2018/Berlin/datalake

Digital Archiving was a session at IndieWebCamp Berlin 2018.

Notes archived from: https://etherpad.indieweb.org/datalake

---

IndieWebCamp Berlin 2018
Session: Digital Archiving
When: 2018-11-03 14:10

Participants

• Sebastion Dümcke
• Calum Ryan
• Peter Molnar
• Jeremy Keith
• David Shanske
• Sebastian Greger
• Sneha Belkhale
• Toni Mattis
• Sven Knebel
• Greg McVerry

Notes

How to store

• media vs format
• licensing is interesting: public, permissive licensed has chance to get additional copies outside
• media:
  • reliablitiy of media vs reliability of interfaces
  • Tape is awesome in reliability/storage duration, but readers are expensive and need to be maintained too
  • optical: CD is random, DVDs (outside maybe M-Disk?) less good. readers long-term available. Blue-Ray: archival grade available, cheap ones organic storage layers. Expensive media too.
  • Flash (SD, SSD, CF, ...) loose data (charge) if unpowered for extensive period (~5 years) (SSDs I've seen numbers in the months if not stored in cold places)
  • HDDs are relatively cheap, sturdy even if stored powered off
• digital:
  • text is great
  • common, simple formats are good (jpg, png)
  • PDF is complex - if you do want to store to print, LaTeX (https://www.latex-project.org/) or PostScript
  • potentially: archive software too if possible (trickier for modern, clound-connected software) - if archiveing source code, archive compiler, and all dependencies as well

how to archive (structure)

• who wants to find it
  • indexed from structure vs search based, flexible data
  • automatic tagging can work surprisingly well nowadays for images
  • petermolnar: images archived by topic, data, occasion
  • storage of metadata in files themselves is attractive (easy to move etc), but can be tricky with non-text format
  • simple index evolved into ontology/tagging

How to find stuff

• again audience: if external users, central, somewhat standardized hierarchy is useful.
• full-text search

• 4 How to share/publish
• 5 how to gather all
  • browser hierachy would be great worldbrain.io, webmemex

References

• "As we may think" article describing idea of Memex machine
• Zotero also works for websites

See Also

• 2018/Berlin/Sessions