line Processor..(new).................. 15
Trivia status screen........................... 15
OPERATING HINTS AND SHORTCUTS...................... 16
ERROR MESSAGES..................................... 17
ARCHIVE FILE FORMATS AND GENERAL INFORMATION....... 19
Packing (RLE).................................. 19
Imploding...................................... 20
Huffman coding (SQUEEZING)..................... 21
LZW Compression.......... ..................... 25
DETAILS OF STORAGE VERSIONS (ARC).................. 27
Archive file Header Structure.................. 29
HASHING............................................ 30
CRC - calculations............................. 31



NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.


ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
FEATURES UNIQUE TO NARC
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ


1) allows random access to ZIPfiles. ALL of the other archive
utilities allow ONLY serial access.
2) allows viewing DOC files before extraction. This feature
allows you to determine which, if any of the subfiles that
you want to extract. This can save considerable time and
disk clutter when you extract only the necessary files.
3) has a view feature that allows page up, page down, home
end and arrow key movement through a file. Binary files
can also be viewed in HEX mode. Also includes a text
search function.
4) allows printing files from an archive without extracting
the file first. Supports many printing options.
5) allows users with floppy systems to load NARC once and
examine many archives from all floppy drives. Other
archive utilities must reload (and reread the arc) for
each command.
6) directory tree allows one to easily move from directory
to directory or drive to drive.
7) large archives are much easier to deal with in NARC since
you may page through the subfiles.
8) mouse support for all functions.
10) First and still the ONLY menu driven decompression package
with all data compression routines built in !





























NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 1

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
WHAT IS SHAREWARE
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

Shareware is a concept for marketing software, whereby the
author takes the only risks. The purchaser is allowed to take
the software home without spending a dime and trying it out.
If the software is found to be useful and is used, then the
author expects the end user to buy it. From the author's
standpoint, he is so confident that you will like his
package, that he is willing to GIVE you a copy on the chance
that you will take it home and find it useful and want to buy
it. There are many advantages to registering shareware
software. First and foremost is the fact that unless
shareware authors are supported, they cannot maintain the
high standards in software that they do, and the prices
cannot be maintained at low levels. It takes financial
support for any company to continue to develop quality
software. The benefit to the author is that he can distribute
his software very economically, thus allowing low prices on
the software. Another benefit of shareware, is that shareware
authors are much more accessible than commercial authors and
that means better, more informative support to you. When was
the last time you called Micropro and got to speak to the
author of WordStar ?

How many times has Microsoft sent a nice piece of software to
your local BBS and asked you to try it out ? Shareware is a
vital and benefical concept, but it requires your support.
Look at it this way, when you send in a shareware
registration for $20, ask yourself just how many hours of
that authors time are you buying with that money, and where
else could you get such a bargain ?

























NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 2

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
WHAT IS IT ANYWAY ?
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ


NARC is a menu driven de-archive facility, written entirely
in assembler. NARC allows you to easily move from ZIP file to
ZIP file, with the option of viewing, printing, extracting or
deleting the subfiles from the archive. The program may be
operated from the mouse or the keyboard. Menus are of the
musical popup variety to add a little "TechNoFlash" to the
proceedings. NARC is the culmination of about 18 months of
frustrating effort and 13000 + lines of 8088 source code.
NARC was developed using the Norton Editor and the Microsoft
Macro assembler on a Kaypro PC and an ALR 386/20.

Why....


Because I use a lot of ZIP files and PKzip and the clones are
reminiscent of the early Ward Christensen CP/M days in user
interface etiquette,I wanted something a little more flexible
and friendly to use. I would like to pause here for a second
and give a little credit to Mr. Christensen ( the Don Garlits
of CP/M ) for the fine FREE utilities he has given to ALL of
us over the years. The next time you do a modem transfer, you
can thank him and Keith Petersen for the original XMODEM from
which all others have transpired.

Why NARC...

It seemed like a good idea. Short for uN-ARC. The idea was
originally Bob Freed's.

Acknowledgments..

I would like to thank my parents for not killing me at any
early age and my wife for not killing me at a later age...

I would like to thank Bob Freed for his allowing me to
examine his Z80 code before writing NARC. Bob wrote UNARC for
the CP/M world. A special thanks to the following kind folks.


Curt Edwards Jerry Taylor
Chuck Crumpton Frank Roemer
Paul Bowling Dr. Matthew Witten
Paul Clarke Sherman Able
Carlisle Crutcher H.W. Straley
John R. Gernert Dudley Hoffman
Dr. Melvin Maron Bob Fergus







NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 3

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
ABOUT ZIP FILES
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ


The ZIP file format was developed as the natural extension of
the older LBR and ARC files. This new format allows much
greater data integrity than we have ever had before with
archive utilities. The 32 bit CRC implemented plus the use of
both a central directory and a distributed directory, offers
the very best archive integrity.

NARC ZIP 3.0 AND ABOVE SUPPORTS BOTH ARC AND ZIP FILES

The compression algorithms used in ZIP files are shrinking
and reducing. The shrinking algorithm is basically the same
as squashing and crunching used in ARC files, except, that
the algorithm has been fine tuned to eliminate the wasted
bytes in earlier implementations. The code length
responsibility has been shifted from the decompressor to the
compressor, which again offers an increase in compression
ratios. The total string table clearing encountered in ARC
files, has been replaced with a much more intelligent partial
clear, which, instead of eliminating all strings that have
been learned in the table, only parts of strings are
eliminated and used again.

The reducing algorithm is really a combination of two
distinct algorithms, statistical compression and run length
encoding (packing), just as crunching was a combination of
LZW and packing. The statistical compression method takes
advantage of certain statistical character pair occurences
common in the English language, but the algorithm also fairs
well with other languages and binary files.

The ZIP file format was developed by IDC and PKware and
several other individuals after much discussion. The LZW
algorithm improvements were readily visualized by all
parties, but the reducing and imploding algorithms were
developed by Phil Katz and PK-friends and has shown
themselves to be very powerful, as we have seen files
compressed as much as 99%.
















NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 4

ÍÍÍÍÍÍÍÍÍÍÍÍÍ
COMPATIBILITY
ÍÍÍÍÍÍÍÍÍÍÍÍÍ


NARC is compatible with all known "skrunching" algorithms,
that is up to and including Squashing. NARC is compatible
with PKarc, PKpak and PKzip. NARC supports both .ARC and .ZIP
files and recognizes the .ARK extension soon to be prevalent
in the CP/M world via Bob Freed's CP/M archive facility,
NOAH.


ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
ARCHIVE STORAGE METHODS SUPPORTED BY NARC
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ



Packing - all versions
Squeezing - Huffman Coding
Crunching - all versions (LZW encoding)
Squashing - one version


ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
ZIP STORAGE METHODS SUPPORTED BY NARC
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ



Reducing - statistical compression + packing
Shrinking - very intelligent LZW algorithm
Imploding - Shannon-Fano coding of sliding
dictionary length and distance
pairs


Note: LZW stands for Lempel-Ziv-Welch



















NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 5

OverView...

NARC first searches the default path for ZIP, ARC and ARK
files and if any are found they are displayed in the ZIP-
window on the left side of the screen. The arrow keys (or the
mouse), PgUP,PgDN, Home and End may be used to move the
cursor bar up and down the window, there are two ways to
select the highlighted ZIP file.

(1) Hit the ENTER key
(2) Press the left mouse button


After selecting the sub-file of interest, NARC displays all
of the ZIP sub-files and their statistics on the screen. You
are also given a menu bar at the bottom of the screen. You
may use the arrow keys or the mouse to move the cursor bar to
the desired selection and then select with the ENTER key or
the left mouse button. As the cursor bar is moved, you are
also given a brief description of the highlighted command. A
file may be located in the list by the now common ALT-letter
key combination. This applies to all menus in NARC. The
commands will now be discussed individually.

Note: You may also select any option from the command bar by
entering the first letter of the command.


The ESCape key will abort any operation or window.


Video Modes


NARC will determine whether a monochrome or color video
adapter is being used and act accordingly. If a CGA or EGA
card is detected, then NARC will select 80 x 25 color mode,
unless color has been turned off with NARCCFG. The users
video mode and cursor shape is saved when NARC is invoked and
restored upon exit or shell. On EGA/VGA equipped systems, the
F2 key may be used to toggle NARC into and out of 25 and 43
line modes.

Mouse Support

NARC will determine if a mouse driver is present. If so, then
NARC's mouse support routines are enabled and working at all
times within the program. The left mouse button ALWAYS
emulates the ENTER key (for selecting highlighted items) and
the right mouse button ALWAYS emulated the ESC key (for
aborting operations). Highlight bar movement can always be
accomplished with mouse motion. The sensitivity of mouse
motion can be changed with NARCCFG. The mouse routines were
written with the Microsoft mouse in mind and any mouse that
is compatible with the Microsoft mouse should work.



NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 6

Sound

NARC makes use of the 8253 programmable interval timer chip
and speaker port to create some gurgles,bleeps,whistles and
burps in general, that some folks like and some don't. The
chip is programmed for square wave output in mode 3. The
sound may be permanently disabled with NARCCFG. NARC
determines microprocessor speed and adjusts the sound
routines accordingly so that the sounds are pretty much the
same on all machines.


ENVIRONMENT VARIABLES


If you don't know what an environment variable is, then
continue reading. If you DO know, then skip to "The Goods"
section below.

Note that all of the environment information can be
configured directly into NARC zip 3.0 and above with NARCCFG.
However, this information can be overridden with the
environment variables. The environment variables are given
precedence so that if you take your NARCCFG'd version to
another machine, you can do a quick temporary set up with the
environment variables.


What are they ?

An environment variable is a convenient way for programs to
communicate with DOS or with each other through a commonly
shared spot(s) in memory. If you type "SET" at the
DOS prompt, DOS will respond by showing you the current spot
of shared memory which we call the environment. These text
strings can be created,read, modified and deleted at any time
by any program. Each variable has a certain format that must
be adhered to. The format is as follows,

variable_name = string

An example would be IDCED=C:\UTILITY\WS

Spaces are allowed before and after the equal sign. This
variable will mean absolutely nothing to DOS, but, to NARC,
it will show the way to your editor. The variable name here
is IDCED and the string is C:\UTILITY\WS which is the
pathname for finding WordStar on the disk. You can change
this string to anything you want by simply typing --- SET
IDCED= (and type your pathname for your editor). NARC uses
several of these environment variables to communicate with
you and you should set them up in your AUTOEXEC.BAT file.

What do I do if DOS says, "Out of Environment Space" ?

DOS reserves only so much space for the environment strings,
and if you have a long search path or a lot of strings, then

NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 7

you may run out of space. I can offer two suggestions, one
for DOS 3.1 and one for DOS 3.2. There is a "switch" that you
can add to your CONFIG.SYS file (If you don't have one, then
you should !). One of the lines in your config.sys file
should read Shell=C:\COMMAND.COM or something similar. You
can add the following to the end of this line to increase
your environment space.

DOS 3.1 - /E:nn where nn = 10 up to 62. This
number times 16 is the length
of your environment.

example: Shell=c:\command.com c:\ /P/E:10

This would reserve 160 bytes.

DOS 3.2 - add /E:nnnn where nnnn is the actual
number of bytes to reserve.

example: Shell=c:\command.com c:\ /P /E:160

This would reserve 160 bytes.


Environment variables can be set up in your Autoexec.bat file
as follows.


Set IDCARCPATH=c:\utility\idcshell.exe
Set IDCED=c:\wordproc\editor.ext



The Goods ..

IDCED=c:\editors\ws.exe

Use this variable to tell NARC where your
editor is located, so that when the F9 key is
pressed to invoke the editor, NARC will know
where to locate it. The drive and path and
editors name must be given (in upper or lower
case). The file extension (.COM or .EXE) MUST
be given.

IDCARCPATH=c:\utility\idcshell.exe

Use this variable to tell IDCshell where to
find IDCshell when the "I" key is pressed. As
with IDCED, the drive and path and filename
and extension must be given.

NARCHELP=c:\utility\narchelp.hlp

Use this variable to tell NARC where to find
the NARC help file when the F1 key is pressed.


NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 8

IDCTEMP=e:\ramdisk

Use this variable to tell NARC where it can
place temporary work files. If you have a
ramdisk set up, then this would be a good use
for it since all file operations will be MUCH
faster from the ramdisk.
THE HELP SCREEN


From the files screen, it is always possible to get a context
sensitive help screen. Hitting the F1 or F10 key will cause a
full screen of help to pop up that has information specific
to the command that the cursor bar was pointing to.

Once the help window is opened up, it is then possible to get
help on any of the commands. The Home key will take you to
the help index (as will the "I" key). The window may be put
at any time with the ESC key.

NOTE: There are several commands that are listed under
miscellaneous commands in the INDEX section of the
help screens, so be sure to check them out.


ÍÍÍÍÍÍÍÍ
COMMANDS
ÍÍÍÍÍÍÍÍ


ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
Extract Command
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

Selecting this option will cause another prompt to be
displayed, asking whether you wish to extract the highlighted
file or tagged files. (Files are tagged with the space bar).
"Point and shoot" here again as before. The ESC key or the
right mouse button will abort the operation. If the disk
becomes full, you will be informed and have the option of
aborting or continuing.


Highlighted File

When EXTRACT is selected, you will be asked for a drive/path
to extract the file to. If you simply hit ENTER or the left
mouse button, the file will be extracted to the default
drive/path. You may also enter any valid DOS drive/path. The
ESC key or the right mouse button will abort the operation.

NOTE: Control-R can be used to recall the last path that
files were extracted to.





NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 9

Tagged Files

The Space bar (or F3 key) is used to TAG the current file.
When a file is tagged, a diamond will be displayed on the
line with the current file in column 80. The space bar will
also unTAG a file, thus it is a "toggle". When the space bar
is pressed, a diamond will appear as described above and the
cursor bar will move to the next file. When the "TAGGED"
option is selected from the command line, all files that have
been tagged, will be extracted to the SAME drive/path.

After the file is extracted, it's date and time are set to
those contained in the ZIP file. The file is also checked for
size and CRC, if both of these do not match exactly what was
contained in the ZIP file header, then an error has occurred
and the user is notified. The files will also remain tagged
after the extraction.

If a zipfile contains stored subdirectory names, each
file will be displayed with the INFINITY symbol
(sideways 8) next to it. When you wish to extract a file
that has this "tag", NARC will ask you how you wish to
handle the stored subdirectory names. You have several
options..


(A)bort - aborts the extraction
(P)rompt - prompt and ask before creating any
subdirectories
(C)reate - create subdirectories as needed
(I)gnore - ignore the stored pathnames


If the (C) or (P) options are chosen, then directories
that are created, will be created starting with the
extract path that you have selected as the top
directory.


ÍÍÍÍÍÍÍÍÍÍÍÍ
View Command
ÍÍÍÍÍÍÍÍÍÍÍÍ

This option will display the currently highlighted file on
the screen. The PgUP, PGDN, Home and End keys, as well as the
cursor keys allow movement through the file. The file to be
viewed is first extracted to the default drive (or to
whereever the IDCTEMP variable points) to a file called
NARC.TMP. This file is deleted when the view is ended. The
extraction is performed due to the sequential nature of ZIP
files, which makes it nearly impossible to perform the page
up,down operations on the compressed file itself. If NARC
finds that there is not enough disk space or directory space
to create NARC.TMP, you will be asked for another drive where
the temporary file can be created.



NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 10

The newest addition to the view command in NARC 2.1 is the
text search command. Simply hit "S" and you will be prompted
for a string to find. When located, the text will be
highlighted on the screen, otherwise a message is issued
stating that the text was not found. The search is case
insensitive, that is you may enter upper or lower case
letters and NARC will find a match.

The mouse is also supported now in the view function.

ÍÍÍÍÍÍÍÍÍÍÍÍÍ
Print Command
ÍÍÍÍÍÍÍÍÍÍÍÍÍ

The print option will print the currently highlighted file.
After selecting the print option, you will be asked which
character pitch you want to print in. Enter the number that
you wish (or 0 for the default pitch) and the printer will be
set to that pitch.


NOTE: The printer strings that come installed with NARC are compatible
with EPSON printer strings. If you wish to install NARC with
your own strings, see NARCCFG.DOC for complete instructions.

After selecting selected the printer pitch, you will be shown
three more options for formatting the printout. Use the arrow
keys to move left and right. The space bar is used to toggle
the options ON or OFF. When you have finished and are ready
to print, hit ENTER (or left mouse button). The ESC key or
the right mouse button will abort at any time. The options
are;

Format - YES - This option causes NARC to format the
output with page breaks and page numbers.
NO - NARC does not format the file.

The following two options work independently of the Format option.

Strip High - YES - NARC will strip the high bit off each
character before it is sent to the
printer. Some word processors set this
high bit on some characters as a means of
text formatting. These characters will
print as garbage usually.
NO - NARC will not strip the high bit.

Strip Control - YES - NARC will strip all control characters
from the file before it is printed. This
is useful on files that have embedded
formatting characters, and you wish to
have NARC provide the formatting.
NO - NARC will not strip the control chars.

NOTE: On all of the following windows, the PG UP, PG DN,
HOME and END keys in addition to the cursor keys
allow movement through the window.

NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 11


ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
ZIP-wind Command
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
This option will display the window containing all of
the ZIP, ARC or ARK files in the current sub-directory.
Move the cursor bar up and down with the mouse or arrow
keys and select with the ENTER key or left mouse
button.

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
DRV-wind Command
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

This option will pop up a window that contains all of
the logical drives that DOS reports to NARC. Select as
before and the ZIP-window will be popped up so that you
can then choose a ZIP file to examine.


ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
DirTree Command
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

This command is totally new with release 2.1. The old
window approach won a lot of criticism and rightfully
so. I had a tree routine that I wrote some time ago,
that I just wasn't happy with speed-wise and didn't want
to stick that thing in NARC. I did have some nice ideas
for a new routine that would be much faster and slicker
all around and finally found the time to write it and
NARC release 2.1 has it, you may judge for yourself.

A visual tree is presented that displays all
subdirectories on the disk. Once the tree has been read,
NARC will not read it again, unless you have changed
drives.

From the tree, you can create and delete subdirectories,
select a new drive or subdirectory or return with no
action at all. The ESC key will abort the tree
operations.

ÍÍÍÍÍÍÍÍÍÍÍÍ
Quit Command
ÍÍÍÍÍÍÍÍÍÍÍÍ

The exit options have been changed so that now you will
be prompted for where you wish to exit the program. You
have the choice of exiting to the current directory or
the original directory where NARC was invoked from. This
can be made totally automatic with the NARCCFG program
in the registered version.





NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 12

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
ALTERNATE COMMANDS
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

The extra commands can be located on the help screens,
which are invoked by the F1,F10 or ? keys.

F1 - key
Invokes the context sensitive help system.

F2 - key
This key allows toggling between 25 and 43
line modes on EGA/VGA systems. On CGA/MONO
systems, it will have no effect.

F3 - key
This key will tag a subfile and move the
cursor bar on to the next subfile. This key
also has other functions, and they are also
shown on the screen when necessary.
F4 - key
The F4 key will print an image of the screen
less the advertisement and command lines.

F5 - key
Invokes the NARC-DOS command processor. You
may then enter any valid DOS command. When
finished, simply hit ENTER by itself and you
will be returned to NARC. You may also enter
"COMMAND" which will invoke a second copy of
COMMAND.COM, if the file COMMAND.COM is in
your search path. To return to NARC, you would
then type "EXIT".

F6 - key
This key will tag all of the subfiles in the
archive for subsequent extraction.

F7 - key
This key will invert all of the tags on the
subfiles, that is all files that were tagged
will become untagged and vice-versa.

F9 - key
This key will call your editor. Note that you
MUST have first set up the environment
variable IDCED before this function can be
used -OR- in NARC 3.0 and above, you may also
set the path to the editor permanently into
NARC with NARCCFG in the registered version.
See the section on environment variables.

F10 - key
This key will call up the context sensitive
help system.



NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 13

ALT-F10 - key
This sequence displays the trivia screen,
where your serial number and other information
is located.


(F)ind command.
Will prompt for a wildcard filename to find in
the sub-file list. Any number of characters
may be used, for example, you may enter a
single character and NARC will find the first
file whose name begins with that character.
Alternatively, you may enter a complete
wildcard specification and NARC will attempt
to find a match.

(K)ill file command.
Will remove the currently highlighted sub-file
from the archive. No additional disk space is
required for temporary files.

PgUP,PgDN,Home and End
These keys do what you might expect. These
functions work in all windows.

(S)et search spec command.
When NARC searches the current directory for
archive files, it normally searches for ZIP,
ARC or ARK files. The .ARK may be changed to
another extension (or wildcard extension) with
this command. All subsequent searches will use
the .ZIP and .ARC extension and the newly
defined mask. This command may be used from
the ZIP file window or from the subfile
screen.

The extension may also be passed from the
command line. Enter NARC /h for complete
information.

(I)DCshell command.
This command will run IDCshell.EXE. Note
that the environment variable IDCARCPATH must
be set up before this command will work,
unless IDCshell is in the current directory.
IDCshell is the companion program to NARC and
was be released in Aug - 88. This command
provides a simple link between the arcing
functions of IDCshell (and its copy / rename /
delete/print/view etc...) functions. Note that
NARC can also be called from IDCshell via its
"N" command along with a filename. Note that
in NARC zip 3.0 and above, the path for
IDCshell can be permanently configured into
NARC with NARCCFG in the registered version
instead of using the environment variables.


NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 14


Command line syntax:

NARC [drive:\path\subfilemask]
If the drive:path\subfilemask is given, then NARC will
display all files that match the given specification in
the zipfile window when NARC is first invoked. If no
subfilemask is given, then NARC defaults to displaying
all ZIP/ARC/ARK files if no subfilemask is not given.

If the drive and/or path is given, then NARC will search
the given drive or path for files that match the
subfilemask.


TRIVIA STATUS SCREEN

This screen shows some useful information about the
current NARC environment. It is invoked with the ALT-F10
command sequence from the subfiles screen.






































NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 15


Operating Hints and Shortcuts



When NARC is first invoked, it pops up the ZIP-window
showing all of the ZIP/ARC files in the current directory.

NOTE: When the ZIP file window is onscreen, you may also
choose another file extension beside .ARK with the "S"
command. See the section entitled "Using extensions
besides .ZIP and .ARC".

When there are no windows popped up, the "A" key or the
"Z" key will pop up the ZIP file window, regardless of
where the command line cursor bar is. This is handy when
you have a lot of ZIP's that you want to thumb through,
with just the 2 mouse buttons or ENTER and F2 keys, you
can look through a whole directory of ZIP files in
nothing flat ! Also along these lines, when the ZIP
window is on the screen, the right mouse button or the
ESCape key will exit the program.

NARC buffers 64k of the input file at one time, thus
speeding up file operations. The output buffer is 32k
and should be larger in the next version. NARC requires
about 220K of RAM to operate.






(This prime advertising space for rent)
























NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 16


Note that the arrows keys may be used to move the error
message box up and down the screen. This feature is provided
so that if the message covers up something on the screen that
you would like to see, you may move the box.


ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
Error Messages.
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

Memory Allocation Error.
- NARC allocates memory when it is invoked,
this says that DOS told NARC that there was
not enough memory left to run the program

ERROR: Extraction Failed due to CRC error, Hit ENTER
- After a file is extracted, the CRC contained
in the ZIP header is compared to the CRC that
NARC calculates, this message says that the
two were different. This is the CRC-16
polynomial in ARC/ARK files and CRC-32 in ZIP
files.

ERROR: Extraction Failed due to FileSize error
- Same as above, except with filesize

ERROR: Disk File Inconsistency. Hit ENTER
- This will usually mean that the user has
swapped disks just after telling NARC to
View,Print or Extract and NARC does not
recognize the file, or it may also be
displayed as the result of reading/extracting
a damaged ZIP file.

ERROR: Incompatible Crunch Format
- Says that either the stowage code for this
file is not supported by NARC -OR- there is an
error in the ZIP header

ERROR: Extraction Failed due to Lack of Disk Space
- Abort will stop tagged extraction, continue
will try to fit next file.

Squeezed File Has a Diseased Decode Tree.
- When unsqueezing a file, NARC has found a
bad entry in the decode table.

ERROR: No directory space on destination!
- Self explanatory

Bad Path Name, Hit ENTER
- The destination path that the user entered
for extraction is not a valid DOS pathname,
re-enter.



NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 17

Requires DOS version 2.0 or above.
- NARC requires DOS 2.0 or above to operate.

Invalid archive file format
- NARC could not find any ZIP headers, this is
probably not a ZIP or ARC file.

Warning: Bad archive file header, bytes skipped = xxxxx
- If an entry has a bad header, NARC will
examine the next 64k bytes looking for a good
header. This is to maintain compatibility
with self-unpacking ARC files.

Unexpected end of archive file
- Says that NARC couldn't find the last ZIP
header

No matching file(s) in archive
- ZIP file is empty


Cannot create work file, enter drive for temporary file
-there was not enough directory space or disk
space to create NARC.TMP for viewing. Enter a
drive letter where NARC can create the
necessary work file. The file will be deleted
when the view has ended.































NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 18

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
ARCHIVE FILE FORMATS AND GENERAL INFORMATION
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ


For Those With a Little More Curiosity...


The following are the currently supported stowage methods.

1 unpacked (obsolete)
2 unpacked
3 packed
4 squeezed (after packing)
5 crunched (obsolete)
6 crunched (after packing) (obsolete)
7 crunched (after packing, using faster hash algorithm)
8 crunched (after packing, using dynamic LZW variations)
9 Squashed c/o Phil Katz (no packing) (var. on crunching)

NOTE: LZW is Lempel-Ziv-Welch compression algorithm

A little about the stowage methods.

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
Packing (or run length encoding)
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

This is the simplest of the storage methods. Suppose
that you have a line of text and at the end of the
line, you have 40 spaces. These 40 spaces are compressed
into 3 bytes in the ZIP file by the "packer". The first
byte is the actual character to be expanded (in our case
a space). The second byte is a special "flag" byte that
indicates that we need to expand these bytes when we
extract the file. The third byte is the count byte (in
our case it would be 40). So you can see that any time
the compressor finds repeated bytes like this, it can
compress them into 3 bytes. The anomalous case to watch
out for here is when the count byte is the same
character as the "flag" byte, this proved to be a
difficult roach to kill !
















NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 19

ÍÍÍÍÍÍÍÍÍ
IMPLODING
ÍÍÍÍÍÍÍÍÍ


The imploding algorithm scans the input file for string
matches. As the algorithm proceeds through the input
file, it looks "backward" toward the start of the file
to see if any character strings can be found to match
the string it is currently examining. If a match is
found, then the LENGTH of the match, as well as the
DISTANCE "backwards" in the file can be sent to the
output file, rather than the string itself. For example,
if our input file looked like this;


"This is the input file. This is the last sentence."


We note there are 50 characters in our input file. As
the algorithm scans this file, it notices that the
phrase "This is the " is repeated, so we need not send
this string to the output file, but instead we send the
LENGTH and DISTANCE values. When the algorithm is
looking at the second "This is the " string, it finds a
match at 24 characters backwards in the file (e.g. the
first string), this is the DISTANCE value; and the match
LENGTH is 12. The output file would now look like this;


"This is the input file. LENGTH,DISTANCE "last sentence."


The imploding algorithm, however, takes this "sliding
dictionary" output (LENGTH,DISTANCE pairs) and further
encodes them using a Shannon-Fano tree. Just as the
Huffman algorithm translates one character set into
another based upon frequency of occurrence, so the
imploding algorithm does with the LENGTH,DISTANCE pairs.
A table of values is built determined by the frequency
of occurrence of the LENGTH,DISTANCE pairs, that is the
number of times each LENGTH,DISTANCE pair is used in the
encoding process. The table is then sorted by frequency
of occurrence and unique variable bit length codes are
assigned to the table entries. The net result is that
the more frequently used pairs will be assigned the
shortest codes (measured in bits). There are up to three
tables used in the encoding process, one for LENGTH, one
for DISTANCE and one for literal characters. These
tables are stored in the zipfile in BCD compressed
format. For a complete description of Shannon-Fano
coding, see "DATA COMPRESSION", Held, Wiley Pub.






NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 20

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
HUFFMAN CODING (SQUEEZING)
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ



It does, at first, seem that making a file smaller would
be an impossible task. I will make an attempt here to
shed a little light on this subject since that is a
question that I hear pretty frequently and it is not a
two minute discussion question. It does require some
thought.

To compress a file with the Huffman algorithm, commonly
called squeezing, the first thing that must be done is
to read the file completely and count the occurrences of
each character. That is you count the "A" 's and the "B"
's and so forth. There are 256 characters in the
extended ASCII character set, of which approximately 90
are "printable", that is you can see them on the screen.
The IBM set has more "printables", but that is of no
consequence, since the squeezer deals only with the
numbers and doesn't care whether or not the file is an
ASCII text file or an EXE file. Once the squeezer has
counted the occurrences of each character, thus the
frequency of occurrence, it scans the table for the
characters that appear the least number of times and
forms an imaginary link between them, called a node.
Somewhere else in the tree, we will later develop a
pointer that points to this node. When you start putting
all of these things together, you will form a binary
tree in memory. Confused enough ? Let us try an example.

We have a file that is 100 bytes long and has 6
different characters in it. We have counted the
occurrence of each of the characters and found the
following.


quantity character

5 - D
10 - A
10 - F
20 - B
25 - E
30 - C

The spelling in the file wasn't very good, but we don't care.
Now we take these numbers and will call them the
frequency of each character. We then arrange the table
as below. This is an arbitrary arrangement, but it is
useful here so as to make our tree readable on the
screen. The arrangement makes no difference.




NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 21

Frequency 20 10 5 10 30 25

Character B A D F C E


We then examine the table to find the two characters
with the smallest frequency of occurrence. In our case,
it is obvious that one of them is 5,but which 10 do we
choose. As it turns out, it doesn't matter which one you
choose, we will arbitrarily choose the F. We draw lines
from the D and the F to form our node (the box below).


Frequency 30 10 5 10 20 25

Character C A D F B E
\ /
\ /
ÉÍÍ»
º15º = 5 + 10
ÈÍͼ


The number in the box is the sum of the frequencies of
the D and F characters. Now we again look for the lowest
two frequencies, except, this time we do not consider
the D and F characters individually, we instead consider
the node. The lowest two now are the A and the node,
that is 10 and 15. We again do some artwork.


Frequency 30 10 5 10 20 25

Character C A D F B E
\ \ /
\ \ /
\ ÉÍÍ»
\ º15º = 5 + 10
\ ÈÍͼ
\ /
\ /
ÉÍÍ»
º25º = 10 + 15
ÈÍͼ


We look at the table again for the next two lowest
frequencies and now find B and E. We continue in this
fashion until the entire "tree" is built, that is until
it all condenses to one node. The leaves are the actual
characters at the top of the tree and the nodes
represent branch joints with the root at the bottom.






NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 22

Frequency 30 10 5 10 20 25

Character C A D F B E
\ \ \ / \ /
\ \ \ / \ /
\ \ ÉÍÍ» ÉÍÍ»
\ \ º15º º45º
\ \ ÈÍͼ ÈÍͼ
\ \ / /
\ \ / /
\ ÉÍÍ» /
\ º25º /
\ ÈÍͼ /
\ / /
\ / /
ÉÍÍ» /
º55º /
ÈÍͼ /
\ /
\ /
ÉÍÍÍÍ»
ºROOTº
ÈÍÍÍͼ


Now that our tree is made up, we can encode the file. We
start at the root (always). To encode the first
character (leaf) of the tree (the letter C), we trace up
the tree until we hit the letter C at the top. Along our
journey, if we make a left turn, we record a 0 bit, and
a 1 for a right turn. So for the C, we would go left to
55 (and record a 0) and then left again to the letter C
(and record another 0),so the Huffman code for our
letter C is 00. For A we go left to 55, right to 25 and
left to A and it becomes 010. By doing all of the
letters this way, we find the following.


C = 00 ( 2 bits )
A = 010 ( 3 bits )
D = 0110 ( 4 bits )
F = 0111 ( 4 bits )
B = 10 ( 2 bits )
E = 11 ( 2 bits )

Mind that the zeroes and ones above are bits and not
bytes. Each character was represented in the original
file by 8 bits (one byte) and since we have reduced the
number of bits needed to represent each character, we
therefore reduce the size of the file. The savings add
up as follows,







NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 23

character frequency original bits squeezed bits savings

C 30 30 x 8 = 240 30 x 2 = 60 180
A 10 10 x 8 = 80 10 x 3 = 30 50
D 5 5 x 8 = 40 5 x 4 = 20 20
F 10 10 x 8 = 80 10 x 4 = 40 40
B 20 20 x 8 = 160 20 x 2 = 40 120
E 25 25 x 8 = 200 25 x 2 = 50 150
ÍÍÍÍÍÍÍÍÍÍ ÍÍÍÍÍÍ ÍÍÍÍÍ ÍÍÍÍÍ
Totals 100 800 240 560
³ ³
original file size ÄÄÄÄÄÄÙ ³
squeezed file size ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÙ


240 is 30% of 800, so we have compressed this file by
70%. Golly Wally, that seems pretty good. The rub lies
in the fact that in order to reconstruct the original
file, we must have access to the decode tree and since
each tree will be different for each file, we must
therefore save the tree with the file. It turns out that
the tree can have only 256 nodes in a bytewise
compression technique and each node will hold 4 bytes as
pointers,a full table will be about 1k long. The table
in our example has 5 nodes plus the 6 leaf nodes (where
our characters are), totaling 11. 4 times 11 is 44 and
if we add a few bytes for storing the node count and
some other statistics, our table is about 50 bytes long.
If we look at the 240 in the above table this gives the
total number of bits that it will take to encode the
file, divide 240 by 8 to get the number of bytes (30)
and add it to our 50, we get a compressed file size of
80 bytes. Since our original file was 100 bytes, we have
achieved a 20% reduction in file size. Not bad. What we
have really accomplished is a translation of character
sets, with our new set requiring less space than the
original ASCII set.

How far can we go ?

If we look at the maximums that we can obtain for the
different bit combinations in a optimally skewed tree, that
is a tree that is not exactly symmetrical, we find that we
can have only 4 - 2 bit codes, 8 - 3 bit codes, 16 - 4 bit
codes, 32 - 5 bit codes, 64 - 6 bit codes, 128 - 7 bit codes,
the remaining 4 will be 8 bit codes.


2 - 1 bit codes
4 - 2 bit codes
8 - 3 bit codes
16 - 4 bit codes
32 - 5 bit codes
64 - 6 bit codes
128 - 7 bit codes
--------
254

NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 24

And since we have a total of 256 different bytes to
encode, the remaining 2 characters must have 8 bit
codes. If we add the number of bits that this
represents,we find a total of 1554 bits or 195 bytes. So
at maximum, we have compressed the 256 bytes to 195 or
33%, thus the idealistic maximum that can be achieved
with the Huffman algorithm is 33% when using a byte
level implementation.

One final note; The Huffman scheme requires the input
file to be read twice, once to count characters and
frequencies and then again to do the actual encoding.
The major differences in Huffman coding and crunching
lie in the fact that crunching is a one pass operation
and does not require the table to be stored with the
file. Both, however, are extremely vulnerable to errors,
for example, imagine what would happen if you skipped
one bit when squeezing the file, all of the remaining
characters in the file would become the proverbial
garbage, since we are looking at the file on a bit
level.


NARC uses the method described in K. & R. pp. 130 for
setting up the binary tree with several modifications.
The simple binary tree is acceptable for this, since the
tree never grows and therefore will never become
unbalanced.

If you followed that, now go back and read the second
paragraph again, maybe it will make sense this time.

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
CRUNCHING/SQUASHING/SHRINKING
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ


Crunching began with an article by J. Ziv and A. Lempel
in IEEE Trans. Information Theory, May 1977, where the
method was originally described. Terry A. Welch wrote a
definitive application article in IEEE Computer, June
1984 which described in detail how to apply the
algorithm and some common problems encountered. Thus
the name LZW compression.

Crunching takes the Huffman coding method a step further
as it does not include a table with the crunched file.
The crunching algorithm also "learns" as it proceeds
through the file. If it finds repeated strings in the
file, they will be encoded into a table. This table is
set up (in NARC's implementation) as a 4096 by 3 table.
Each entry is formatted as ,, where PREF
is a 2 byte pointer to another entry. SUFFIX is the byte
for this entry. Representing the PREF's as pointers
rather than values speeds up most operations in NARC.
This idea came from Bob Freed and is very trick.


NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 25

One obvious benefit of crunched files is the fact that
there is no need to include the encoding table in the
compressed file as was the case with squeezing. Another
great benefit is the fact that crunching is a one
pass operation as opposed to the two pass situation in
squeezed files.

Crunching begins by creating an "atomic" table, that is
a table in RAM that contains 256 entries, one for each
character in the extended ASCII set. The values range
sequentially from 0 to 255. The table entries are
arranged as follows.

Prefix Pointer (2 bytes) and Suffix byte (1 byte)

In this initial table setup, the Suffix bytes are the 0
through 255 mentioned above. These are the "atomic"
characters, in that they must be in the table before
crunching or uncrunching can begin, since all files
contain some portion of this character set. We do not
know which characters will be included in any given file
and which ones will be excluded,so we must include them
all in our initial table. Once this table is set up, we
can begin crunching.

The Prefix pointer will contain a value that is a
pointer to another string. When the table is initially
set up, there are no other strings, so this Prefix
pointer is set to a special "Null" string, that is it
points nowhere. We must be careful when crunching the
file, to look for this special pointer and act
accordingly when we encounter it.

This Prefix and Suffix business is used to "build" long
strings. If we read the input file and found that the
first character was the letter "I", we would look this
letter up in the string table by hashing (computing an
address). If we found the letter in the table (which we
certainly will on the first character), then we output
it's "hashed" address as a code to the output file (the
crunched file). Suppose then, that the next character
input from the file was the letter "D", the cruncher
would then look at the I and the D together to see if
they exist as a string in the table. Well of course,
since this is the second character of the file, we know
that it doesn't, so the cruncher forms a new entry in
the string table. This entry has as its' Prefix
pointer, a value that "points" to the letter "I" entry
in the table, that we made a minute ago. The suffix byte
in this case would be the letter "D". Now another code
is output to the crunched file, representing the letter
"D". Well this is great, we have now turned 2 bytes
from the input file (16 bits) into 3 bytes in the output
file (24 bits). You are correct, crunching begins by
"not crunching" , but it is a crazy world ! The real
value becomes apparent when we run into this same
sequence "ID" in the input file again. Now we will find

NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 26

an entry for it in the string table and we can output a
single 12 bit code that stands for "ID", thus saving 4
bits. The cruncher continues "learning" strings like
this until the file is crunched. It should be noted that
the string table is dynamically changing as the input
file is processed.

The early versions of crunching implemented, stopped
"learning" once the string table was full. This gave a
very poor compression ratio in some files. Versions 8
and 9 have an additional feature called adaptive reset,
where the string table is cleared and crunching begins
all over again ! This capability really helps the larger
files more than smaller files.




Details of Storage Versions


NARC supports all of the current "skrunching"
algorithms. A brief description of each follows.

Version 1

- "STORED" File is simply stored (obsolete now, 25 byte
header)

NOTE: Files stored with this version are rare.

Version 2

- "STORED" Current version of simple storage. This
version was implemented with the implementation of file
compression. The main difference in version 1 and 2 is
the ZIP header (see header section below), version 1
has a header length 4 bytes smaller than any of the rest
of the storage methods since in version 1 there was no
need to store the original file length separately from
the stored file length since they were the same.

Version 3

- "PACKED" Repeated bytes are packed into a three byte
string (see Packing above)

Version 4

- "SQUEEZED" after packing. The file is first packed as
described above and then squeezed

Version 5

- "CRUNCHED" This is the first implementation of LZW
released. Uses fixed length codes and a complex hashing
function. (obsolete now) (See hashing below)

NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 27


NOTE: Files compressed with this version are hard to
find. Version was released only one month when next
version came out.


Version 6

- "CRUNCHED" after packing. The file is first packed and
then crunched. Uses fixed length codes and the same
hashing function as version 5.

Version 7

- "CRUNCHED" after packing. Same as version 6 except a
faster hashing function is used.


Version 8

- "CRUNCHED" after packing. Uses variable length codes
in the crunched file (9 to 12 bits) and has a faster
hash function (actually not hashing at all, but for the
sake of consistency, we will call it that). This version
also resets the string table when it becomes full which
benefits the compression ratio of larger files. This
resetting is commonly called an adaptive reset.


Version 9

- "SQUASHED" (variation on crunching scheme). This
version uses the same hashing function as version 8 but
varies the crunching codes from 9 to 13 bits. There is
also no packing, which affords the string table the
opportunity to "learn" longer codes and thus improve the
compression ratio (benefits "real large" files).





















NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 28


ARC file header structure is same for both DOS and CP/M

Byte number Value(s) Meaning
ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ
1 1Ah Header Flag
2 0-9 Compression Version
3-15 --- ASCIIZ compressed filename
16-19 --- Compressed file size in bytes
Low Word, High Word with each word
in LoHi format
20-21 bits DOS date format
15-9 Year
8-5 Month
4-0 Day (all zeroes means no date)
22-23 bits DOS time format
15-11 Hours (military)
10-5 Minutes
4-0 Seconds
24-25 --- CRC-16 in LoHi format of uncompressed
file. ------- This is important.
26-29 --- Original uncompressed file size
NOTE: Version 1 files are not compressed
so the length can be found with
bytes 16-19, also the header len
for version 1 files is 25 bytes.

































NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 29

Hashing.....

Hashing is simply an arithmetic way of coming up with an
address in a table. You have a set of input numbers
(codes) that will map one-to-one with the output codes
in an ideal situation. That is, each time you input a
certain number, you can rest assured that the output
will always return the same output number. This is not
quite the case in the current versions of .ARC files.
The reason is that the algorithm would require a MEG or
so of RAM for implementation. The utilization of a
smaller string table in all of the ARC programs
introduces the possibility of producing the same output
number for 2 or more input numbers. This is called a
hash collision. This is handled separately in .ARC files
with what is called a "collision table", which helps to
locate the correct table entry.


There are three versions of "hashing" used in ARC files.

Hash1 - Key = upper 12 bits of lower 18 bits of unsigned square of
(prefix code + suffix byte) OR 800h

Used in stowage versions 5 and 6

Hash2 - Key = lower 12 bits of unsigned product of
(prefix code + suffix byte) * 15073

Used in stowage version 7

Hash3 - Key = next available address in table.

Used in stowage versions 8 and 9
























NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 30

CRC calculations - ARC files

NARC does not use the traditional table lookup method
for calculating the CRC of the file. The table approach
requires the table to be in RAM and takes up more space.
NARC calculates the CRC on the fly, which requires no
table and saves space. The algorithm is taken from the
definitive article by Aram Perez in IEEE Micro, June
'83. The polynomial is X^16 + X^15 + X^2 + X^1 which is
not compatible with the Xmodem CRC.


- ZIP files

The CRC calculated in ZIP files is a 32 bit CRC value,
which offers much greater error detection capabilties.
NARC builds a table and uses a table lookup method of
calculating the CRC values.

Note that when NARC displays the CRC-32 used in ZIP
files, only the lower two bytes are displayed, but the
full 32 bits are used in all crc calculations.




































End of NARC.DOC Copyright (c) 1987-90 Infinity Design Concepts, Inc.

Page 31