The Fred Fish Collection 1.5/ffcollection-1-5-1992-11.iso/ff_progs/diskutil/superdup.lzh/SuperDuper.tex

Text File | 1992-01-05 | 55.7 KB | 1,095 lines

%
% This plain TeX document is a typeset version of SuperDuper.guide
% that was kindly prepared by Reinhard Spisser.
% Use AmigaTeX or another (bleeech!) TeX system for compiling it.
%
\font\twelverm=cmr12
\font\twelvett=cmtt12
\twelverm
\font\it=cmti12
\font\bf=cmbx12
\font\sf=cmss12
\font\big=cmbx10 scaled\magstep5
\font\bigss=cmssdc10 scaled\magstep5
\font\twelvess=cmssdc10 scaled\magstephalf
\def\SuperDuper{{\twelvess SuperDuper}}
\def\ToggleClick{{\twelvess ToggleClick}}
\def\SetTDDelay{{\twelvess SetTDDelay}}
\def\SDBootInstall{{\twelvess SDBootInstall}}
\centerline{\bigss SuperDuper}
\vskip 0.2cm
\centerline{\sf Version 2.01}
\vskip 1cm
\centerline{\big A Universal Copier}
\vskip 1cm
\parindent=40pt
\parskip=10pt plus 1pt minus 1pt
\baselineskip=15pt



                          \centerline{\copyright\ 1991 Sebastiano Vigna}

\vskip 2.5cm
\SuperDuper\ is a disk copier/formatter that tries to be to disk handling what
{\twelvess Mostra} is to IFF displaying: a fast, compact, system-friendly tool which
combines speed, features, and some bells and whistles to make your life
easier.

By ``fast" I mean exactly what you're hoping---blazingly fast. A disk is
usually copied and verified in less than 100s. Without verify, the time
drops to 69s. You can buffer a disk in RAM in less than 36s, and then making
a verified copy takes 67s, while a non-verified copy takes less than 36s.
Adding another destination drive increases verified copy times by 34s, but
hardly changes non-verified copy times (the Amiga can write more than one
drive at a time; I just need a few tenths of a second in order to measure
the drive speed and step the heads). Thus, if you really trust your drives
and your media you can make four copies in 38s. These timings can vary with
the system configuration, the multitasking overhead, the disposition of the
blocks on the surface of the disk, the state of the {\twelvett DATE} option (which
requires a separate write on the root block track for each disk) and the DMA
access of the custom chips (previous users of \SuperDuper\ might think this
release is slower than the previous one: it is really faster, but \SuperDuper\
1.0 was a little bit optimistic about its copy times---the motor on/off
delays were erroneously skipped).


\vskip 2cm
\centerline{\big Changes}
\vskip 1.5cm

\SuperDuper\ 2.01 has almost no visible changes with respect to \SuperDuper\
2.0, apart from the extension of the ARexx macros, which is now {\twelvett
supdup}
instead of {\twelvett sd} in order to avoid conflicts with other programs.

However, it was discovered that many flakey drives have power supply
problems when four of them are connected to an Amiga. Sometimes the head of
a drive won't step, and this error cannot be caught even by verifying, since
the drive doesn't know where the head is---its position has to be tracked
via software.

In order to prevent this annoying phenomenon, which was the only known
source of bad copies, the head moving strategy was slightly changed.
\SuperDuper\ no longer steps multiple heads at the same time. This marginally
increases (about 3 tenths of second for each destination) the non-verified
copy times, but gives you a 100\% reliability even on out-of-specs systems.

If something wierd happens in spite of this patch, it can be tracked at
recalibration time. When a mismatch between \SuperDuper 's internals and the
drive signals is detected, a requester {\twelvett Error while recalibrating} is
issued. In this case, you can try to slow down the head moves using the
\SetTDDelay\ utility which is supplied with \SuperDuper .

The only other noticeable enhancement is the fact that now \SuperDuper\ checks
the {\twelvett NOCLICK} flag separately for each drive. If you have some drives which
support {\twelvett NOCLICK} and somes which don't, \SuperDuper\ will click only the
allowed drives. Previously, the information in the public unit of the drive
0 was used for all the drives. The utility \ToggleClick\ which is supplied
with \SuperDuper\ allows to hush selectively any unit. Moreover, a new {\twelvett NOCLICK}
ARexx command allows to force no-clicking selectively even under 1.2/1.3.

It should be remarked that \SuperDuper\ is much less tolerant than
trackdisk.device. Bad drives can work (almost) perfectly with
trackdisk.device, because of its many, frequent cross checks. For instance,
at each disk insertion some track is read by AmigaDOS, and if the track number
doesn't match with trackdisk's internals, a recalibration is started.
\SuperDuper\ instead doesn't read anything before copying (for speed reasons);
thus, if your drive has a flakey DSKTRACK0 signal \SuperDuper\ could believe
it's on track 0 while it isn't.

\vskip 2cm
\centerline{\big Overview}
\vskip 1.5cm
Main features:

\item{$\bullet$}{\SuperDuper\ copies, formats and checks from/to any combination of Amiga
drive(s).}

\item{$\bullet$}{\SuperDuper\ has a switchable 880K RAM buffer that allows for any number of
duplications while reading the source disk only once. The combination of
destination drives can be changed at each pass. If you have a hard disk, you
can create on it an image file that will act as a buffer. This file can be
saved and reused many times. Also, all kinds of virtual disks are supported
for buffering (VD0:, RAD:, FMS:,$\ldots$). Morever, a count is kept of the copies
generated by a buffered disk.}

\item{$\bullet$}{\SuperDuper\ checksums the RAM buffer. If some badly written program is
trashing your memory, you are alerted. Thus, buffered copies are as safe as
direct copies.}

\item{$\bullet$}{\SuperDuper\ also checks its internal DMA buffers at each write.}

\item{$\bullet$}{\SuperDuper\ can allocate a buffer of less than 880K. In this case, it will
use real-time compression in order to do multiple pass copies with maximum
efficiency. Most disks can be wholly buffered on a 1MB machine. You can also
make multiple copies with multiple passes. Copy times are (almost)
unaffected.}

\item{$\bullet$}{\SuperDuper\ will automatically retry tracks which produce a verify error.
The number of retries is programmable. A simple visual clue is given to the
position of the error, but on request detailed error information printing is
available.}

\item{$\bullet$}{\SuperDuper\ is highly system-friendly---the use of CPU time is negligible,
so you can multitask efficiently.}

\item{$\bullet$}{\SuperDuper\ has the option of incrementing the creation date of the copy
so AmigaDOS doesn't get confused. If, however, the option is switched on and
the disk is not an AmigaDOS disk, \SuperDuper\ won't increment the date.}

\item{$\bullet$}{\SuperDuper\ is faster than diskcopy---actually it pushes the drives to
their limits. At the time of this writing, \SuperDuper\ is the fastest Amiga
copier both from a ``pure" (physical time) and from a ``per-copy" (real time
for each copy when a big number of copies of the same disk is produced)
point of view.}

\item{$\bullet$}{\SuperDuper\ alerts the user with sound (and optionally voice) about the
operations in progress---so you can really be doing something else!}

\item{$\bullet$}{\SuperDuper\ can format both OFS and FFS disks.}

\item{$\bullet$}{\SuperDuper\ displays a list of the last few disks copied. If you do a lot
of copying, you'll find this feature more than a little useful.}

\item{$\bullet$}{\SuperDuper\ can manage the Amiga drives without help from the
trackdisk.device. Through the supplied utility \SDBootInstall, you can create
a boot disk which will keep the system away from your drives, giving you
back more than 30K per unit. This is very useful when doing intensive
buffered copying on a 1M machine.}

\item{$\bullet$}{\SuperDuper\ can automatically start any copy or format operation by
monitoring the disks' extraction and insertion.}

\item{$\bullet$}{Almost all elements of the 3-D, 2.0-like graphical user interface have
keyboard equivalents. When possible, 2.04 features like {\twelvett TAB} gadget cycling
and window zooming were supported.}

\item{$\bullet$}{The start/end cylinder of a copy is programmable.}

\item{$\bullet$}{Unique numbered names can be automatically generated while formatting.}

\item{$\bullet$}{\SuperDuper\ works under 1.2, 1.3 and 2.0.}

\item{$\bullet$}{\SuperDuper\ has a time indicator.}

\item{$\bullet$}{\SuperDuper\ has a beautiful name. {\twelvett 8\^{ })}}

\item{$\bullet$}{If this is not enough, an ARexx interface allows any kind of
customization. In particular, a startup ARexx script lets you set up a
custom configuration. Since \SuperDuper\ can turn off its graphical user
interface via a command line switch, it is possible to use \SuperDuper\ as a
CLI command by writing a suitable ARexx macro. A switch allows you to shut
down ARexx in order to gain memory. ARexx macros can be launched via a file
requester (asl.library, arp.library and req.library are supported).}

\vskip 2cm
\centerline{\big First Steps}
\vskip 1.5cm

To use \SuperDuper , you simply double-click on its icon. You will see five
rows of gadgets. The first one has at most one gadget selected: it's the
source. The second one lets you select the destination drive(s). The other
lines contain option and action gadgets.

Every gadget can be activated via mouse or keyboard (using the underlined
letter). The line of destination drives can be controled by pressing {\twelvett SHIFT}
together with the underlined number. You can use {\twelvett Q} or {\twelvett ESC} to exit, instead
of hitting the close gadget. Three of the string gadgets have underlined
letters which activate them. Moreover, if you're running under 2.04, you can
use {\twelvett TAB} or {\twelvett SHIFT-TAB} to pass from a string gadget to another one.

To make your first copy, if you have two (or more) drives simply select in
the first line the gadget for the drive which contains the source floppy,
and in the second line the gadget(s) for the drive(s) containing the
destination(s) (for the time being do not choose the same drive both as
source and as destination). Then hit the {\twelvett GO} gadget. After a while, the
display will flash, a beep will be generated, and the copy will be finished.
As each cylinder is copied, the elapsed time indicator is updated. Note that
a first beep will be generated when the copy is almost finished, so you have
time to prepare yourself.

If you have only one drive, select it both as source and as destination.
Then hit the {\twelvett BUFFER} gadget, thus creating a RAM buffer. Depending on the
memory available, it will be a full 880K buffer or a partial buffer. In the
latter case, real-time compression will let \SuperDuper\ get the best out of
it. Now put in the source disk and hit the {\twelvett READ} gadget: the buffer will be
filled with the contents of the disk. If the progress bar reaches its
maximum length, then the whole disk has been buffered. Pull out the source
disk, put in the destination, and hit the {\twelvett GO} gadget. The buffer will be
written to the disk. If only a part of the source disk was buffered, put it
in again, buffer it again (note that now the progress bar starts where it
stopped before) and write it again. This process must be repeated until the
whole disk has been copied. It is safer to set the write protect tab on the
source disk, in order to avoid the unpleasing side-effects of
source/destination mismatches.

\vskip 2cm
\centerline{\big The action gadgets}
\vskip 1.5cm
Four gadgets control \SuperDuper 's copy/format/check operations:

\item{$\bullet$}{{\twelvett STOP} stops any operation. If pressed while the multi-pass real-time
compression buffer is selected and no operation is in progress, it will
empty the buffer and reset the pass count, thus allowing you to buffer
another source even if the previous one wasn't finished (see the description
of the buffering system). If you {\twelvett STOP} immediately after starting a copy
operation and nothing has been drawn in the progress bar, nothing has been
written to the destinations.}

\item{$\bullet$}{{\twelvett READ} can be used only when a buffer is selected; it fills the buffer by
reading from the source drive.}

\item{$\bullet$}{{\twelvett GO} initiates a copy operation. If no buffer is selected, the source is
copied to the destination(s). If a buffer is selected, the content of the
buffer is written on the destination(s). If {\twelvett FORMAT} is selected, the
destination drive(s) are formatted.}

\item{$\bullet$}{{\twelvett CHECK} is basically a {\twelvett READ} without buffering. The source disk is scanned
for errors. No buffer is needed to use it. Note that \SuperDuper\ will detect
trackdisk.device related errors, but it won't find DOS checksum errors (for
this purpose, for instance, you can use FixDisk).}

When \SuperDuper\ starts an operation which involves reading a disk, i.e.,
{\twelvett READ}, {\twelvett CHECK} and non-buffered {\twelvett GO}, it scrolls
up the name list and marks the current drive as {\twelvett <UNKNOWN>}. This
happens because it can't know if the disk is a DOS disk before reading track 0.
After less than a second, the track will be read, and the name will be changed
to {\twelvett <NDOS>} if the disk is not a DOS disk. Otherwise, as soon as the
track 80 is read (the progress bar is in the middle) the name of the disk will
be displayed. However, if for any reason the name is incorrect (wrong format,
read error, etc.) \SuperDuper\ will name the disk {\twelvett <BAD NAME>}. In
this case, it is very likely that the root block is a little bit scrambled, so
it's probably a good idea to turn off the {\twelvett DATE} option gadget. Beware: if you
are using a multi-pass buffer, the name of the disk could be unavailable at the
first pass.

If \SuperDuper\ finds an error on read (or verify), it will retry reading
(writing) the track, each time incrementing the first number of the {\twelvett
Rtry:Err} indicator. If after a number of retries specified in the gadget
{\twelvett Retry\#} the error remains, \SuperDuper\ will increment the second
number (the error counter), restore the original retry counter and continue. A
little rectangle in the progress bar will point out where the error occured. It
will be positioned horizontally, proportionally to the track number, and
vertically, proportionally to the unit number (the first line of rectangles
shows errors on unit 0 and so on).

Note that while retrying \SuperDuper\ can't
be stopped: don't set the {\twelvett Retry\#} gadget to 99 unless you really know
that's what you want to do. At the end of the copy, the first number shows
how many retries leading to a successful write were done, while the second
one points out the number of tracks with an actual error. If you want to get
a very detailed error report, you can set {\twelvett PRINTERRORS ON} from ARexx. A
console window will appear, and every wrong read, write or retry will
generate a message explaining what doesn't work. Usually you will get bad
checksums, but if a track is really scrambled \SuperDuper\ could be unable to
get the first sector after a gap, in which case nothing at all is recovered.

The progress bar is drawn in a different color if you're doing an {\twelvett READ}, a
{\twelvett FORMAT} or a {\twelvett COPY} operation---so you can be sure you read the new chunk in
the buffer, and so you can avoid formatting your floppies when you think
you're copying something to them. The gadget corresponding to the action
currently executing will remain highlighted in order to remind you what
you're doing. Note also that the progress bar and the elapsed time indicator
are not updated if something locks the screen (like using menus). The update
is delayed until the screen is unlocked (thus \SuperDuper\ won't get stuck as
will almost all programs which do any rendering to their windows).

The volume of the beeps produced by \SuperDuper\ while copying can be set with
the ARexx {\twelvett VOLUME} command.

If you specify start/end cylinders different from 0/79 in the {\twelvett SC} and {\twelvett EC}
gadgets, only the part of the disk specified will be copied. The main use of
this option is for retrying some lazy disk (usually on the last tracks) if
you're not satisfied with the number of retries issued by \SuperDuper . Please
refer to the section on the buffering system for some subtle interactions
between the RAM/HD/VDisk buffer and the start/end cylinder selectors.

While doing buffered copies, at each successful copy (that is, without
errors) the {\twelvett Cpy\#} indicator will be incremented. Thus you can know precisely
how many disks you copied. Moreover, the counter will be incremented only if
the operation ended on the last track of the disk and started from the first
track of the buffer. This allows you to manually retry spare tracks by
changing the {\twelvett SC/EC} gadgets without getting spurious increments, and if a
multi-pass copy is in progress only the last pass will actually increment
the counter.


\vskip 2cm
\centerline{\big Copy Options}
\vskip 1.5cm

Five gadgets control the copy/format options.

\item{$\bullet$}{{\twelvett VERIFY} turns verify on and off (you can also format without verifying).
However, turning off verify is not recommended.}

\item{$\bullet$}{{\twelvett DATE} toggles on or off the change of the date of an AmigaDOS disk. This
change is necessary so AmigaDOS can distinguish otherwise identical disks;
if two truly identical disks are inserted in the drives, AmigaDOS gets
confused and crashes. However, if for some reason you want a ``physical"
copy, you would turn off this option. {\twelvett DATE} will be ignored for a
non-AmigaDOS disk.}

\item{$\bullet$}{{\twelvett FORMAT} enables formatting. When you hit {\twelvett GO}, all destination drives will
be formatted. To copy again, you must deselect {\twelvett FORMAT} by clicking it again.
If {\twelvett VERIFY} is selected, the format process is verified. Note that when you
hit {\twelvett READ}, {\twelvett FORMAT} is automatically deselected. This happens in order to avoid
the unpleasing error of thinking you're writing a buffered disk, while
actually formatting it.

The gadget prefixed by {\twelvett Label:} allows you to choose a name that \SuperDuper\
will use while formatting. The name must be chosen before clicking {\twelvett GO}---it
is disabled (ghosted) during the formatting.}

\item{$\bullet$}{{\twelvett INCNAME} makes easy to format a bunch of disks with different, unique
names. If this gadget is selected while formatting, \SuperDuper\ will scan the
Label string gadget searching for a numeric pattern (i.e., one or more
digits) and will increment the pattern value for each disk formatted.

In case more than one pattern is present, the last one is used. For instance,
if you format four disks with label {\twelvett Foobar.000}, the disks will be named
Foobar.001, Foobar.002,... and at the end of the copy the label gadget will
contain Foobar.004, thus being ready for the next formatting. The more
digits, the more unique names. Since you can start from any number, and
after 99...9 the numeration wraps around, if you need to start with 00...0
you can put in something like {\twelvett Foobar.999}: The first disk will be labeled
with Foobar.000.}

\item{$\bullet$}{{\twelvett FFS} enables the formatting of FFS disks; for
copying it is ignored.}

\vskip 3cm
\centerline{\big The Buffering System}
\vskip 1.5cm

Three gadgets control the full-featured buffering system of \SuperDuper .
Buffering is useful when you have to do a lot of copies: you read a disk
only once, and then you can make as many copies as you want without
rereading it. It also has other uses: if you have to create distribution
disks (for instance for a commercial package) you can create them using high
speed virtual floppies, such as Commodore's RAD: or Matt Dillon/Jim
Cooper's FMS: disk. \SuperDuper\ can then read from those virtual disks and
make many copies on floppies at high speed.

Since data integrity is a primary issue, \SuperDuper\ checksums the RAM
buffer. The possibility of writing a munged track is very low. Strict
control is also kept on the validity of the buffer---you can't write random
data on your disks inadvertantly.

\item{$\bullet$}{{\twelvett BUFFER} allocates a RAM buffer. \SuperDuper\ will try to get an 880K buffer:
if you don't have enough memory, a warning will be issued, showing the
number of buffers allocated (each buffer is 11K) and warning you that the
real-time compression system is activated. Beware of the fact that many
programs tend to crash under low-memory conditions, so if you have 1MB or
less you should close everything you can before hitting {\twelvett BUFFER}, and you
should possibly also use {\twelvett KILLSYS}.}

\item{}{The memory allocated will be used as a buffer to make multiple pass copies.
If \SuperDuper\ can find 880K, the process is very simple and uses very little
of the CPU, but if (for instance on a 1MB machine) it's impossible to buffer
a whole disk this way, \SuperDuper\ will use a real-time compression
algorithm. As the disk is read in the buffer, it is compressed in a special
format. The gain in size is usually 35\% for empty tracks, 20-30\% for text,
15-25\% for programs and 5-10\% for IFF ILBM images. Tracks which can't be
compressed are simply stored. The only disks which can't really be
compressed are disks filled with compressed files, like {\twelvett .lzh} or
{\twelvett .zoo} files, but for the others the size gain is enough to buffer a
whole disk on a 1MB machine. {\twelvett 8\^{ })}}

\item{}{Of course, the compression overhead eats a lot of CPU power. The algorithm
has been devised in such a way that compression and decompression are
absolutely real-time, i.e., you will notice no slowdown. However, beware of
the fact that while doing compression \SuperDuper\ always fully uses the CPU.
Even moving the mouse can slow down the operation in progress. Anyway, if
you have all of your memory allocated for the buffer, it is definitively not
a good idea to do anything besides waiting for the copy to finish.}

\item{}{A little side-effect of the allocation of all of the available RAM is that
some requester could be turned into an alert, or could even disappear
without waiting for the user to acknowledge it.}

\item{$\bullet$}{{\twelvett HDBUF} creates an 880K file in the current directory of \SuperDuper\ , and
uses this file as a buffer, exactly like {\twelvett BUFFER} does with RAM. Of course you
must use it only if you have a hard disk, and you started \SuperDuper\ from
it. The file contains the 1760 blocks which form a disk in their natural
order. The {\twelvett READ} operation will be a little slower, but if you have a good
hard disk you should be able to make copies as fast as with a RAM buffer.
The file is named {\twelvett SD\_Buffer}, and it's accessed only during the copy
operations. This means that you can read or write it using the CLI commands,
or the Workbench (but you will have to supply an icon). You can easily write
an ARexx macro which retrieves/stores binary images of a disk from/to
{\twelvett SD\_Buffer}. Then \SuperDuper\ will use the new contents when writing to
floppies.

If you put a file named {\twelvett SD\_Buffer} in \SuperDuper 's directory {\it before}
clicking {\twelvett HDBUF}, then \SuperDuper\ will assume this is a buffer file and will
use it. You can even write directly to floppies without reading anything.
Note that usually the buffer file is deleted when the {\twelvett HDBUF} gadget is
deselected, but if you supply a buffer file before activating the gadget
your file will be left untouched.}

\item{$\bullet$}{{\twelvett VDBUF} is probably \SuperDuper 's most esoteric
feature. By typing a device name in the string gadget named {\twelvett VDName},
you can select any device (\SuperDuper\ needs the Exec device name, e.g.,
{\twelvett ramdrive.device} for the RAD: AmigaDOS device). The unit number is
taken from the gadget with the label {\twelvett VDUnit\#}. The device you
specified will be used as a buffer for your disks. \SuperDuper\ expects the
device to behave like the trackdisk.device, namely it must be able to write
data at specific offsets. The main devices you can use, with their respective
names, are:

 \item{}{RAD: - the recoverable RAM drive. Configure it in your mountlist as a
floppy, and you can use it as a buffer (Exec name: ramdrive.device).}
 \item{}{FMS: - Matt Dillon/Jim Cooper's virtual floppy-on-hard disk
(Exec name: fmsdisk.device).}
 \item{}{VD0:, etc. - other recoverable, sector-oriented RAM drives.}}

The device you specify is checked on opening to see if it has enough space
to contain a full disk. The check is done simply on the number of sectors
available---if there are enough sectors, and they are arranged differently
than on a floppy, you will be able to use the device as a buffer, but don't
expect AmigaDOS to get anything meaningful from it.

{\bf WARNING:} many of these devices are buggy and return no error on
unsuccessful opening or failed size test. Some of them in this case will trash
your memory. Be sure that the device is configured properly---try an AmigaDOS
command on it first.

Of course, many people will find incredible ways to use this feature (for
techies: if you want try something weird, consider that \SuperDuper\ reads 512
bytes at offset 900608 on opening to test for size, and then reads 1760
chunks of 512 bytes, one for each sector, for every copy. The sectors are
read sequentially as they are distributed on the disk, so if the device
ignores the offset indication, you can feed it with 880K of a continuous
bytes stream. Buffering is another story though---the offset indication is
important because \SuperDuper\ places the blocks on the device ``in the right
place" as soon as it encounters them).

A {\twelvett BUFFER} is considered non-valid as soon as allocated, because it will
contain random info. To make it valid, you must read in a floppy. {\twelvett VDBUF} and
{\twelvett HDBUF} instead assume the buffer is always valid, because it could be
externally fed. This mechanism allows you to prepare, for instance, a
distribution disk at high speed in RAD: or in your hard disk using FMS:, and
then to copy it to floppies directly.

In the same vein, \SuperDuper\ will act slightly differently when determining
if a buffer contains a DOS disk (if not, the incrementing of the date is
inhibited even if selected). At read time, the information is recorded, but
if at write time the pass starts from track 0, \SuperDuper\ will re-fetch the
DOS mark from the buffer and check it again. This way if for instance you
externally feed a ramdrive.device with a diskcopy command \SuperDuper\ will be
aware of it and will increment the date if requested to do so.

Some care must be taken in order to obtain what you really want when mixing
the buffering features and the selection of the start/end cylinder.
\SuperDuper\ implements a reasonable mean of flexibility and reliability for
these kinds of operations.

When using {\twelvett VDBUF} or {\twelvett HDBUF}, the read/write operations start and end exactly
where you specify with the start/end cylinder gadgets. Since \SuperDuper\ has
no control over what you do to the virtual disk while it's not accessing it,
it has to assume you made it right.

When using a RAM buffer, \SuperDuper\ can clearly make some assumptions on its
validity. In particular, just after allocation or a stopped {\twelvett READ} it assumes
the buffer is not valid.

If you have a valid buffer and you change the start/end cylinders, there are
two cases: either the buffer range and the start/end range do not intersect,
in which case an error message is issued if you try to write the buffer, or
there is a non-empty intersection, in which case the intersection will be
written, i.e., the starting track will be the greatest of the start of the
buffer and the start cylinder, while the ending track will be the least of
the end of the buffer and the end cylinder. Example: if you read something
with {\twelvett SC=20}, {\twelvett EC=30}, then you set {\twelvett SC=10}, {\twelvett EC=25} and hit {\twelvett GO}, the range 20-25
will be written.

There are however two subtle differences between the behaviour of a complete
(880K) RAM buffer and a partial one. First of all, the track range chosen
for {\twelvett READ}ing in a complete RAM buffer is always the full start/end cylinder
range, while if reading in a partial buffer \SuperDuper\ will start from the
last track of the previous buffer (of course, if the last track is past the
end cylinder, it will start from the start cylinder). Moreover, if a long
range of tracks is skipped (for instance, you read in a buffer range of 0-79
and you write 70-79) a few (less than 10) seconds will pass while \SuperDuper\
unpacks the data you don't want to write---they have to be decompressed
anyway.

If all this scares you, don't fear: the buffer/range interaction will simply
work just as you intuitively expect. I hope at least {\twelvett 8\^{ })}.

\vskip 2cm
\centerline{\big User Interface Gadgets}
\vskip 1.5cm

\item{$\bullet$}{{\twelvett TALK} activates \SuperDuper 's ability to give its status by voice.
Currently only English is supported.}

\item{$\bullet$}{{\twelvett AUTO} activates automatic operation starting. \SuperDuper\ will monitor disk
insertion and ejection. When all destination(s) have been ejected and
re-inserted, a GO operation is started. If {\twelvett FORMAT} is selected, the
destination(s) are formatted. Else, if a buffer is selected, it is written
to the destination(s). If neither formatting nor buffering is requested,
\SuperDuper\ will monitor the source, too, and will start a disk-to-disk(s)
copy as soon as the source and all destination(s) have been ejected and
re-inserted. {\bf WARNING:} especially on one-drive-only systems, {\twelvett AUTO} can be
extremely dangerous. You'd better write-protect your source disks.}

\item{$\bullet$}{{\twelvett KILLSYS/RESTORE} closes the Workbench and voice, flushes the memory and
opens a very small screen with only two colors. Moreover, the window is of
SIMPLE\_RE\-FRESH type rather than SMART\_REFRESH. This way, the maximum amount
of memory for your system is at your disposal (unfortunately, under 1.3 the
window can be refreshed incorrectly because of an Intuition bug). If the
Workbench can't be closed for some reason, a warning is issued (usually some
application has a window opened on the Workbench screen). When you want to
get back, hit the gadget again (this time it will be named {\twelvett RESTORE}). This
feature is very powerful if coupled with \SDBootInstall\ and with the CLI
option {\twelvett LOWMEM}.}

{\bf WARNING:} If you grab the disk.resource (by selecting a source and/or a
destination) just after a disk was inserted, it's likely the Workbench will
be locked, waiting for you to unlock the drive in order to load the icon of
the disk. If in this moment you hit {\twelvett KILLSYS}, you will lock the entire
system, since \SuperDuper\ will be waiting for the Workbench to close, while
the Workbench will be waiting for you to release the disk.


\vskip 2cm
\centerline{\big String Gadgets}
\vskip 1.5cm

The string gadgets have been more or less discussed in the previous
sections. They are gathered here for sake of clarity.

\item{$\bullet$}{{\twelvett SC,EC} select the start and the end cylinders, respectively, for any
operation.}

\item{$\bullet$}{{\twelvett Label} lets you choose a name for the disks formatted by \SuperDuper\ . See
also the paragraph about the {\twelvett INCNAME} gadget.}

\item{$\bullet$}{{\twelvett VDName, VDUnit\#} select the name and the unit number of the Exec device
that \SuperDuper\ will use as a virtual disk if the {\twelvett VDBUF} gadget is selected.}

\item{$\bullet$}{{\twelvett Retry\#} selects the number of read/verify retries on each track.}


\vskip 2cm
\centerline{\big Special Requesters}
\vskip 1.5cm

When \SuperDuper\ needs to inform the user about something, usually a
requester with a message appears (if the {\twelvett TALK} option is on the message is
also read out loud). While most of the requesters are self-explanatory, some
of them need a more detailed description.

{\twelvett Can't get disk.resource}

The disk.resource is the Exec way of controlling the access to the low-level
disk hardware. \SuperDuper\ can't access the resource, probably because
someone is already using it. If you suspect a particular program, close it
and try again to select a disk gadget.

{\twelvett Please free disk.resource}

(See also previous requester). If the disk.resource can't be grabbed, Exec
won't give back the message passed by \SuperDuper\ until the resource is free.
Thus, until that moment \SuperDuper\ can't exit.

{\twelvett Checksum error: buffer munged.}

Someone wrote over \SuperDuper 's RAM buffer. The buffer is no longer valid,
and the current copy is probably munged, too. You should probably reboot,
because if something writes on someone else's memory it's likely it will do
it again.

{\twelvett A track buffer has been munged.}

Someone wrote on one of \SuperDuper 's track buffers. The same comments of the
previous requester apply.

{\twelvett ARexx server not active}

In order to use ARexx macros, the ARexx server has to be activated. Type
{\twelvett RexxMast} at a CLI prompt (if it's not in your path, you should locate it
easily).

{\twelvett Error while recalibrating unit x.}

\SuperDuper\ found an error while recalibrating a drive head. The head was
moved to track 0, but the drive signal DSKTRACK0 wasn't activated. This
means that either your drive has lazy signals, in which case there's nothing
to worry about, or that some head step wasn't actually performed (possibly
because of power supply reasons) in which case the last copy could be bad,
even if {\twelvett VERIFY} is on. Better {\twelvett CHECK} it. Try also to increase the step and
calibrate delays of the drive with \SetTDDelay.


\vskip 2cm
\centerline{\big ARexx Interface}
\vskip 1.5cm

ARexx is the system macro language of the Amiga. It was originally developed
by Bill Hawes (to whom every Amiga owner owes much more than he probably
realizes) and was then included in the release 2.0 of the operating system.

ARexx is a beautiful interpreted language, with unique features such as
syntax/se\-mantics collapsing (for instance, you can ask the value of a
variable given its name as a string) and, overall, the ability to interface
itself with external applications. A single ARexx script can control several
different programs and make them interact.

The ARexx interface consists of a port, which is used for communications,
and a set of commands that ARexx can issue to the application. For
\SuperDuper, the port name is \SuperDuper , and the command set is described
below. ARexx scripts written for \SuperDuper\ should have extension {\twelvett
supdup},
like {\twelvett foobar.supdup}. This is in order to distinguish ARexx scripts written
for different applications.

ARexx provides at little or no implementation cost a powerful macro language
which substantially increases the performance and the versatility of an
application. Maybe some feature you would like to have is not in \SuperDuper\
at this time, but it's very likely you'll be able to put it in via a
suitable ARexx script.

Besides being able to execute commands issued by an ARexx macro, \SuperDuper\
is also able to start an ARexx macro. This is indeed the purpose of the
AREXX gadget (the last one in the last row). The gadget is activated if 1)
the rexxsyslib.library is somewhere in your LIBS: directory and 2) you have
a file requester. \SuperDuper\ is able to recognize and use the ASL file
requester (under 2.0), the ARP file requester or the req.library file
requester (the first available in this order will be used). You can start
any number of macros at the same time (beware of wild interactions though).

\noindent {\bf General issues}

\SuperDuper\ commands generally correspond to gadgets, and are similarly
named: for instance, the command {\twelvett CHECK} will check the source drive, while
{\twelvett VDUNIT 4} will set the virtual disk buffer unit number to 4. Commands are
case insensitive, and only the first two or three letters are significant.
So you can write {\twelvett CH} instead of {\twelvett CHECK} but you have to write {\twelvett REA} for {\twelvett READ}, or
you could make confusion with {\twelvett RESTORE} or {\twelvett RETRY}.

ARexx needs a console by which it communicates with the user. If you started
\SuperDuper\ from the CLI, the your original CLI will be used. Otherwise, a
console window will be opened. Under 1.3, this window appears at the start
of any ARexx macro and gets closed when there is no macro running. Under 2.0
it's always open, but it's an {\twelvett AUTO} console window, so you can close it if
you wish: it will be reopened as soon as something is printed into it.

\noindent {\bf Action commands}

The commands {\twelvett GO, READ, CHECK} and {\twelvett STOP} act just like their gadget
counterparts, starting a copy (buffering, formatting) process or stopping
it. The first three return at the end of the operation. However, for
instance, if another task sends a {\twelvett STOP} command while a copy is in progress,
the copy is interrupted and the {\twelvett GO} command returns. You can then check the
{\twelvett RC} variable to see what happened (see the ARexx manual).

The pair {\twelvett KILLSYS} and {\twelvett RESTORE} work like the corresponding gadget. The
operations which are nonsense have no effect (i.e., if you send {\twelvett KILLSYS} and
the system has already been killed, nothing happens).

\noindent{\bf Selection commands}

I list here for sake of completeness the whole group of selection commands.
They could be easily deduced anyway from the gadget names, apart from
{\twelvett VOLUME}, {\twelvett PRINTERRORS} and {\twelvett RX} which are
available only through the ARexx interface. Here {\twelvett <string>} is a
string of characters and {\twelvett <n>} is a nonnegative number. When
{\twelvett on/off} is specified as an argument, you have two ways of invoking
the command: {\twelvett <command> on} will switch the thing on, and {\twelvett
<command> off} will switch it off. Note that the ARexx interface of
\SuperDuper\ is rather lazy about syntax---strings too long will be silently
truncated, and passing a non-numerical argument where {\twelvett <n>} is
required will usually produce a value of 0. {\parindent=4.1cm \item{\twelvett
SOURCE <n>}{Selects drive n as source;} \item{\twelvett SOURCE off}{Turns
off source drive;} \item{\twelvett DEST <n>}{Selects destinations using n
as a bit mask. For
                instance, 0 selects no drive, 1 selects drive 0,
                5 selects drives 0 and 2, 15 selects all destinations;}
\item{\twelvett BUFFER on/off}{Controls the RAM buffer;}
\item{\twelvett HDBUF on/off}{Controls the hard disk image file buffer;}
\item{\twelvett VDBUF on/off}{Controls the virtual disk buffer;}
\item{\twelvett VERIFY on/off}{Turns on/off verify;}
\item{\twelvett DATE on/off}{Turns on/off date adaptation;}
\item{\twelvett INCNAME on/off}{Turns on/off name increment while formatting;}
\item{\twelvett FFS on/off}{Selects Fast File System or Old File System while formatting;}
\item{\twelvett TALK on/off}{Toggles talk mode;}
\item{\twelvett AUTO on/off}{Toggles auto mode;}
\item{\twelvett LABEL <string>}{Sets the disk label}
\item{\twelvett RETRY <n>}{Sets the number of retries;}
\item{\twelvett VDUNIT <n>}{Sets the virtual disk unit number;}
\item{\twelvett VDNAME <string>}{Sets the virtual disk unit device name;}
\item{\twelvett SCYL <n>}{Sets the start cylinder;}
\item{\twelvett ECYL <n>}{Sets the end cylinder.}

\noindent The following commands are only available through the ARexx interface:

\item{\twelvett VOLUME <n>}{Sets the volume of the beeps (0$<$=n$<$=63);}
\item{\twelvett PRINTERRORS on/off}{Opens/closes \SuperDuper 's detailed error report window;}
\item{\twelvett RX <string>}{Executes the ARexx macro named {\twelvett <string>};}
\item{\twelvett NOCLICK <n>}{Forces \SuperDuper\ to not click the drives specified
                by {\twelvett n} as a bit mask (the same format of {\twelvett DEST}).}}
\vskip 2cm
\noindent {\bf Return codes}

Commands issued by ARexx to an application should return useful values in
order to tell what really happened. Generally, a command which fails returns
an error level, while a successful command returns an error level of zero
and, upon request of the caller via the {\twelvett OPTIONS RESULTS} command, a result
string which can be parsed in order to get useful information.

\SuperDuper\ returns an error code of 10 if the syntax of the command was
wrong. This will cause ARexx to complain with an error message. An error
code of 1 is returned if the syntax was right but the command couldn't be
executed, but there is no real failure (for instance, if you send {\twelvett GO} while a
copy is already in progress or if you try to select a ghosted gadget). An
error of 30 is returned in extreme cases, for instance when you hit the
close gadget and there are still some commands pending. No strings are ever
returned, since we have only a few cases to differentiate. Return codes with
special meanings are returned by the following commands:

\vskip 0.5cm
\halign{\hfil\twelvett#&\quad\hfil\twelvett#&\quad#\hfil\cr

SOURCE,DEST&           2: &The selected drive is not connected.\cr
                &5: &The disk.resource is not available.\cr\cr

BUFFER,
HDBUF,
VDBUF&           5: &The buffer cannot be allocated.\cr\cr

BUFFER          &2: &A full buffer cannot be allocated. Compression is on,\cr
                &   &and there is the possibility of multi-pass copies.\cr\cr

GO, READ,
CHECK           &2: &This pass is not the last one.\cr
                &3: &Something is wrong with the chosen source, destination \cr
                &&and buffer options. For instance, you're trying to copy \cr
                &&from df0: to df0: without a buffer.\cr
                &4: &The buffer is not valid.\cr
                &5: &A unit is empty.\cr
                &6: &A unit is write-protected.\cr
                &7: &The start/end cylinders chosen are meaningless. This can\cr
                &&happen if the numbers are out of range, or (for a RAM-buffered\cr
                &&{\twelvett GO}) if there is no intersection with the current buffer.\cr

                &8: &There were errors.\cr
                &9: &There were errors. Moreover, this pass is not the last one.\cr
                &20: &Someone munged the RAM buffer or the track buffer.\cr\cr

TALK            &5: &The voice system cannot be activated.\cr\cr

KILLSYS,
RESTORE         &20: &The current window has been closed, but it was impossible\cr
                &&to open the new one. The program exits in this case.\cr}


{\bf What can I do with ARexx?}

Basically you can expand \SuperDuper 's capabilities and/or make it interact
with other programs. A couple of examples of the first case could be a
{\twelvett CheckAll.supdup} macro which checks all drives in sequence. The ``native"
\SuperDuper\ can only check one drive at a time, but if you have two or more
drives you can check many drives using a macro like
\vskip 0.3cm
{\obeylines\twelvett\parskip=0pt
/* CheckAll */
do i = 0 to 4
   source i
   if rc==0 then check
end}

After checking you should of course look at the return codes in the {\twelvett
RC} variable and decide upon appropriate actions.

Suppose now you have four drives and you want to make a copy of two
different floppies. You can put the sources in drives 0 and 1, the
destinations in drives 2 and 3, and then

{\obeylines\twelvett\parskip=0pt
\vskip 0.3cm
/* DoubleCopy */
source 0
dest 4
go
source 1
dest 8
go}

(of course I'm assuming \SuperDuper\ is in its default configuration). This
will produce the two copies in a completely unattended way.


\vskip 2cm
\centerline{\big CLI Line Options}
\vskip 1.5cm

When you start \SuperDuper\ from the CLI, you have the chance to specify an
option. The possible options are printed in the standard Amiga template
format if you type {\twelvett SD ?}. In this case, the following line

{\twelvett NOGUI/S,LOWMEM/S}

\noindent will be displayed. Its meaning is that {\twelvett NOGUI} and {\twelvett LOWMEM} are switches that
you can activate. For instance, the command line {\twelvett SD NOGUI} will invoke
\SuperDuper\ in its {\twelvett NOGUI} mode. The two flags are mutually exclusive---if both
are specified, only the first one counts.

{\twelvett NOGUI:} \SuperDuper\ won't open its graphical user interface but you
can then control it through the ARexx interface. This makes possible to write
an ARexx macro allowing you to use \SuperDuper\ from the shell much as the
diskcopy command. Moreover, the startup file {\twelvett Startup.supdup} is not
executed, so that in your ARexx macro which calls \SuperDuper\ directly you can
expect to get the standard configuration.

{\twelvett LOWMEM:} This switch shuts down the ARexx port and the sound system.
\SuperDuper\ won't open either the ARexx port/rexxsyslib.library pair, or the
audio.device. This mode is provided for user with 1M or less who want to
have as much free memory as possible (read also the section about
\SDBootInstall).

These options are only available from the CLI.


\vskip 2cm
\centerline {\big Startup File}
\vskip 1.5cm

At startup time, \SuperDuper\ checks if ARexx is available, and in this case it
tries to start an ARexx macro named {\twelvett Startup.supdup}. This file
should contain your usual settings: for instance, it's a very good place where
to put a {\twelvett VOLUME} command. The startup file is a regular ARexx macro,
just like any other one started by the {\twelvett AREXX} gadget or by the
{\twelvett RX} command. However, a couple of conventions were implemented in
order to get a better behaviour on systems without ARexx. In particular, the
absence of the ARexx server or the ARexx error message {\twelvett Program not
found} will {\it not} be displayed if caused by the startup file. Notice that
the last message can also be caused by the first line of {\twelvett
Startup.supdup} not being a comment (every ARexx macro must start with a
comment).

\vskip 2cm
\centerline{\big SDBootInstall}
\vskip 1.5cm

When your system boots up (at power on or after a reset), the operating
system searches for available drives, and creates some trackdisk.device
tasks accordingly. These tasks take a lot of memory for their buffers
($>$30K), but \SuperDuper\ doesn't use them at all, because it has its internal
routines.

If you have to do intensive copy work, and you have 1MB of memory or less,
you could find it useful to boot up your system in a special configuration
that will shut down almost all trackdisk.device tasks, thus freeing a lot of
memory.

To accomplish this, do as follows:

\item{1.}{Make a copy of your usual Workbench 1.2 (or greater) disk (from now on we
work on the copy).}

\item{2.}{Delete some programs to make room---preferences, diskcopy and format are
good candidates. Moreover, delete the file {\twelvett Disk.info}.}

\item{3.}{Copy \SuperDuper\ to the disk root directory (by dragging its icon on the
disk icon or using the CLI).}

\item{4.}{Edit the {\twelvett startup-sequence} of the disk (it's in the
{\twelvett s} directory). Delete it entirely, and substitute it with

{\obeylines\twelvett\parskip=0pt\vskip 0.3cm
SetPatch >NIL:
Run >NIL: <NIL: SD LOWMEM
EndCLI >NIL:}

If you're under 1.2, don't put in the first line (you don't have a {\twelvett
SetPatch} command).}

\item{5.}{Now put the disk in df0:, and run the utility \SDBootInstall. A special
bootblock will be installed on the floppy. When booting from it, the
operating system (and you) will be able to access only drive 0---the other
ones will be for \SuperDuper 's use only. To get back to normality, a reboot
is necessary. You will gain 30/40K per drive using this method (for techies:
it is perfectly legal---the bootblock simply AllocUnit()s the drives with
ID$>$0).}


\vskip 2cm
\centerline{\big SuperDuper and Your System}
\vskip 1.5cm

\SuperDuper\ has been written keeping in mind that a good program doesn't have
to eliminate everything from the system in order to work. The Amiga has a
very efficient multitasking kernel which allows for resource arbitration.

When \SuperDuper\ is started, it won't allocate anything from your system. As
soon as a source/destination gadget is clicked, it will inhibit all of the
drives (so don't select a gadget while reading or writing to floppies) and
then will grab the disk.resource. Until the resource is released, {\it no one
else} can access the Amiga drives. This is necessary in order to avoid
unpredictable collisions with the system or other programs. Inhibiting the
drives is not enough, since some other file system (like CrossDOS) could
access them.

If you need to temporarily access your drives, you must simply deselect all
\SuperDuper\ source/destination gadgets: the disk system will be restarted (it
will be re-grabbed on a gadget selection of course). The heads will be moved
to their original position, so that you no longer need to eject the drives
under 1.3.

The CPU use of \SuperDuper\ is almost unnoticeable. You can do anything else,
and you shouldn't notice any slowdown. In particular, if no
source/destination is selected \SuperDuper\ is completely asleep.

This however is not true if you use compression. In this case, not only will
the system be slowed down (a priority 0 task will almost always be active),
but {\it any} operation (including moving the mouse pointer) will slow down
\SuperDuper.

If you use the utility \ToggleClick\ distributed with \SuperDuper\ (or any other
utility which legally kills drive clicks under 2.0) \SuperDuper\ won't click
empty drives (drive clicking is necessary for monitoring disk insertion;
using \ToggleClick\ is good but you must be sure your drives won't try to move
past track 0 if asked to do so). Anyway, you always have the chance to
selectively force {\twelvett NOCLICK} via the corresponding ARexx command.

You should avoid running \SuperDuper\ while a 16 color hi-res screen (or a
4-color ECS productivity mode screen) is displayed. The video DMA access
will interfere with the disk/CPU/Blitter access to the point that copy times
will rise to incredible values---reading and compressing a disk in the
buffer can take more than 100s.


\vskip 2cm
\centerline{{\bigss SuperDuper }\big and You}
\vskip 1.5cm

``Well," you could say, ``\SuperDuper\ is a great copier---but how can I trust
it for making my copies? This guy diddles with hardware---maybe I should use
the system {\twelvett DiskCopy} command."

This is not a good idea. First of all, \SuperDuper\ is {\it incredibly} picky
about verifying. You will get more verify error messages than with the
standard copy commands (for techies: \SuperDuper\ verifies also the MFM timing
bits, not only the data bits; this means a 200\% efficiency improvement in
catching verify errors and generally bad media).

Moreover, both the 1.3 and the 2.0 trackdisk.device have unpleasant
side-effects on frequently read/written tracks. These side-effects are
cleared when you do a copy of the disk with \SuperDuper\ (for techies:
trackdisk.device doesn't check for MFM bits being read in correctly, and
doesn't re-MFM the track before writing it; it just re-MFMs the changed
sector. If a MFM timing bit is read wrong, it will stay wrong forever,
possibly causing read errors; but \SuperDuper\ re-MFMs every track it copies,
thus restoring every MFM timing bit to its correct value).

Finally, if you don't like coffee-breaks during your copies, you'd better
use the fastest copier available---namely \SuperDuper . Note that if you have
four drives and you use top-quality disks, so you can skip verify, the
buffer system allows you to get a per-copy time of 9$1\over 2$ seconds, which is
definitely not bad.


\vskip 2cm
\centerline{\big Copy Protection}
\vskip 1.5cm

\SuperDuper\ won't copy protected disks (or if it will it's just a
coincidence). I do not believe in copy protection. Scrambled tracks will
produce random data on the destination. If the read error goes beyond a
simple checksum error don't expect anything meaningful to be written on the
destination disks.

However, \SuperDuper\ will faithfully reproduce data block checksum errors
({\twelvett Disk foo has a read/write error}) or DOS checksum errors
({\twelvett Key 880 checksum error}) on the source disk in disk-to-disk copies
(header checksum errors are fixed when renumbering the sectors). Thus, if you
got the typical {\twelvett Key <n> checksum error}, you can make a copy of the
disk before fixing it. \SuperDuper\ won't do any surgery: use a good tool (such
as {\twelvett DiskSalv} or {\twelvett FixDisk}) for that. Avoid {\twelvett
DiskDoctor}. On the other hand, during buffered copies data block checksums
will be silently fixed by recalculating the right checksum.


\vskip 2cm \centerline{\big Acknowledgments} \vskip 1.5cm

The first person I must thank a thousand times is Dirk Reisig. It was by
means of his suggestions that I sped up \SuperDuper\ to the current, amazing
level. I wrote him a letter which he answered gently with a long explanation of
the optimizations performed by {\twelvett PCopy}. The first time I read the
letter it seemed greek to me, but little by little I learned all the mysteries
of MFM encoding and disk direct hardware driving. Moreover, I learned from the
source code of {\twelvett TrackSalve} the usage of the blitter for MFM encoding
and many other subtle things. In other words, without the help of Dirk you
would have never seen anything after DFC5 (for release 2.0, a new optimization
was introduced; it was suggested by Dan Babcock).

The second guy behind the birth of \SuperDuper\ is Tom Rokicki. He pushed me to
write a substitute for {\twelvett TurboBackup}, and overall suggested the main
thing---that on the Amiga it is possible to write many disks at the same time.
Without this trick, you could never do four non-verified copies in 38s. Tom
also tested all pre-whatever-greek-letter versions, always giving useful
comments... and risking the life of his drives {\twelvett 8\^{ })}. Moreover, I
had time to work on \SuperDuper\ because the Amiga\TeX\ system is so incredibly
efficient I got a lot of spare time while writing math papers...

Last but not least, Randell Jesup at Commodore drove me through the
labyrinth of non-specified-specs, hardware quirks, strange behaviors, and
system esoteric features. Without his help \SuperDuper\ could probably work...
but I wouldn't trust it for {\it my} copies {\twelvett 8\^{ })}.

The name \SuperDuper\ popped up during a rather intensive BIX discussion. Many
other names were proposed, but in the end I chose this one---it has
symmetry, correctly defines the product and has a simple shortening (SD).
Thus, a thousand thanks to Kent Kalnasy and Dan Barrans for suggesting this
name. Many features were not my ideas. An incredible number of BIX users came up
with excellent suggestions, many of which were actually implemented. Thanks
to them you have support for buffering on any device (I never use RAD: nor
FMS:, so I didn't think it could be useful).

But, as always, the biggest {\it thanks} goes to the beta-testers of \SuperDuper :
Dennis Atkin, Michele Battilana, Vittorio Calzolari, Jim Cooper, Doug
Erdely, Charlie Fair, Blaine Gardner, Robert Jenks, John Jones, Kent
Kalnasy, Robert Kesterson, Paul King, Randy Menzer, Linda Munson, Davide
Repetto, Tom Rokicki, Sergio Ruocco, Carlo Santagostino, Reinhard Spisser,
Jeff Todd, Carlo Todeschini, Michael Scott Velez and Marco Zandonadi.
Beta-testing a copier is different from anything else---if it doesn't work
you won't get a marginally corrupted picture on your display: rather, the
Fish Disks it took an hour to copy could be unusable. A special kind of
patience is needed under these conditions {\twelvett 8\^{ })}.


\vskip 4cm
\centerline{\big Disclaimer and Author Info}
\vskip 1.5cm

\SuperDuper\ is \copyright\ 1991 Sebastiano Vigna and it's freely distributable as
long as all of its files are included in their original form without
additions, deletions, or modifications of any kind, and only a nominal fee
is charged for its distribution. This software is provided ``AS IS" without
warranty of any kind, either expressed or implied. By using \SuperDuper , you
agree to accept the entire risk as to the quality and performance of the
program; don't come to me if you destroy your entire Fish Disk library with
it! Of course, it was tested rather extensively before it was released...

Comments, complaints, desiderata are welcome.
\vskip 2cm


{\obeylines\parskip=0pt
\sf  Sebastiano Vigna
    Via Valparaiso 18
    I-20144 Milano MI
\vskip 1cm
    BIX: svigna
    UUCP:seba\[e-mail]
         ...$\{$uunet$|$piramid$|$rutgers$\}$!cbmvax!cbmehq!cbmita!sebamiga!seba
    FIDO: 2:331/301.6 (aka 2:21301/6)}
\bye