Infuse
For a better scientific conversation
Sébastien Lerique
Web
Improvements, isolated
Symptoms of
lack of standards
Annotations
Web
Bibliographic search
- No citation-level search
- Unknown keyword problem
List the papers which the qualified person for this query would give me
Initiatives to integrate
- Notebooks as publications
/ 
- Blog posts as publications
- Post-publication peer review
- Karma-based interactions (à la
) - Wiki pages as publications
- Wiki-like fluid authorship
Enclosure of knowledge
Timothy Gowers on his blog:
(...) something is deeply wrong with the current system of academic publishing. The basic point, which has been made innumerable times by innumerable people, is that the really hard parts — the writing of papers, and the peer review and selection of the ones to publish — are done voluntarily by academics, and modern technology makes things like typesetting and dissemination extremely cheap. And yet publishers are making more money than ever before.
2014 Elsevier costs:
| University | Cost | Enrolment |
|---|---|---|
| Bristol | £808,840 | 19,220 |
| Cambridge | £1,161,571 | 19,945 |
| Edinburgh | £845,000 | 31,323 |
| ICL | £1,340,213 | 16,000 |
| Leeds | £847,429 | 32,510 |
| Manchester | £1,257,407 | 40,860 |
| Newcastle | £974,930 | 21,055 |
| Nottingham | £903,076 | 35,630 |
| Oxford | £990,775 | 25,595 |
| UCL | £1,381,380 | 25,525 |
Symptoms of
lack of governance
Cross-community barriers
The unknown keywords problem
Large minimal contributions
Disincentives to cross-community work
Content incentives
Infuse
Wiki-like article reader
Powered by background PDF processing
Results stored in the permanent distributed web 
Catalyse cognitive support for ideation
PDF processing
Constraints
- Distributed processing
- Grow a community
Grobid
- Java → JS using
... ✈💥 - Performance and optimisation of generated JS
- Messy code ... and just Java
Rust 
- Amazing community & process
- Rust → Wasm momentum
- Process more difficult PDFs
Goals
Lift a roadblock for other initiatives to federate
Make reading research nicer from day 1
Bring all publications into the permanent web
- Broader discussion on metrics
- Open competition for search
- Shared annotations everywhere
- Relax conversation formats