1.2 Lesson 1

Almost every Linux distribution offers a pre-installed set of default applications. Besides those pre-installed applications, a distribution has a package repository with a vast collection of applications available to install through its package manager. Although the various distributions offer roughly the same applications, several different package management systems exist for various distributions. For instance, Debian, Ubuntu and Linux Mint use the dpkg, apt-get and apt tools to install software packages, generally referred as DEB packages. Distributions such as Red Hat, Fedora and CentOS use the rpm, yum and dnf commands instead, which in turn install RPM packages. As the application packaging is different for each distribution family, it is very important to install packages from the correct repository designed to the particular distribution. The end user usually doesn’t have to worry about those details, as the distribution’s package manager will choose the right packages, the required dependencies and future updates. Dependencies are auxiliary packages needed by programs. For example, if a library provides functions for dealing with JPEG images which are used by multiple programs, this library is likely packaged in its own package on which all applications using the library depend.

The commands dpkg and rpm operate on individual package files. In practice, almost all package management tasks are performed by the commands apt-get or apt on systems that use DEB packages or by yum or dnf on systems that use RPM packages. These commands work with catalogues of packages, can download new packages and their dependencies, and check for newer versions of the installed packages.

Suppose you have heard about a command called figlet which prints enlarged text on the terminal and you want to try it. However, you get the following message after executing the command figlet:

That probably means the package is not installed on your system. If your distribution works with DEB packages, you can search its repositories using apt-cache search package_name or apt search package_name. The apt-cache command is used to search for packages and to list information about available packages. The following command looks for any occurrences of the term “figlet” in the package’s names and descriptions:

The search identified a package called figlet that corresponds to the missing command. The installation and removal of a package require special permissions granted only to the system’s administrator: the user named root. On desktop systems, ordinary users can install or remove packages by prepending the command sudo to the installation/removal commands. That will require you to type your password to proceed. For DEB packages, the installation is performed with the command apt-get install package_name or apt install package_name:

At this point the package will be downloaded and installed on the system. Any dependencies that the package eventually needs will also be downloaded and installed:

After the download is finished, all files are copied to the proper locations, any additional configuration will be performed and the command will become available:

In distributions based on RPM packages, searches are performed using yum search package_name or dnf search package_name. Let’s say you want to display some text in a more irreverent way, followed by a cartoonish cow, but you are not sure about the package that can perform that task. As with the DEB packages, the RPM search commands accept descriptive terms:

After finding a suitable package at the repository, it can be installed with yum install package_name or dnf install package_name:

Once again, the desired package and all its possible dependencies will be downloaded and installed:

The command cowsay does exactly what its name implies:

Although they may seem useless, commands figlet and cowsay provide a way to draw the attention of other users to relevant information.

The same commands used to install packages are used to remove them. All the commands accept the remove keyword to uninstall an installed package: apt-get remove package_name or apt remove package_name for DEB packages and yum remove package_name or dnf remove package_name for RPM packages. The sudo command is also needed to perform the removal. For example, to remove the previously installed package figlet from a DEB-based distribution:

After confirming the operation, the package is erased from the system:

A similar procedure is performed on an RPM-based system. For example, to remove the previously installed package cowsay from an RPM-based distribution:

Likewise, a confirmation is requested and the package is erased from the system:

The configuration files of the removed packages are kept on the system, so they can be used again if the package is reinstalled in the future.

Office applications are used for editing files such as texts, presentations, spreadsheets and other formats commonly used in an office environment. These applications are usually organised in collections called office suites.

For a long time, the most used office suite in Linux was the OpenOffice.org suite. OpenOffice.org was an open source version of the StarOffice suite released by Sun Microsystems. A few years later Sun was acquired by Oracle Corporation, which in turn transferred the project to the Apache Foundation and OpenOffice.org was renamed to Apache OpenOffice. Meanwhile, another office suite based on the same source code was released by the Document Foundation, which named it LibreOffice.

The two projects have the same basic features and are compatible with the document formats from Microsoft Office. However, the preferred document format is the Open Document Format, a fully open and ISO standardized file format. The use of ODF files ensures that documents can be transferred between operating systems and applications from different vendors, such as Microsoft Office. The main applications offered by OpenOffice/LibreOffice are:

Both LibreOffice and Apache OpenOffice are open source software, but LibreOffice is licensed under LGPLv3 and Apache OpenOffice is licensed under Apache License 2.0. The licensing distinction implies that LibreOffice can incorporate improvements made by Apache OpenOffice, but Apache OpenOffice cannot incorporate improvements made by LibreOffice. That, and a more active community of developers, are the reason most distributions adopt LibreOffice as their default office suite.

For most users, the main purpose of a computer is to provide access to the Internet. Nowadays, web pages can act as a full featured app, with the advantage of being accessible from anywhere, without the need of installing extra software. That makes the web browser the most important application of the operating system, at least for the average user.

The main web browsers in the Linux environment are Google Chrome and Mozilla Firefox. Chrome is a web browser maintained by Google but is based on the open source browser named Chromium, which can be installed using the distribution’s package manager and is fully compatible with Chrome. Maintained by Mozilla, a non-profit organization, Firefox is a browser whose origins are linked to Netscape, the first popular web browser to adopt the open source model. The Mozilla Foundation is deeply involved with the development of the open standards underlying the modern web.

Mozilla also develops other applications, like the e-mail client Thunderbird. Many users opt to use webmail instead of a dedicated email application, but a client like Thunderbird offers extra features and integrates best with other applications on the desktop.

Compared to the available web applications, desktop applications are still the best option for the creation of multimedia content. Multimedia related activities like video rendering often require high amounts of system resources, which are best managed by a local desktop application. Some of the most popular multimedia applications for the Linux environment and their uses are listed below.

There are also many applications dedicated to multimedia playback. The most popular application for video playback is VLC, but some users prefer other alternatives, like smplayer. Local music playback also has many options, like Audacious, Banshee and Amarok, which can also manage a collection of local sound files.

When a web browser loads a page from a website, it actually connects to a remote computer and asks for a specific piece of information. In that scenario, the computer running the web browser is called the client and the remote computer is called the server.

The server computer, which can be an ordinary desktop computer or specialized hardware, needs a specific program to manage each type of information it will provide. Regarding serving web pages, most servers around the world deploy open source server programs. This particular server program is called an HTTP server (HTTP stands for Hyper Text Transfer Protocol) and the most popular ones are Apache, Nginx and lighttpd.

Even simple web pages may require many requests, which can be ordinary files — called static content — or dynamic content rendered from various sources. The role of an HTTP server is to collect and send all the requested data back to the browser, which then arranges the content as defined by the received HTML document (HTML stands for Hyper Text Markup Language) and other supporting files. Therefore, the rendering of a web page involves operations performed on the server side and operations performed on the client side. Both sides can use custom scripts to accomplish specific tasks. On the HTTP server side, it is quite common to use the PHP scripting language. JavaScript is the scripting language used on the client side (the web browser).

Server programs can provide all kinds of information. It’s not uncommon to have a server program requesting information provided by other server programs. That is the case when an HTTP server requires information provided by a database server.

For instance, when a dynamic page is requested, the HTTP server usually queries a database to collect all the required pieces of information and sends the dynamic content back to the client. In a similar way, when a user registers on a website, the HTTP server gathers the data sent by the client and stores it in a database.

A database is an organized set of information. A database server stores contents in a formatted fashion, making it possible to read, write and link large amounts of data reliably and at great speed. Open source database servers are used in many applications, not only on the Internet. Even local applications can store data by connecting to a local database server. The most common type of database is the relational database, where the data is organized in predefined tables. The most popular open source relational databases are MariaDB (originated from MySQL) and PostgreSQL.

In local networks, like the ones found in offices and homes, it is desirable that computers not only should be able to access the Internet, but also should be able to communicate with each other. Sometimes a computer acts as a server, sometimes the same computer acts as a client. That is necessary when one wants to access files on another computer in the network — for instance, access a file stored on a desktop computer from a portable device — without the hassle of copying it to a USB drive or the like.

Between Linux machines, NFS (Network File System) is often used. The NFS protocol is the standard way to share file systems in networks equipped only with Unix/Linux machines. With NFS, a computer can share one or more of its directories with specific computers on the network, so they can read and write files in these directories. NFS can even be used to share an entire operating system’s directory tree with clients that will use it to boot from. These computers, called thin clients, are mostly often used in large networks to avoid the maintenance of each individual operating system on each machine.

If there are other types of operating systems attached to the network, it is recommended to use a data sharing protocol that can be understood by all of them. This requirement is fulfilled by Samba. Samba implements a protocol for sharing files over the network originally made for the Windows operating system, but today is compatible with all major operating systems. With Samba, computers in the local network not only can share files, but also printers.

On some local networks, the authorization given upon login on a workstation is granted by a central server, called the domain controller, which manages the access to various local and remote resources. The domain controller is a service provided by Microsoft’s Active Directory. Linux workstations can associate with a domain controller by using Samba or an authentication subsystem called SSSD. As of version 4, Samba can also work as a domain controller on heterogeneous networks.

If the goal is to implement a cloud computing solution able to provide various methods of web based data sharing, two alternatives should be considered: ownCloud and Nextcloud. The two projects are very similar because Nextcloud is a spin-off of ownCloud, which is not unusual among open source projects. Such spin-offs are usually called a fork. Both provide the same basic features: file sharing and sync, collaborative workspaces, calendar, contacts and mail, all through desktop, mobile and web interfaces. Nextcloud also provides private audio/video conferencing, whilst ownCloud is more focused on file sharing and integration with third-party software. Many more features are provided as plugins which can be activated later as needed.

Both ownCloud and Nextcloud offer a paid version with extra features and extended support. What makes them different from other commercial solutions is the ability to install Nextcloud or ownCloud on a private server, free of charge, avoiding keeping sensitive data on an unknown server. As all the services depend on HTTP communication and are written in PHP, the installation must be performed on a previous configured web server, like Apache. If you consider installing ownCloud or Nextcloud on your own server, make sure to also enable HTTPS to encrypt all connections to your cloud.

Communication between computers is only possible if the network is working correctly. Normally, the network configuration is done by a set of programs running on the router, responsible for setting up and checking the network availability. In order to achieve this, two basic network services are used: DHCP (Dynamic Host Configuration Protocol) and DNS (Domain Name System).

DHCP is responsible for assigning an IP address to the host when a network cable is connected or when the device enters a wireless network. When connecting to the Internet, the ISP’s DHCP server will provide an IP address to the requesting device. A DHCP server is very useful in local area networks also, to automatically provide IP addresses to all connected devices. If DHCP is not configured or if it’s not working properly, it would be necessary to manually configure the IP address of each device connected to the network. It is not practical to manually set the IP addresses on large networks or even in small networks, that’s why most network routers come with a DHCP server pre-configured by default.

The IP address is required to communicate with another device on an IP network, but domain names like www.lpi.org are much more likely to be remembered than an IP number like 203.0.113.165. The domain name by itself, however, is not enough to establish the communication through the network. That is why the domain name needs to be translated to an IP address by a DNS server. The IP address of the DNS server is provided by the ISP’s DHCP server and it’s used by all connected systems to translate domain names to IP addresses.

Both DHCP and DNS settings can be modified by entering the web interface provided by the router. For instance, it is possible to restrict the IP assignment only to known devices or associate a fixed IP address to specific machines. It’s also possible to change the default DNS server provided by the ISP. Some third-party DNS servers, like the ones provided by Google or OpenDNS, can sometimes provide faster responses and extra features.

All computer programs (client and server programs, desktop applications and the operating system itself) are made using one or more programming languages. Programs can be a single file or a complex system of hundreds of files, which the operating system treats as an instruction sequence to be interpreted and performed by the processor and other devices.

There are numerous programming languages for very different purposes and Linux systems provide a lot of them. Since open source software also includes the sources of the programs, Linux systems offer developers perfect conditions to understand, modify or create software according to their own needs.

Every program begins as a text file, called source code. This source code is written in a more or less human-friendly language that describes what the program is doing. A computer processor can not directly execute this code. In compiled languages, the source code is therefore be converted to a binary file which can then be executed by the computer. A program called compiler is responsible for doing the conversion from source code to executable form. Since the compiled binary is specific to one kind of processor, the program might have to be re-compiled to run on another type of computer.

In interpreted languages, the program does not need to be previously compiled. Instead, an interpreter reads the source code and executes its instruction every time the program is run. This makes the development easier and faster, but at the same time interpreted programs tend to be slower than compiled programs.

Here some of the most popular programming languages:

C and Java are compiled languages. In order to be executed by the system, source code written in C is converted to binary machine code, whereas Java source code is converted to bytecode executed in a special software environment called Java Virtual Machine. JavaScript, Perl, Shell script, Python and PHP are all interpreted languages, which are also called scripting languages.