Problem threads

The 28 long-running problems the catalogue follows. Each thread tells one problem as a lineage of works, oldest first: who posed it, what was proposed, who objected, what changed, what is still open, and where work on it happens today.

Grouped by where each problem stands now: open (21), reopened (3), largely settled (4), abandoned (0). Within each, threads are grouped by the design issue group they mainly follow (technical, bridge or human), and each card names its issue.

open 21 threads

No accepted answer yet; work on it continues.

Across all issues 1

Agents acting for us 1998–2026 all issues 7 steps
Will software agents read the Web's data and act on people's behalf, and what do they need from it?

Technical 8

Names that outlive their hosts 1994–2022 Identification 9 steps
How can a name keep working after the server, the company or the domain behind it is gone?
Same thing, many names 2004–2026 Identification 6 steps
When two datasets name the same thing, how do we say so without claiming more than we mean?
Who issues your name online? 2000–2026 Identification 11 steps
Should people be known online by accounts that platforms issue, or by names they control themselves?
Who runs the registry? 1957–2026 Identification 10 steps
When everyone depends on one set of identifiers, who issues them, on what terms, and what happens when the registry runs out, changes hands or closes its code?
Can readers write back? 1960–2018 Linking 7 steps
Can readers attach comments, corrections and links to someone else's page, where other readers will see them?
Graphs, tables and property graphs 1968–2026 Data model 11 steps
What is the right shape for connected data: tables, triples, or graphs whose nodes and edges carry properties?
Hand-built meaning or learned meaning 1959–2024 Reasoning 11 steps
Do machines get meaning from knowledge people write down, or from patterns in large amounts of data?
Querying a web, not a database 2008–2025 Scale & performance 7 steps
Can you query data spread across thousands of independent servers as if it were one database?

Bridge 3

From catalogue card to web data 1876–2024 Vocabulary 11 steps
Can library catalogue records, built for card drawers and MARC, become first-class data on the Web?
One vocabulary or many 1984–2016 Vocabulary 9 steps
Should everyone share one ontology, reuse a few common vocabularies, or map between many local ones?
Ontology against folksonomy 1876–2018 Vocabulary 11 steps
Should shared terms be designed by experts in advance, or emerge from the words people actually use?

Human 9

Central against decentralized 1964–2026 Topology 12 steps
Where should data live: with a few large hosts, or with the people and groups it belongs to?
Who vouches for it? 1991–2022 Trust 9 steps
How can a reader tell whether a claim on the Web is true, and whose word to take for it?
Can the Web forget? 1998–2019 Governance 8 steps
Should what is published about a person stay findable forever, or can they have it removed?
Who controls data about me? 2002–2026 Governance 8 steps
Can people decide what happens to data about them, and can machines read and respect those decisions?
Who owns what the crowd built? 1998–2025 Governance 8 steps
When volunteers build a database together, what stops whoever hosts it from closing it, selling it or shutting it down?
Who sets the standard? 2001–2024 Governance 8 steps
Who gets to decide the Web's data standards: a consortium, a grassroots community, or the largest consumers of the data?
Why would anyone add structure? 1998–2026 Adoption & incentives 12 steps
Who bears the cost of publishing structured data, and what makes it worth their while?
Who pays for public data infrastructure? 1995–2025 Sustainability 9 steps
Who pays to keep shared identifiers, query services and knowledge bases running after the launch money is spent?
Can ordinary developers use it? 1987–2024 Usability 12 steps
Can people who are not specialists build with linked data, or does it stay a tool for experts?

reopened 3 threads

Settled once, as a step in the thread shows, then brought back by new technology or new needs.

Technical 2

One-way or two-way links 1945–2018 Linking 7 steps
Should a link be a cheap one-way pointer that can break, or a maintained two-way relationship?
Expressiveness against tractability 1975–2023 Reasoning 10 steps
How much can a knowledge language say before reasoning with it becomes too slow, or impossible to complete?

Human 1

On what terms may anyone reuse it? 2005–2026 Governance 10 steps
What licence, if any, should published data carry, and who gets to set the terms of its reuse?

largely settled 4 threads

An answer most of the field now accepts, though objections remain.

Technical 4

Page or thing? 1998–2012 Identification 9 steps
When you look up an http URI, have you got the thing it names, or a document about that thing?
Statements about statements 1999–2026 Data model 7 steps
How do you say who made a statement, when, and how sure they are, in a data model built from single statements?
Rigour against developer taste 1998–2015 Serialization 10 steps
What does a data format have to look like before ordinary web developers will use it?
Unknown or false? 1978–2026 Query & validation 8 steps
When a fact is missing, is it unknown or false, and how do you check data in a world that assumes it is merely unknown?

abandoned 0 threads

Dropped without an answer.

None yet.

Step roles

What a work does in a thread. The same work can play different roles in different threads.

precedent: Comes before the question but shapes it: an earlier idea the thread inherits.
poses: States the problem for the first time, or sharply enough that later work starts from it.
proposes: Offers an answer to the problem.
critiques: Argues that an earlier answer is wrong or incomplete.
responds: Answers a critique or an earlier answer.
workaround: Sidesteps the problem in practice without resolving it.
settles: Becomes the accepted answer, for now.
reopens: Brings a settled question back.
measures: Gives evidence on how things actually stand.