ISAD/SAD's user interface consists of one window and a number of
panels. When you first start the program, you'll be presented with
the Knowledge window. If you wish to save what SAD learn's from
your machine, you will need to fill in the bottom field, if you
want to use an old knowledge file, you'll need to fill in the top
one and if you wish to use knowledge files all the time, you should
fill both with the same name; in this case, SAD is smart enough
to backup your old file before overwriting it. Using knowledge files
is optional.
Then, you can select the paths you like to search or to skip.
ISAD/SAD will always use what is in the knowledge file if one
was used, add files in the paths you provide, and then remove
all files that are to be skipped.
You will then have to decide which mode you want ISAD/SAD to function
at. In the ISAD mode, all images specified in the paths you provided
will be scanned and signaturized and in the SAD mode, only basic
file information is queried from the Operating System. (Note:
Like most other fields in this program, these buttons will remember
their old values, so you can simply press the "Continue >>" button
a few times and skip all these windows)
SAD will then start scanning your drives and if required, examining
the image files. This can take as few as a couple of seconds or
many hours, depending on the values you've specified and the content
of your file system. Generally, 10 seconds is the time needed
for each Gigabyte of drive space (HPFS) in SAD mode and about
3 seconds average per image additional in the ISAD mode.
Now, SAD is ready to search its knowledge (ie. your system) for
files or duplicates.
The
searchable items are:
Expression
search (similar to OS/2 Warp's Find)
- Match
Expression. Searches all of your system for a file pattern,
for example use *.zip to list all zipped files. Once you have
reached this window, you can expect each expression search to
take less than a few seconds, even for the largest file systems.
This field can be used with other search methods.
Searches
based on name
- Duplicate
names. Searches for duplicate file names. For example, select
this button and enter an expression *.EXE and ISAD/SAD will
find all duplicate executables on your system.
- Ignore
image extensions. Searches for duplicate file names while ignoring
their extensions; used to find various formats of the same image,
for example apple.gif and apple.jpg.
- Smart
renames. Searches for duplicate file names while changing some
similar characters, for example apple_1.gif and apple-1.gif
would be considered the same.
Searches
based on size and content
- Duplicate
file sizes. Searches for files which are of the same size. If
used w/o an expression, an average OS/2 system will have most
of the files match in size with other files. You will want to
use this with an expression similar to *.zip.
- Identical
contents. Select this to search for files which are 100% identical
to each other. this is the absolute best way to find duplicate
files and works real well when used with expressions similar
to *.zip, *.jpg, ...
Searches
based on image information
- Image
Dimension. Search for images which are of the same size, only
good for small image collection of non-standard (i.e. not 640x400,
...) sizes.
- Scaled.
Search for images which are of the same size or a proportional
scales of each other. Also only good for small image collections
of non-standard sizes.
- RGB
Signature. Find image's whose RGB signature is the same in additon
to similar dimensions (or scaled dimensions)
- Grayscaled.
Similar to above but it will also consider images which may
have been grayscaled.
Searches
based on image type
- Other
methods. Search for a certain image types, for example images
which are the size of WWW banners, images which are very small
or very large.
Once you have selected your search, it will be processed.
And the search results displayed. Now, simply press the Red buttons
to destroy images (be careful, you won't be asked to confirm it)
or the Green buttons to keep them. You can move between the groups
of similar images by pressing the left or right arrow buttons
and perform more searches by pressing the "Done" button.
A
few example results are shown below:
A
search based on duplicate file sizes. This was a bad move
since most files on the system matched other files. In fact, 34244
files had duplicate sizes.
Search
based on duplicate file names. This is a better search. The
system contained many files which had the same file name, still,
an expression such as *.exe or *.zip would have reduced the number
of matches found. And exclusion of multiple OS installs would
have also been a good move.
Search
based on an expression. This search looked for any files which
ended in *.zip. There was at least 20 Megs saved by this one alone.
Search
based on image size and signature. This search looked for
images which had the same RGB signature, may have been rescaled,
or may have been grayscaled. This type of search does not usually
find as many matches as searching for size and content but is
great for finding similar images. In this case, Claudia's picture
was found in two different sizes as well as grayscaled.
What
are the System Requirements
ISAD/SAD's
requirements are
-
Warp 3.0 (FP 17 or later) or Warp 4.0. (no FP required)
- A 486 or better processor. (See note below)
- 16M RAM minimum.
- 20M Swapper minimum.
- 800x600 graphics card.
- Internet connection with access to WWW. (To register product)
Note:
In the image mode, at least for the first time, ISAD/SAD will
need to load and examine each image file. A fast processor (or
you may wish to have it be done during lunch or overnight), 32
Meg RAM and much more Swapper space is required if there are many
images to be examined. Using the knowledge file feature is recommended
in this case.
How
much does it cost and who sells it?
MD+F ISAD/SAD for OS/2 Warp is now available for FREE. To download,
visit the [link removed] on this site. If you have previously
purchased MD+F ISAD/SAD, hold on to your SN for the future.