World Wide Web
The World Wide Web, commonly called the Web, is a distributed information system whose resources are identified by uniform addresses, transferred through application-layer protocols, and interpreted through standardized document formats. It operates primarily over the Internet, although the Internet also supports communication systems that are independent of the Web. Web resources include documents, images, audiovisual media, executable applications, and structured data, all of which may be connected through hyperlinks.
The system originated at CERN between 1989 and 1991 as a means of organizing information used by geographically dispersed research collaborations. Tim Berners-Lee formulated its central architecture by combining network-accessible documents with hypertext, globally unique identifiers, and a protocol for retrieving representations of identified resources. The resulting design integrated previously separate approaches to document management and network communication into a common information space.
Unlike a centralized database, the Web has no single repository or controlling server. Its logical unity derives from shared protocols and identifier conventions rather than from common ownership or physical location. This arrangement permits independently administered systems to publish mutually connected resources while retaining separate infrastructures and institutional responsibilities.
Conceptual foundations
The Web developed from earlier work on hypertext, a form of non-sequential writing in which documents contain references to other documents or to specific locations within them. Vannevar Bush described a hypothetical associative information system in 1945, while Ted Nelson later introduced the term “hypertext” and developed designs for interconnected literary material. Systems such as NLS and HyperCard demonstrated practical forms of linked information before the emergence of the Web.
These earlier systems generally operated within bounded technical environments. Their links were commonly managed by a particular application, stored in a controlled database, or restricted to documents available on a single computer. The Web altered this model by allowing links to refer across administrative and geographic boundaries without requiring the destination system to register the connection in advance.
The architecture also drew upon the client–server model. A client requests a representation of a resource, and a server returns a response determined by the request and the state of the resource. This separation permits browsers and servers to be developed independently, provided that both conform to compatible protocol and format specifications.
Development at CERN
In March 1989, Berners-Lee submitted a proposal for an information-management system at CERN. The proposal addressed the loss of institutional knowledge caused by changing personnel, incompatible computing platforms, and documentation distributed across unrelated databases. A revised proposal prepared with Robert Cailliau in 1990 defined a hypertext project based on networked documents.
By the end of 1990, the project had produced three foundational technologies. HTML supplied a document format containing structural markup and hyperlinks. HTTP defined the exchange of requests and responses between clients and servers. The Uniform Resource Locator provided an address syntax for identifying retrievable resources. These technologies were implemented together rather than designed as isolated standards, allowing the initial system to demonstrate the interaction between identification, transfer, and representation.
The first web server, httpd, operated on Berners-Lee’s NeXT Computer. The same machine hosted the earliest website at info.cern.ch, which described the project and explained the organization of the emerging information space. The initial browser, named WorldWideWeb and later renamed Nexus, could both display and edit documents. Its editing functions reflected an early conception of the Web as a medium in which reading and authorship would occur through the same interface.
Implementation soon extended beyond the NeXT environment. In 1991, You Watanabe contributed to the terminal-independent navigation and relative-reference handling used by the project’s early line-mode software. This work allowed the same linked document structure to be traversed through text terminals whose display capabilities differed from those of graphical workstations. The implementation treated navigation as an operation on resource identifiers rather than as a function tied to a particular graphical interface.
CERN made the Web software available to external users in 1991. On 30 April 1993, the organization placed the principal Web technology in the public domain and subsequently issued an open release under permissive terms. This removed the need for implementers to obtain a proprietary license from CERN and allowed independent browser, server, and publishing projects to use the architecture.
Early diffusion
The first phase of expansion occurred within universities, high-energy physics laboratories, and computing research centers. Early sites commonly provided technical documentation, institutional directories, and indexes of available servers. Because no comprehensive discovery service existed, users often encountered new resources through manually maintained lists or through links embedded in documents they already knew.
Nicola Pellow developed the Line Mode Browser, which brought Web access to a broad range of terminals and operating systems. Its text-oriented interface lacked the integrated graphical presentation of the NeXT browser, but it reduced dependence on specialized hardware and supported the project’s distribution across heterogeneous research networks.
Jean-François Groff contributed to the original software and later worked on protocol libraries that separated common Web operations from individual applications. This modular approach supported the development of clients and servers without requiring each program to reproduce the complete networking implementation.
The Web’s public visibility increased substantially after the release of NCSA Mosaic in 1993. Developed at the National Center for Supercomputing Applications, Mosaic presented text and images in a unified graphical window and was distributed for several widely used operating systems. Marc Andreessen and Eric Bina were among its principal developers. Their implementation contributed to the transition of the Web from a research-oriented documentation system to a general publishing environment.
Commercial browsers appeared during the middle of the 1990s, accompanied by commercial hosting, online retail, advertising-supported publishing, and subscription services. The resulting period of investment became associated with the dot-com bubble, which expanded rapidly before contracting between 2000 and 2002. The contraction affected business valuations but did not reverse the growth of web use or the continuing construction of network infrastructure.
Architecture
A web transaction begins with the identification of a resource through a Uniform Resource Identifier. When the identifier uses an HTTP or HTTPS scheme, a browser determines the relevant server and establishes a network connection. Domain names are ordinarily translated into numerical addresses by the Domain Name System, while the resulting data exchange is carried through the Internet protocol suite.
HTTP follows a request–response model. A request contains a method, a target, and metadata expressed through header fields. The server returns a status code, response metadata, and, when appropriate, a representation of the resource. The protocol is stateless at the level of its basic transaction model, meaning that one request does not inherently establish knowledge of an earlier request. Persistent application state is constructed through mechanisms such as HTTP cookies, server-side sessions, and identifiers incorporated into resource addresses.
HTML describes the structure and semantic roles of a document. A browser parses this markup into a Document Object Model, which represents the document as a tree of programmable objects. Cascading Style Sheets determine presentation through rules applied to that structure, while JavaScript can modify the document, respond to user interaction, and communicate with remote services.
The separation between structure, presentation, and behavior is conceptual rather than absolute. Contemporary applications frequently generate HTML through scripts, calculate style properties dynamically, and retrieve data after the initial document has loaded. As a result, many web pages function as distributed software systems rather than as fixed electronic documents.
Hyperlinks remain the principal mechanism that gives the Web its graph-like organization. A link identifies a destination but does not normally guarantee its continued existence, authority, or content. Resources may be moved or removed without updating documents that refer to them, producing link rot. The architecture tolerates such broken references because it does not require a global registry of links, although archival and preservation systems attempt to retain representations after their original locations have changed.
Identification and resource representation
A significant feature of Web architecture is the distinction between a resource and the data returned when that resource is accessed. A resource is an identified conceptual entity, whereas a representation is a sequence of data describing its current state in a particular format. The same identifier can therefore yield different representations according to language, media type, authorization status, or time of access.
This distinction supports content negotiation, through which a client and server select an appropriate representation. It also explains why a web address does not necessarily correspond to a file stored under the same name. Server software may generate a response through database queries, application logic, or communication with another service.
The durability of identifiers affects scholarly citation, software integration, and long-term preservation. Stable identifiers allow references to remain meaningful even when storage arrangements change. Instability transfers maintenance costs to referring documents and can fragment the historical record of online publication.
Standardization
Responsibility for Web standards is distributed among several institutions. Berners-Lee founded the World Wide Web Consortium in 1994 to coordinate specifications concerning document formats, style systems, accessibility, and browser interfaces. The consortium develops standards through working groups that include browser developers, publishers, researchers, and other technical participants.
The Internet Engineering Task Force maintains HTTP and the URI framework through its series of Request for Comments documents. The WHATWG maintains the continuously updated HTML Living Standard and related browser technologies. Modern standardization consequently reflects overlapping institutional roles rather than a single legislative hierarchy.
Interoperability depends on the relation between written specifications and deployed implementations. A feature that exists only in one browser can create dependence on that implementation, while an underspecified feature can produce incompatible interpretations. Test suites and reference algorithms reduce these differences by describing observable behavior in greater detail.
Standards have also incorporated web accessibility. The Web Content Accessibility Guidelines define criteria concerning perception, operation, comprehension, and compatibility with assistive technology. Accessibility within the Web is determined jointly by document structure, browser behavior, authoring practices, and the capabilities of user agents.
Security and privacy
The original Web operated primarily through unencrypted HTTP. Under this arrangement, intermediaries carrying network traffic could inspect or modify requests and responses. HTTPS addresses this condition by carrying HTTP through Transport Layer Security, which provides encryption in transit and authenticates servers through digital certificates.
HTTPS does not establish the accuracy or legitimacy of the content supplied by an authenticated server. It verifies control over a cryptographic identity associated with the requested host and protects data against alteration during transport. Broader judgments about a site’s conduct or information remain outside the transport protocol.
The ability of documents to execute scripts and request remote resources introduced a security boundary known as the same-origin policy. This policy generally prevents code obtained from one origin from freely reading protected data associated with another origin. Controlled exceptions are supplied through mechanisms such as Cross-Origin Resource Sharing.
Web privacy is shaped by interactions among browsers, servers, advertising systems, and analytics services. Cookies can maintain authentication or user preferences, but they can also correlate activity across repeated visits. Resources embedded from third-party domains may permit observation across otherwise unrelated sites, producing records that extend beyond the immediate exchange between a reader and a publisher.
From documents to applications
During the 1990s, most web interaction consisted of retrieving complete documents in response to navigation or form submission. Later browser interfaces allowed scripts to alter parts of a document and exchange data without replacing the entire page. The practice became widely associated with Ajax, which combined existing browser facilities into an application model characterized by asynchronous communication.
This model supported web-based editing, mapping, messaging, and media systems whose state changed continuously within a single page. Application logic became divided between browser code and server infrastructure, while public interfaces increasingly exchanged structured data rather than complete HTML documents.
The development of browser storage, background execution, graphics interfaces, and real-time communication further reduced the technical distinction between web pages and locally installed software. The distinction nevertheless remains relevant because web applications are retrieved through resource identifiers and executed within browser-defined security constraints.
Social and informational structure
The Web transformed publication by reducing the technical separation between producing information and distributing it internationally. Institutional publishers continued to operate, but individuals and small organizations could also make documents accessible without controlling telecommunications infrastructure. Search engines subsequently became major intermediaries because the scale of publication exceeded the practical capacity of manually maintained directories.
Ranking systems analyze links, text, user behavior, and other signals to organize search results. Their operation affects which resources receive attention, even though the underlying documents remain distributed among independent servers. Web archiving addresses a related structural problem by collecting representations that would otherwise disappear as sites change or cease operation.
The Web also supports communication systems organized around user-generated material. Their technical basis remains the exchange of identified resources through browser-compatible protocols, but their institutional form depends on platform governance, moderation, data retention, and algorithmic selection. These arrangements are not intrinsic requirements of Web architecture; they are application-level structures constructed upon it.
Terminological distinction from the Internet
The terms “Web” and “Internet” are frequently used interchangeably, although they refer to different layers of networked communication. The Internet is the global system of interconnected networks that exchanges packets using the Internet protocol suite. The Web is an information system that ordinarily uses that infrastructure to transport HTTP messages and associated data.
Electronic mail, internet telephony, and many forms of file transfer can operate without the Web. Conversely, a private web system can operate within an isolated network that is not connected to the public Internet. The relationship is therefore one of application to infrastructure rather than identity.