arc521.doc

汇编源代码大全

archive, since the files are not analyzed or
compressed, but merely stored straight.  The intent is
to allow users to build an archive "in pieces", and
then compress all of the entries at once with the
Convert command.  The conversion is much faster if you
take advantage of RAMdisk support.

A minor bug was detected in our handling of date/time
stamps which occasionally resulted in stamping an
archive with the wrong date and time.  This has been
corrected.





ARC                                             Page 27
CHANGES IN VERSION 5.0

The Move command used to delete the files as it went.
It now waits until it is finished updating the
archive, and deletes them all at once.  (You *did* know
that Move is just an Add where the file gets deleted,
didn't you?)  This, along with the changes made in
version 4.5, means that it is now much safer to
interrupt ARC while it is working.

The Print command no longer prints the name of each
file.  Instead, it prints a formfeed after each file.

The Run command now supports BASICA programs.  Also,
the filename extension is no longer required on the
Run command.

The Garble option was added.  It provides a convenient
means of low level data security for the casual user.
Use it sparingly, if at all.

ARC no longer tests for the presence of $ARCTEMP.CRN
before creating a new one.  If you interrupt ARC a
lot, you'll find this much more convenient.  If you
happen to have a file named $ARCTEMP.CRN which you
want to keep, too bad.

Improved error recovery was added when reading an
archive.  ARC now has a good chance of recovering the
data from a corrupted archive (the corrupted entry is
still lost, of course).

Path support has been added for all commands, though
it doesn't do anything on most of them.  For example,
there isn't much we can do with a path in the List
command.  But many users will be glad to know that a
path can be used when extracting a file, and specifies
where the file is to be placed.

Support for the TEMP environment string was added.  If
ARC doesn't find an environment string named ARCTEMP,
then it looks for one named TEMP to use instead.
Several packages already use the TEMP string for
exactly this purpose.  With any luck, maybe we can get
a standard going.









ARC                                             Page 28
ARC is now using a different variation of Lempel-Zev
coding, courtesy of Kent Williams, who found it on
USENET and adapted it to the IBM PC.  The new method
differs from the old in several respects.  The most
significant differences are:

1) Where our previous implementation used a fixed code
   size of twelve bits, the new one starts with a code
   size of nine bits and increases it as it needs to.

2) The earlier method tended to "choke" on large files
   when its string table filled up.  The new method
   has a rather ingenious scheme its authors call
   "adaptive reset."  When it notices that its string
   table has filled, and its compression ratio is
   starting to suffer, it clears the table and starts
   defining new strings.

Our benchmarks show an improvement in compression on
the order of 10% when crunching is used.
Additionally, ARC 5.0 is on the order of 23% faster at
adding a file when crunching is used, or 13% faster
when squeezing is used.  Extracting a file crunched
with the new method is 27% faster than it is with the
old method.  Extraction of any other type of file
(including those crunched with the older method) is no
faster than before.  These figures are based on our
own benchmark tests; your results may vary.

The previous implementation of Lempel-Zev coding is no
longer used to pack files.  The "V" (Verbose listing)
command distinguishes between the two by referring to
the older method as "crunched" (with a lower-case
"c"), and the newer method as "Crunched" (with a
capital "C").

ARC 5.0 can still read archives created by earlier
versions of ARC, but once again it creates archives
which older versions cannot read.



CHANGES IN VERSION 5.1

Rick Moore discovered that ARC was occasionally adding
an archive to itself.  This would only happen when the
archive is in the same directory as the files being
added, and its name comes last.  This bug has been
fixed, though it is still possible to fool ARC into
adding an archive to itself by getting tricky with
path names.



ARC                                             Page 29
Dana Montgomery found the upper limit on how many
files can be added to an archive.  There's always been
an upper limit, but it depends on memory, and used to
be larger than anyone could possibly want (knock on
wood).  However, the added memory requirements in
version 5.0 lowered this limit into the realm of
possibility, somewhere around 300 files.  We change
some things around, and effectively, there is no
longer a limit on how many files you can add at once.
ARC will add the files in batches of as many as it can
handle at one time.

We've also introduced a new packaging method for ARC
and its documentation, since we are in the unique
position of being unable to use ARC for this purpose.
We've created a program called ARC51.COM which, when
executed, attempts to create the ARC program and
manual.  You must be in the same directory as
ARC51.COM when you run it.



CHANGES IN VERSION 5.2

We've made some changes to Lempel-Zev coding to
further improve its compression ability on "poor fit"
files (mainly .EXE and .COM files).  The result is
that crunching now works better than squeezing on
almost all files, and comes close on most of the rest
(within 5%, by our benchmarks).  Accordingly, we have
dropped Huffman encoding in this version.  Our
benchmarks show a speed improvement of around 14% when
adding a file to an archive.

A low-level file copy routine was implemented to speed
up bulk data moves.  Deleting or adding a file to a
large existing archive is now much faster.

The Run command has been modified to allow passing
command line arguments to the program being run.

The temporary file introduced in version 4.3 has been
eliminated.  Crunched output generated in the analysis
phase is now placed directly in the output archive,
and is rewritten if crunching turns out not to be the
best choice.  This can, in rare circumstances, cause
an archive to be slightly larger than it should be,
but it also makes ARC considerably faster at adding
files to an archive.





ARC                                             Page 30
  PROGRAM HISTORY AND CREDITS



In its short life thus far, ARC has astounded us with
its popularity.  We first wrote it in March of 1985
because we wanted an archive utility that used a
distributive directory approach, since this has
certain advantages over the then popular central
directory approach.  We added automatic squeezing in
version 2 at the prompting of a friend.  In version
2.1 we added the code to test for the best compression
method.  In less than a year we found that our humble
little program had spread across the country, and
seems to have become a new institution.

We are thankful for the support and appreciation we
have received.  We hope that you find this program of
use.


If we have achieved greatness, it is because we have
stood upon the shoulders of giants.  Nothing is
created as a thing unto itself, and ARC is no
exception.  Therefore, we would like to give credit to
the following people, without whose efforts ARC could
not exist:


Brian W. Kernighan and P. J. Plauger, whose book
"Software Tools" provided many of the ideas behind the
distributive directory approach used by ARC.

Dick Greenlaw, who wrote the public domain SQ and USQ
programs, in which the Huffman squeezing algorithm was
first developed.

Robert J. Beilstein, who adapted SQ and USQ to
Computer Innovations C86 (the language we use), thus
providing us with important parts of our squeezing
logic.

Kent Williams, who graciously allowed us to use his
LZWCOM and LZWUNC programs as a basis for our Lempel-
Zev compression logic, and who continues to make
valuable contributions.

David Schwaderer, whose article in the April 1985
issue of PC Tech Journal provided us with the logic
for calculating the CRC 16 bit polynomial.




ARC                                             Page 31
Terry A. Welch, whose article "A Technique for High
Performance Data Compression", IEEE Computer Vol 17 No
6 (June 1984) seems to have started all the research
on Lempel-Zev coding.

Spencer W. Thomas, Jim McKie, Steve Davies, Ken
Turkowski, James A.  Woods, and Joe Orost, who are the
authors of the UNIX compress utility.

Alex Jacobs, who in June of 1985 sent us the very
first shareware contribution we ever received.


And many, many others whom we could not identify.








































ARC                                             Page 32
        BULLETIN BOARDS



ARC is distributed mainly through shareware channels.
Among other things, this means that ARC is available
from many bulletin board systems.  In fact, the system
operators (sysops) of many bulletin boards have taken
to storing almost all of their downloadable files in
archives to save themselves disk space and to save
their users time.

This also makes things more convenient for the
bulletin board users, since one archive may contain
several programs, any related data files, and the
documentation.  Many shareware authors have taken to
distributing their software in archives to help ensure
that the users receive everything.

Obviously, we can't do that with ARC.  As a result,
many of our users have ARC, but don't have the manual.
Most of our customer support calls come from people
who have never seen the manual, and in many cases
didn't even know that one exists!

To solve this problem we developed what is in essence
a self-unpacking archive.  We distribute this as
"ARCxxx.COM", where "xxx" is the current version
number.  For example, ARC version 5.20 would be
distributed as "ARC520.COM".  This program, when run,
unpacks itself into a copy of ARC plus its
documentation.

Now that we've gone to all that work, we'd really
appreciate it if you would use this program when you
distribute ARC.


















ARC                                             Page 33
         SITE LICENSES



Corporate users may wish to obtain a site license for
the use of ARC.  Please use the order form in this
manual to order a site license.  Site licenses are
granted as of when we receive your payment.  License
fees vary depending on the number of computers on
which ARC will be used, as follows:

     1 to  9 machines  $35 each
    10 to 24 machines  $25 each
    25 to 49 machines  $20 each
    50 to 99 machines  $15 each
    over  99 machines  $1500 one time fee



Enclosed is a site license agreement, which should be
signed and sent with your payment when ordering a
commercial site license.


A commercial site license does not include additional
copies of the ARC program disk and the ARC manual.
Instead, you make your own copies of the disk and
manual as you need them.  If you wish, you can order
additional program disks and manuals from us.

Additional program disks cost $10 each.  Additional
manuals cost $20 each.  If you wish, you can order
program sets for $40 each.  A "program set" is a
program disk and manual enclosed in a vinyl binder.
There is a $50 minimum on all orders.



















ARC                                             Page 34
             SITE LICENSE APPLICATION



The use of ARC in a commercial environment or government organization
is granted under the following terms:

1.  Payment of the license fee must be made to System Enhancement
    Associates, Inc.  The fee is based on the number of computers
    which will be used to run ARC, as follows:

          1 to  9 machines  $35 each
         10 to 24 machines  $25 each
         25 to 49 machines  $20 each
         50 to 99 machines  $15 each
         over  99 machines  $1500 one time fee

2.  You may use ARC on the number of computers included in the license
    fee.  If you have paid the fee for over 99 machines, then you may
    use ARC on any number of computers within your organization.

3.  You may make copies of the program and documentation, in their
    original,  unmodified form,  without restriction.  You may
    distribute these copies without restriction.

4.  If these copies are distributed outside of your organization,  you
    have no obligation to control the use of those copies which are
    outside of your organization.

5.  You may make copies of the program documentation, in both its
    printed form and machine readable form, without restriction.

6.  You may use all future versions of ARC under this license.

7.  You may *not* modify the program or charge a fee for copying or
    distributing the program or its documentation.

8.  It is your responsibility to make the necessary copies and to
    deliver them to the computers which they will be used on.  If you
    wish, you can order additional ARC program disks and manuals.

9.  We are not responsible for *anything* that may happen or not happen
    if you use ARC.  You're on your own.


    I agree to abide by the terms and conditions of this license.

    _____________________________         __________________________
    Signature                             Date

    _____________________________
    Name (please print or type)

    _____________________________
    Title

    _____________________________
    Company


                   ORDER FORM



Check which items you wish to purchase:

    [] Noncommercial license for the use of ARC.

    [] Commercial license for the use of ARC on ___ computers (see
       attached price schedule and terms).

    [] Program disk and documentation (only on orders of $50 or more,
       not counting maintenance contracts).

    [] ____ additional program disks at $10 per disk.

    [] ____ additional manuals at $20 per manual

    [] ____ additional program sets (disk, manual, and binder) at $40
       per program set.

    [] One year maintenance agreement at $50 per year.

    [] Payment of $_____ is enclosed (check or money order).

    [] Please charge $_____ to my [] Visa or [] MasterCard:

      Card number:  _______________________________

      Expiration date:  ___________________________

      Signature:  _________________________________

    _______________________________________________
    Name

    _______________________________________________


    _______________________________________________
    Address

    ______________________   ________  ____________
    City                     State     Zip

    _______/_______
    FidoNet address


Send this completed form to:

                  System Enhancement Associates, Inc.
                     21 New Street, Wayne NJ 07470

For program disk orders outside the U.S., please add an additional $5,
and enclose an international money order payable in U.S. currency.

For commercial site license orders, please enclose a signed copy of
the site license agreement.