Design issues
The open questions any system for publishing and linking information has to answer, named after Tim Berners-Lee's Design Issues notes. Each is a neutral question, not a solution; every work is filed under the one it mainly addresses. They fall into technical questions, which a specification or better engineering could settle, and human ones, which persist even with a perfect specification because they depend on behaviour, incentives or power. People and organizations are listed under the issues of the works they are linked to.
Technical
Questions a specification or better engineering could settle: what the data is, and how it is handled.
Representation How things are named, connected, shaped and written down.
- Identification 51 works What is a thing called? Names that locate (URLs) against names that stay stable (content hashes, DOIs, DIDs); whether a web address names a page or the thing it describes (httpRange-14). from Barcode (1949), Satellite Catalog Number (1957), International Designator (1963)
- Linking 31 works How do things connect? Easy, one-way, breakable links (the Web) against two-way, unbreakable ones (Xanadu). from As We May Think (1945), Memex (1945), Project Xanadu (1960)
- Data model 37 works What shape is the data? Flexible graphs against strict structure: RDF against property graphs against tables. from Semantic network (1956), Canada Geographic Information System (1963), Semantic Memory (1968)
- Serialization 17 works How is it written down? Formal precision against developer familiarity: RDF/XML against JSON-LD. from MARC (1965), Two-line element set (1980), Text Encoding Initiative (1987)
Computation What can be worked out from the data and asked of it, and how fast.
- Reasoning 27 works What follows from it? Expressiveness against tractability: richer logics say more but cost more to compute, or cannot be computed at all. from Programs with Common Sense (1959), Semantic Information Processing (1968), Tobler's first law of geography (1970)
- Query & validation 18 works How do we ask and check? Open world against closed world: a missing fact is unknown, or it is false. from On Closed World Data Bases (1978), SPARQL (2008), Semantics and Complexity of SPARQL (2009)
- Scale & performance 9 works Does it work at the size needed? Completeness against speed: storing, indexing and answering over billions of statements, and queries that reach across many sources at once. from Virtuoso (1998), MapReduce: Simplified Data Processing on Large Clusters (2004), Apache Hadoop (2006)
Bridge
Technical machinery for a social agreement: the languages for defining terms can be specified, but the terms themselves have to be agreed.
Human
Questions that persist even with a perfect specification, because they depend on behaviour, incentives or power: who, not how.
Institutional Who hosts it, who vouches for it, and who decides.
- Topology 17 works Where does data live, and who hosts it? Centralized convenience against decentralized control, after Paul Baran's 1964 distinction between centralized, decentralized and distributed networks. from On Distributed Communications: Introduction to Distributed Communications Networks (1964), Thoughts on the Social Graph (2007), Long Live the Web (2010)
- Trust 14 works Who vouches for it? Who counts as an authority; verification against usability. from Pretty Good Privacy (1991), Web of trust (1992), PICS (1996)
- Governance 40 works Who decides, who may use it, and on what terms? Openness against control: who sets the standard, who owns the data, and the licences that say what others may do with it. from UN Register of Objects Launched into Outer Space (1962), CelesTrak (1985), CDDB (1993)
Economic Whether running it makes sense, commercially or as a public good: who uses it, who pays, and what it costs people to work with.
- Adoption & incentives 5 works Why would anyone publish or use it? Effort now against payoff later: who bears the cost of adding structure, who gets the benefit, and the chicken-and-egg of data without applications and applications without data. from Metacrap: Putting the torch to seven straw-men of the meta-utopia (2001), Linked Open Data cloud diagram (2007), Rich Snippets (2009)
- Sustainability 9 works Who pays for it, and for how long? Launch against upkeep: projects, public endpoints and vocabularies that outlive, or do not outlive, the grant or the company that paid for them. from Kessler syndrome (1978), Collision Frequency of Artificial Satellites: The Creation of a Debris Belt (1978), PURL (1995)
- Usability 7 works Can ordinary developers and people use it? Rigour against approachability: what a specification makes possible against what a working developer will actually pick up. from HyperCard (1987), JSON-LD and Why I Hate the Semantic Web (2014), Linked Data Platform (2015)
Spanning all issues (11)
Concepts 3
-
Semantic Web
1998
concept
vision
Tim Berners-Lee's vision of a Web whose data carries machine-readable meaning, built on standards such as RDF and OWL.
-
Semantic Web layer cake
2000
concept
★ seminal
vision
Tim Berners-Lee's diagram of the Semantic Web as layers, first shown in his XML 2000 keynote in December 2000: Unicode and URIs at the base, then XML, RDF and RDF Schema, an ontology vocabulary, logic, proof and trust, with digital signatures alongside. W3C redrew it as standards arrived, adding query and rules.
-
Software agent
concept
vision
Program that acts on a person's behalf, for example finding and booking a service; the payoff promised by the 2001 Scientific American article.
Writings 7
-
Semantic Web Road Map
1998
writing
★ seminal
vision
Tim Berners-Lee's 1998 Design Issues note giving a high-level plan for the architecture of the Semantic Web.
-
Weaving the Web
1999
writing
★ seminal
vision
Tim Berners-Lee's account of how he created the World Wide Web, ending with his plans for a Semantic Web.
-
The Semantic Web
2001
writing
★ seminal
vision
Scientific American article by Berners-Lee, Hendler and Lassila describing software agents that act on machine-readable meaning published on the Web.
-
The Semantic Web Revisited
2006
writing
★ seminal
debate
IEEE Intelligent Systems article in which Shadbolt, Hall and Berners-Lee reassessed the Semantic Web five years after the 2001 article.
-
Where Are All the Intelligent Agents?
2007
writing
debate
Jim Hendler's 2007 editorial in IEEE Intelligent Systems, six years after he co-wrote the Scientific American article that promised Semantic Web agents, asking why those agents had not appeared.
-
A Programmable Web: An Unfinished Work
2013
writing
vision
Aaron Swartz's unfinished short book on Semantic Web ideas, published after his death in 2013.
-
W3C and the State of the Web in the AI Era
2026
writing
vision
October 2026 paper by the W3C Strategy Team, prepared by Philippe Le Hégaret, on how the Web's standards must change as AI agents act for users: agents prefer text and structured data to HTML, JSON-LD, RDF, DCAT and ODRL become more valuable as inputs and guardrails for agents, and Verifiable Credentials and DIDs can identify agents and whom they act for.
Projects 1
-
International Semantic Web Conference
2002
project
Annual research conference of the Semantic Web community, held since 2002 after a 2001 symposium at Stanford.
See also the map of issues × traditions: which communities answered which questions, and when.
Other facets: Kinds of idea · Traditions · published as SKOS in graph.jsonld