Glossary File

[ Editor's Note: This glossary is not intended to be exhaustive nor even that comprehensive, but rather to provide more information about some terms introduced in our articles, columns, and other editorial.]

ABI
Application Binary Interface, a software/hardware environment that allows precompiled (``binary'') programs written for a particular ABI to run directly on any computer that supports this ABI, without any need for porting or recompiling. Compare to API.

absolute path name
An absolute path name identifies a file in a Unix directory hierarchy starting at the root of the file system (identified by a slash, /). Absolute path names list the directories traversed to reach the target file or directory. Absolute path names always start with the slash, as in /home/ray/cmd.filter.

API
Application Program Interface, a software environment that allows source-code or intermediate-code programs written for a particular API to run on any computer that supports this API, without any need for porting. Compare to ABI.

Archie
An Internet-based facility for locating files available via anonymous FTP.

ASCII
ASCII stands for American Standard Code for Information Interchange. ASCII defines the actual numeric codes stored in a computer to represent characters. The ASCII code contains 127 characters. This includes the control characters, upper and lower case alphabetic characters, digits 0 to 9, and all the punctuation characters found on most computer keyboards.

ASCII collating sequence
The order in which computers normally sort ASCII characters, corresponding to alphabetical order in print. Each ASCII character has a value equal to the numeric value of the byte that represents that character, and lower numerical values come earlier in order. The sequence of non-alphabetic characters can be given:

! " # $ % & ' ( ) * + , - . / : ; < = > ? @ [ \ ] ^ _ ` { | } ~

Awk
Originally developed in 1977 by Alfred Aho, Brian Kernighan, and Peter Weinberger--and hence the acronym name, Awk is a pattern-matching and data-formatting language useful for writing short programs to perform common data-processing tasks. A new version--new Awk, or Nawk--was released in 1985 that incorporated additional features, such as multiple input files, dynamic regular expressions, and user-defined functions.

This utility is found on almost every UNIX variant, and the Nawk version is also available for some other operating systems, such as MS-DOS. It provides extensive programming facilities inspired largely by the C programming language.

BIND
BIND or ``Berkeley Internet Name Domain'' names U.C. Berkeley's implementation of a domain name server. This defacto standard Internet name server is also the basis for several commercial name server implementations.

bin directory
Here ``bin'' is short for binary (program). Thus, a ``bin directory'' is a directory containing binary programs. However, in practice, executables besides binary programs, such as shell scripts, might be contained in a ``bin directory.''

CGI or Common Gateway Interface
A specification for software that mediates between a Web server and the outside world. Programs written to this specification are called ``gateways''. Commonly used to process forms data, perhaps returning a page to the browser.

child process
Every Unix process is created from an existing process. The new process is called a ``child'' process of the process that created it. Thus, every process is a child of some process.

CORBA (Common Object Request Broker Architecture)
The CORBA standard introduced by the Object Management Group (OMG) defines a means by which distributed objects on different platforms -- created using different programming languages -- can communicate with one another via method invocations.

core files
A disk file created from the memory image of a process that ran into an error conditions that prevents the process from continuing execution. You can often use a debugger program to examine the core file to help debug the defective process.

crypt command
The crypt command encrypts and decrypts a file using a key supplied by the user. Many vendors have removed the crypt command because the U.S. government will not allow it to be exported. See the crypt manual page for more information

command-search-path
The list of directories the shell searches to locate commands. The list is made available by its definition with the PATH environment variable. Each directory is separated from the next by a colon character, as shown in:

$ echo $PATH
/bin:/usr/bin:.
$ []

cc:Mail
cc:Mail is popular email package from Lotus, Inc. with extensive dial-up capabilities.

$HOME/.cshrc File
$HOME/.cshrc is a file that is interpreted by the C shell whenever it is invoked. Commonly, aliases are defined in this file so they are available for any invoked shell.

encryption key
Encryption algorithms are used to convert plain text (what you can read) into encrypted text (what you can't read). These algorithms employ a key used to reverse the encryption.

daemon
A daemon is a Unix system program that runs in the background without a controlling terminal. Such program are generally started during bootstrap to perform system-specific functions.

defunct process
See zombie process

DNS
The Domain Name System provides the mapping between human-understandable names and the machine-understandable IP addresses.

$HOME/.forward file
When electronic mail is received on many -- especially BSD- derived -- Unix systems, the mail delivery command looks for a file named .forward in the home directory for the recipient. If one is found, the mail is forwarded to any account names listed in the file. You can also place a command in the .forward file, in which case, the mail message will be piped to that command. The account originally receiving the mail will not get a copy unless that account's name is also listed in the .forward file. In that case you'll need to ``escape'' the account name -- with a backslash (\) -- to avoid a forwarding loop.

Dynamic Host Configuration Protocol
DHCP is a TCP/IP protocol that enables hosts, such as a diskless workstation, to obtain IP addresses from an address pool maintained by a central server.

FAT
File Allocation Table is a term used in describing the DOS file system. FAT file systems use an ``8.3'' naming system, eight-character file names with optional three-character extensions. A period is used to separate the name from the extension. FAT file systems use drive letters (A-Z) to identify hard drive partitions, floppy drives, network-mounted devices and other random access storage. Each letter is the root of an independent file hierarchy. Fully-qualified FAT filenames use a colon (``:'') following the drive letter and a backslash (``\'') between directory names. For instance, c:\word\docs\sales.doc is an example of a fully-qualified name.

ferrite
A mixture of ferric oxide and another divalent oxide, like zinc, manganese, copper, or cobalt, pressed-formed into a ceramic-like material that has high magnetic permeability.

File matching metacharacters
Often, files with similar names are located in the same directory. To make it easier to reference these files as a group, UNIX shells recognize file-matching metacharacters: the asterisk matches zero or more characters, the question mark any single character, and square brackets can enclose a set of characters to be matched. Here are some examples of file name specifications:

All files whose names start with a ``F'':

F*

All files whose name begin with ``F'' but end with ``l'':

F*l

All files whose name begins with ``abc'' followed by a single character:

abc?

All files whose name begins with ``abc'' followed by either a ``1'', ``2'', or ``3'':

abc[123]

FQDN (Fully Qualified Domain Name)
The full name of a host, which consists of its local host name and its domain name. For instance, if ``engr'' is the host name, and ``sun.com'' the domain name, the FQDN is ``engr.sun.com''. A FQDN must be unique if the host is connected to the Internet.

GIF
GIF, which stands for Graphics Interchange Format, is a graphics file format that was created to facilitate image transfer and storage for use by CompuServe and its customers. However, now it's in wide use and supported by most graphics-related programs. This format is employed universally on the World Wide Web because it can be displayed as an inline image by most graphical browsers.

Inode
An inode represents most of the information known about a UNIX file: its permissions, the individual owner and group owners, file modification, file access, and inode change times, number of links, size in characters, file type, and the addresses (disk block numbers) that actually hold that data for this file. There are two things an inode does not contain: the file's name or the directory path where this file is stored. The reason for this is that a file can have more than one name and be in more than on directory on a UNIX file system. Inodes are numbered, starting with two (which is reserved for the file system root directory inode). A directory entry is a name, inode pair, which is called a link. A directory link says that a given name is represented by the listed inode.

InterNIC
The InterNIC is a collaborative project of three organizations that work together to offer the Internet community a full scope of network information services. These services include providing information about accessing and using the Internet, assistance in locating resources on the network, and registering network components for Internet connectivity. The overall goal of the InterNIC is to make networking and networked information more easily accessible to researchers, educators, and the general public. (This paragraph obtained from InterNIC's home page).

IRC or Internet Relay Chat
An interactive, realtime dialogue over the Internet among two or more users, in which each participant types in contributions to the conversation a line at a time, and sees other participants' remarks as they are typed in. IRC looks very similar to ``chat'' channels on bulletin board systems, but with IRC the participants need not be logged into the same system. The term may also refer to the protocol by which these conversations are carried on over the Internet.

ISDN
Integrated Services Digital Network

JPEG
JPEG, which stands for Joint Photographic Experts Group, refers to a standards organization, a compression technique, and a file format. It offers excellent compression of deep-pixel images and is widely used for photographic images on the World Wide Web.

4/100/400 rule
Our Gregorian Calendar dictates that every four years is a leap year. Unfortunately, this simple rule overestimates the number of actual leap years by three every 385 years. To compensate, the Calendar says that a year ending in ``00'' is only a leap year if it is divisible by ``400.''

long-directory-listing
The listing style produced by commands, such as ls -l or tar tvf. Here's an example:

 
$ ls -l
total 1731
drwxr-sr-x 2 root 512 Mar 11 06:26 Mail/
-rw-r--r-- 1 root 707 Mar 24 11:59 appt.list
lrwxrwxrwx 1 root 7 May 20 1991 bin -> usr/bin/
-rwxr-xr-x 1 root 531 Jul 13 1991 bkupfilelist*
-r--r--r-- 1 root 110880 May 20 1991 boot
drwxr-sr-x 2 root 512 Nov 1 07:17 cdrom/
drwxr-sr-x 2 root 512 Sep 7 1994 core/
-rw------- 1 root 24301 Sep 25 15:50 dead.letter
drwxr-sr-x 2 bin 8192 Mar 24 04:46 dev/
drwxr-sr-x 10 bin 2560 Mar 24 04:46 etc/
drwxr-sr-x 14 bin 512 Feb 20 16:47 home/
-rw-r--r-- 1 root 61 Nov 26 06:11 li.fmt
lrwxrwxrwx 1 root 7 May 20 1991 lib -> usr/lib/
drwxr-xr-x 2 root 8192 May 20 1991 lost+found/
-rw-r--r-- 1 root 110529 Mar 7 06:48 mbox
drwxr-sr-x 2 bin 512 Oct 12 1990 mnt/
drwxr-sr-x 2 root 512 Jul 14 1991 package/
drwxr-sr-x 2 root 512 Sep 4 1994 pcfs/
drwxr-sr-x 2 root 512 Jul 14 1991 remainder/
-rw-r----- 1 root 4739 Sep 20 1994 rolodex
drwxr-sr-x 2 bin 512 May 20 1991 sbin/
lrwxrwxrwx 1 root 13 May 20 1991 sys -> ./usr/kvm/sys/
drwxr-sr-x 2 root 512 Jul 5 1991 testnfs.rec/
drwxrwsrwt 3 bin 512 Mar 24 13:00 tmp/
drwxrwsr-x 26 bin 1024 Nov 27 09:03 usr/
drwxrwxrwx 9 root 512 Oct 4 14:16 var/
-rwxr-xr-x 1 root 1448313 Apr 13 1994 vmunix*
$ []

mailing list
In computing, a list of e-mail addresses on a server that is programmed to forward incoming messages to a particular e-mail address on that server to all the addresses in the list. The incoming address is called the list reflector address. A mailing list may be unmoderated -- automatically forwarding all incoming messages (or all messages from persons who are authorized to mail to the list) -- or moderated -- no messages are forwarded until they have been approved by the person appointed as the list moderator.

Mounted over the network
Computer disk storage often comes from several different physical disk storage devices connected to the system. On many operating systems, such as DOS and MacOS, the contents of a disk are identified by a drive designation, such as C: on a DOS file system. For UNIX systems, all the disk storage appears as a single file hierarchy that starts at the root directory. To create this hierarchy, file systems are mounted on it at a specific directory. With the advent of networks, disk partitions stored on the server can be mounted on client machines ``over the network.''

NNTP
NNTP, or Network News Transfer Protocol, is used to transfer news articles over the Internet.

padding

1. A method of delaying transmission to an unbuffered peripheral for a few milliseconds after sending certain instructions, to allow the peripheral time to complete tasks that briefly make it unable to receive more characters. The technique is to send a string of no-op characters that the peripheral has been set to ignore. The usual throwaway characters are ASCII null (NUL) or delete (DEL).
2. Insertion of no-meaning characters, such as spaces or nonsignificant zeroes, before or after the meaningful characters of a string or other value, in order to expand it to the standard length required by a particular situation.

parent process
Every Unix process is started from an existing process. The creating process is called the ``parent'' process. Every process has a parent process, except the kernel-invoked init process, which is its own parent.

Perl
Perl, which was developed in the late 1980s by Larry Wall, is a powerful language for manipulating text, files, processes, and communicating over TCP/IP networks. In fact its name stands for Practical Extraction and Report Language, also known more affectionately as the Pathologically Eclectic Rubbish Lister.

Perl provides a more concise way to perform most tasks formerly done with the C language. Also, Perl programs are more portable and efficient that those written with the standard combination of the shell, Awk, Sed, and other UNIX utilities.

Perl is a freely-distributable (free) program that has been ported to most UNIX versions, though it isn't part of many standard UNIX distributions.

Perl is a sophisticated, yet quirky programming language that handles searching, editing, text and numeric data manipulation, provides flexible formatting capabilities, and includes more advanced programming features, such as access to many UNIX system and networking function calls.

In my (Becca Thomas) opinion, the best print publication for learning Perl is named (appropriately enought) "Learning Perl" and is written by Randal L. Schwartz, co-author of the "Programming Perl" book. "Learning Perl" is available from O'Reilly and Associates publishers, ISBN 1-56592-042-2).

You may wish to check out University of Florida, CIS, Perl Archive World Wide Web page, which has links to The Hypertext Perl Man Page provided by rgs@cs.cmu.edu, The Frequently Asked Questions List, or the out of date Hypertext Very slow cross-Atlantic link to Nexor Searchable Index, comp.lang.perl: The Perl Usenet Newsgroup, The Perl FTP Archive courtesy of the University of Florida Computer and Information Sciences, examples from both "Programming Perl" and "Learning Perl", Quotes by the Perl God, Larry Wall, Just Another Perl Hacker : A collection of japhs from Randal Schwartz, What the comp.archives catalog of free compilers and interpreters has To say about perl, and the 99.44% pure Perl Purity Test. There Are also links to these Perl sites: Metronet Perl5 Page, Metronet Gopher Server, NEXOR Ltd Perl Page, A Wais index of comp.lang.perl, and an Index of Perl/HTML archives.

PPP
Point-to-Point Protocol

RFC
The Request For Comment document series describes--among other things--Internet protocols. InterNIC provides both a hypertext index to RFCs and anonymous FTP archive.

RIP
Routing Information Protocol, an Interior Gateway Protocol (IGP) supplied with the Berkeley Unix system, is derived from Xerox Network Standard (XNS). RIP was originally distributed with no formal standard, but was subsequently defined by RFC 1058 (90K text file). RIP II, a much more advanced, but less popular version was introduced in 1993 and is described in RFC 1387 (6K text file), RFC 1388 (16K text file), and RFC 1389 (24K text file).

RPM
RPM is shorthand for ``Red Hat Package Manager''. With it, a program can be compiled into a binary package that can be installed or removed from a Red Hat Linux system very quickly. It also makes upgrading the Linux system software much less painful.

To review the capabilities and features of RPM see the Linux Installation Guide.

open file
A process must open a file before it can be referenced by that process. The operating system populates various kernel-resident data structures with information about the open file and hands the process a file descriptor, which is used by the process to reference the open file.

OSPF
Open Shortest Path First, one of the more advanced dynamic routing protocols, is an Interior Gateway Protocol based on the technique of link-state routing. It has far superior performance and scalability in medium to larger size networks compared to RIP. OSPF is defined by RFC 1131 (781K PostScript file).

pipe
The pipe is a mechanism that allows data streams to be passed between reading and writing processes.

shell script
A shell script is a file containing Unix command lines and shell programming constructs that can be executed by typing the name of this command file on a shell command line. The user must have both read and execute permissions for this file to run it.

SLIP
Serial Line Internet Protocol

SMTP
Simple Mail Transfer Protocol used for Internet e-mail transfer. Defined by RFC 821 (117K text file), with associated message-file format in RFC 822 (103K text file).

tcpdump
Source for the Tcpdump package is available from Lawrence Berkeley Laboratories using anonymous FTP.

UNICOS
Cray computer's UNIX operating system.

UUCP
Unix-to-Unix-Copy-Program

uuencode
The standard Unix mail system can only handle ASCII character text. If you need to send some non-ASCII text through the mail system, you must somehow change it to contain only ASCII characters. The UNIX command uuencode does this. See the uuencode manual pages for more details.

xterm
Xterm is an X Window System terminal emulator program supplied with the X distribution. It opens a window that gives you access to the shell to run applications.

zombie process
A process that has terminated but its entry remains in the kernel-resident process table because its parent process has not sent the proper exit code. Although, they don't consume system resources they could cause the process table to fill up.