D

Delete Key
The Delete key on a keyboard sometimes causes confusion, especially on Macs. In the big picture, there's two delete keys, that sometimes go under different names.

There's the northeast-corner key, above the Return key, that is typically called 'backspace' on PC-style keyboards, but 'delete' on Mac style keyboards. It sends the Backspace ascii character (0x08). This is what you use to delete backwards while you are typing in Interrogator, whether by Touch Targeting, renaming, or into a dialog box of some sort.

Then there's the Delete key on the little keypad above the arrow keys on a desktop keyboard. Sometimes labeled "Del" key, it sends the Rubout ascii character (0x7F). That one deletes files in Interrogator. In some text editors, it deletes characters forward.

various delete key locations

Most Mac Laptops seem to not have these. You can use cmd^T instead for the Delete function in Interrogator. The Backspace key is reserved for textual backspace.

Of course, your keyboard may vary, especially if non-US, old or proprietary, or if it's remapped (eg on X11). If you're not sure, you should experiment around before using these keys.

Depletion
Depletion is when files use up disk space. Also sometimes confusingly called Usage.

Most volumes (disks or disk partitions) have some sort of "total capacity", these days measured in gigabytes. As you make files, install software, and download data, your volume slowly fills up. When full, your software will get error messages that hopefully complain to you that the disk is full (sometimes they say other things). Usually you want to avoid this because sometimes software will fail in unexpected ways. Many of the techniques you would want to use to clean up involve - surprise - using up more disk space.

On a small machine, the total capacity of a volume or disk usually doesn't change unless you back up the data, rip it up, repartition, reformat, and restore the data. Usually you do something else if you need more disk space. Larger machines have ways of requesting more disk space from yet another disk somewhere.

Usually, files deplete disk space in chunks called "fragments" or "blocks" or "clusters". We call this the Depletion Block Size. The depletion block size might be 512 bytes or 1024 bytes or 4096 bytes or something else. The leftover space (sometimes called "slack space") is usually wasted. Because of this, the total depletion of your file is usually not the same as the Content Size in bytes.

For example, say your depletion blocksize is 1024 bytes on an EXT2 volume (typical on Linux) or an HFS+ volume (typical on MacOS).

You make a file that contains just the text "Hello".
Total size: 5 bytes. Total depletion: 1024 bytes.

You make a file that contains exactly 1024 bytes.
Total size: 1024 bytes. Total depletion: 1024 bytes.

You make a file that contains exactly 1025 bytes.
Total size: 1025 bytes. Total depletion: 2048 bytes.

You make a file that contains exactly 1,024,000,001 bytes.
Total size: 1,024,000,001 bytes. Total depletion: 1,024,001,024 bytes. (That's one million and one depletion blocks.)

You make a file that is totally empty.
Total size: 0 bytes. Total depletion: 0 bytes.

The depletion works differently on other kinds of volumes. On UFS volumes, typical on Solaris and BSD variants, small files, and ragged ends of larger files, share blocks, to cut down on this waste.

The depletion is not always more than the file size. Some filesystems support sparse files, where blocks that are composed entirely of zero bytes are not stored and therefore use no depletion. Other filesystems support compression; the total depletion might be half of the file size, or less.

In addition to depletion, a volume might have other resources that run out. Usually there is a fixed maximum number of files that can exist on a volume. On Unix these are called i-nodes. When a volume is created (formatted), some default number of i-nodes are accomodated; this default is usually taken to be very large because it's a hassle to run out. (You might get a poorly chosen error message, something about "out of space", but you'll look and you still have plenty of space left). If you have a lot of small files on a particular volume, such as a million, you might run out of i-nodes. Usually, the only way to get more i-nodes is to reformat and use some option on the mkfs command (or whatever).

Depletion Block Size
The quantum that files deplete disk space by. Also called Cluster Size, Allocation Block Size, or similar names. See Depletion for a full discussion.

Depletion Size
The actual amount of disk space consumed by a file or a directory full of files. This is usually more than the Content Size. See Depletion for a full discussion.

Depletion Uncertainty
When Interrogator tallies depletions on your volume, it has two sources of information:
* the volume itself knows how much it is depleted
* each individual file knows how much depletion it uses up.

To calculate a directory's depletion, Interrogator has to add up the depletion of all of the files and directories inside it. This is a tedious process and you might notice a lot of disk motion sometimes when it is doing this calculation.

After it's done tallying up the depletions of all the files and directories on your volume, it compares it against the volume's total depletion (usage), which the volume keeps track of separately. Usually there is a gap, this gap is called the Depletion Uncertainty. This happens because there are some directories that Interrogator cannot scan through. There are several reasons why:

- If you don't have permission to see in a directory, Interrogator doesn't either, so it skips it. This is the most common reason. (If you run as the superuser, of course, this never happens.)

- If the directory is a mount point for another volume, the original contents are hidden. Usually there are no files in the hidden directory, but there could be some. Interrogator cannot access this hidden directory so it skips it. The depletion of the mounted volume is not included in the depletion figures for this volume, of course.

- If there is some sort of I/O error while trying to read through a directory, Interrogator similarly skips it.

- Since tallying takes time, perhaps several minutes, directories that are as-yet untallied cannot yet contribute to the total. This is frequently true when Interrogator has just started up or just opened a window onto a directory for the first time. The depletion uncertainty is a moving target.

- Any directory that has a subdirectory with one of the above problems will, in turn, have an uncertain depletion.

Directories that cannot be completely tallied show up as two concentric directory icons. The outside icon shows the maximum possible size, the inside icon tells the minimum possible size. Usually the truth is closer to the smaller inside icon.

For instance, lets say Interrogator determines that the depletion uncertainty for a volume is 100 megabytes. There are ten directories that are untallied on that volume. The missing 100 megabytes might be spread out, 10 megabytes per directory, or it could be all lumped into one directory, in which case the other nine are empty. Since we don't have permission, we just don't know the truth, and we cannot guess. The "average" case is when the depletion is spread out, so usually the minimum is a closer guess than the maximum.

Directory
A directory is a file (filesystem object) that groups together regular files, directories, and other files. Directories do not have any restriction on the amount of depletion they can contain, except that the total space on the volume cannot be exceeded. Space not used in the directory is available for use by other files and directories in the same volume.

Every volume has a root directory. This directory contains directories and other files. All of the files on the volume are either in this directory or in one of the directories at the root.

In Interrogator, you can see all the directories on a volume with a Volume window. You can see all of the files in a given directory with a Directory window: just double click on any directory in a Volume or .

A directory is a mechanism for hiding complexity. On the inside, a directory hides files that are outside of it, so that the humans can oncentrate on a small set of them. On the outside, it hides the details of the individual files under one name, so at a higher level humans can deal with them as a group. The computer itself identifies files in a volume by iNode number; at this level, directories are superfluous.

Directory Window a typical directory window
A Directory Window shows the files in a directory as icons with names.

There is one directory window for every directory in your system. If you open another directory, it goes in its own window. Notice that a directory window is bound to its directory; you cannot redirect a window to a different directory as you would a web browser page. If you try to open a given directory a second time, the same window will be brought to the front.

From a directory window, you can open its Volume Window, by typing ^V ^O or, on X11, altV altO. Not only will the volume window for the directory's volume be opened, but the directory itself will be exposed in the hierarchy for your convenience.

There are a lot of commands and viewing options for directory windows; see the Directory Window section for more details.

Disk
An electromechanical device that holds data. Typically there is one or more disk platters, and data is written on one or both sides, sometimes in layers, by heads that read and/or write data using electromagnets, lasers, or other technologies. From the computer's point of view, it's just a big storage area for bytes.

Many kinds of disks are removable, meaning that you can command your computer to disengage and spit out the disk so you can store it on a shelf and insert a different disk. This way, you can have almost unlimited storage. Most removable disks have only one partition and therefore result in just one volume.

Hard disks are usually installed permanently in a computer. They are the fastest of all kinds of disks. Their storage is usually measured in gigabytes. Usually hard disks are partitioned and therefore result in more than one volume. Even a hard disk that appears to be one volume, often has hidden partitions, or else, is partitioned into one partition.

Floppy disks are small, plastic removable square modules that only store about 1.4 megabytes. They are often too small for modern usage.

CDs and DVDs are silvery plastic disks that hold anywhere from 0.6 to 17 gigabytes of data. Some are nonwritable (CD-ROMs and DVD-ROMs); the writable ones (CD-RWs and DVD_RAMs) often write slowly, and somtimes cannot be erased (CD-Rs).

ZIP disks, invented by Iomega corporation, are like large floppy disks that hold 100mb, 250mb or more. Other manuracturers make similar kinds of large floppy disks, magneto-optical disk cartriges, and the like.

Disk Mirroring
A way to use two (or more) disks (or partitions) in concert to make one apparent volume. Data is written identically to both disks in unison. Data can be read from either. By keeping both disks working, read performance is increased, although write performance is a little slower. By keeping two copies of all the data, reliability against mechanical disk failure is increased. This is also called RAID 1.

Disk Striping
A way to use two (or more) disks (or partitions) in concert to make one apparent volume. Data is written alternately to the two disks, one "stripe" at a time. (The stripe size can be any number of blocks.) Data is read back from both at the same time. By employing both disks in parallel, read and write performance are both increased. By breaking up the data between two pieces of hardware, reliability against mechanical disk failure is decreased, because a failure in either will destroy the data. This is also called RAID 0.

Disk Striping With Parity
A way to use three (or more) disks (or partitions) in concert to make one apparent volume. Data is striped as in Disk Striping. In addition, another stripe of parity data (or some other kind of checksum) is also stored. Data is read back from all at the same time. With the extra parity stripe, a mechanical failure of any one disk in the array still allows you to reconstruct the data. Various schemes are called RAID 2 through RAID 5 (they differ in the way they do their parity stripes). RAID 5 seems to be the best of them all and often is the only one of the four implemented in a particular system.

Disklabel
A piece of data at the beginning of a disk or disk partition that describes how that partition is divided up as slices.

DS_Store
See .DS_Store.


Documentation > Glossary >
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z   
0123456789 punctuation
       

                     

T a c t i l e   I n t e r r o g a t o r   W e b s i t e