IPFS in GLAM Informal Call
Tuesday 23 May 2017, 10-11 AM Pacific / 11-12 AM Mountain / Noon-1 PM Central / 1-2 PM Eastern / 6-7 PM BST / 7-8 PM CEST
Video of the Call: https://youtu.be/y2FR_F86Oe4
- Moderator/Gatekeeper: Matt Zumwalt
- Notetaker:
- Speaker/Topic: Enno Meijers on his work with MediaChain
How to Join the Meeting
To be added.
Agenda Items:
- Quick Introductions
- Speaker - Enno
- Logistics for ongoing calls
- Announcements or Opportunities for Collaboration
Attendees
- @flyingzumwalt: Protocol Labs
- @mishmosh: Protocol Labs
- Brendan O'Brien: Data Rescue Project - Environmental Data & Governance Initiative
- Christina Harlow: Repository & Data at Stanford
- Ed Silverton: Universal Viewer lead, concentrating on UV 3D and IPFS
- Enno Meijers:
- Funded to make it easier for Netherlands citizens to access cultural heritage collections
Notes
Mediachain (Enno Meijers)
Focuses on making information more useable
- Chose linked data because content coming from many different kinds of orgs
- Don't want to build yet another big aggregator (europeana, DPLA, etc)
- Interested in IPFS and distributed networking to explore different approach
- Conclusion: it's a bit premature to start using this
- starting by building a registry that points to where the data can be found
- giving directions to relevant info in verious service portals
- looked into options for using IPFS with linked data
- worked with mediachain for ~3 months in 2016
- did this part of a blockchain pilot project run by Dutch Government
- this blockchain project is exploring about 30 different use cases
- brought up IPFS. nobody really knew what it was
- it ws interesting to have a go with mediachain
Slides: Report on Mediachain Pilot for Dutch Heritage Network (PDF)
Focuses:
- visibility
- accessibility
- preservation
"Quite a hassle to set up the right way"
Results:
- Sharing of data is limited (copying & merging only)
- Discovery is limited
Capabilities, esp. with linked data, limited Blockstack was interesting
Observations:
- Mediachain still immature. No further development planned due to Spotify acquisition.
- Docs out of date or overpromised
- Slack channel support was excellent, though
- Developers enthusiastic about this use case
- MC had very pragmatic approach
Conclusions:
- Linked Data support is weak
- Real potential of IPFS can't be evaluated through this pilot due to limited IPFS imp in MediaChain
- Chose more traditional way of getting Linked Data setup working
Discussion:
-
Ed Silverton also tried MC. Got a node running, not sure what to do with it. Considered merging data. Also concerned about lack of IPFS compatibility. MC plan was for IPLD to link data/blogs, but stalled.
-
What's the key difference between blockchain and CRDT?
-
Mediachain was a way to find my stuff on IPFS, otherwise I'd have to remember my hash. Discoverability.
Possible future topics:
- Old Dominion University team to present work with Interplantary Wayback
- Nicola for IPLD overview