Research challenges for decentralized data

The Solid ecosystem decentralizes
by decoupling data from applications.

You can choose where you store
every single piece of data you produce.

You can grant apps and people access
to very specific parts of your data.

Separating app and storage competition
drives permissionless innovation.

The traditional way of building apps
does not work well with decentralization.

Building apps over decentralized data
requires different app techniques.

The Semantic Web research community
lacks sufficient outward focus.

• Works in our basement has always been
  an acceptable narrative.
• When bringing research to the Web,
  many theoretical problems interfere all at once.
• Controlled experimentation does not excuse
  the lack of Web experimentation.

The Semantic Web identity crisis: in search of the trivialities that never were
by Ruben Verborgh and Miel Vander Sande (submitted)

Are we solving all the problems
we should be solving?

Which challenges does FAIR create
within decentralized networks?

Findable

discovery

query

search

…

Accessible

access control

identity

proof

…

Interoperable

shapes

transformations

links

…

Reusable

licensing

provenance

feedback

…

Querying data in
decentralized networks

• How can we query across personal data stores,
  taking into account privacy?
• How can nodes in a decentralized network
  help each other with caching and querying?
• How can we conceal latency in applications
  that need data from decentralized sources?
  • Can we approximate and incrementally improve results?

Realizing interoperability
without centralized agreement

• How to deal with disagreement and inconsistency
  on the public Web?
• Are shapes the new ontologies?
  • What is the role of semantics?
  • What is the role of shapes?
• How to reshape data on the fly?

Actor roles in
decentralized networks

• How to distribute tasks
  between clients and servers?
• What roles can and should
  intermediaries take up?
  • How does this interact
    with privacy and trust?

Facilitating Linked Data
application development

• How can we expose Linked Data to developers,
  hiding the complexities of RDF but
  keeping flexibility and unboundedness?
• How to leverage composable data shapes that
  apps bind to, as opposed to custom data models?
• How to read and write Linked Data
  according to a specific shape?