GIS features

Sparse hierarchical knowledge base

• type system

• typed? facts
• typed? relations between facts (facts, too)

• scripting: mutator/inference

• pull data from periodically updateable public source with diff/mark deleted
• notifies (before/when license expires etc)

• distributed storage for FT/HA
• views and sync relationships

• one-way machine readable exports/reports (txt, xml, csv etc)
• two-way formalized IO (over WebDAV or local files)

• users, groups
• data security
• ACL
  for starters: per view
  then: row- and column-level
  then: conditional:
  field based
  inference based
  data at rest: to be handled by OS
  data in motion: TLS or better
  recommended: ECDHE-ECSDA-PFS
  basic auth over TLS ok
• web view
  admin level editor to access/CRUD raw data
• geospatial datatype
  primitives (points, lines polygons, 3d poly)
  linear relations between primitives (e.g. power grid connectivity graph)
  undirected/directed?
  custom rendering per 'data type'
  R/W KML I/O for geotagged data
• Performance expectations
  Server instance should deliver usable pefrormance given 1-4 i3+ cores, 512-1024 MB user RAM
  requirement non-mandatory for development and alpha builds
  UI must be responsive
  Bandwidth use profiles
  a. 1 Gbps 10 ms max RTT
  b. 1 Mbps 50 ms median RTT
  c. 256 kbps 350 ms median RTT
• Non-functional requirements
  Runs on (at least one)
  Current Debian (Server) x64 (as vanilla as possible)
  Windows Server 2012R2
  VM image (not preferred)
• Use cases
  personal knowledge base
  replacement for Excel/OneNote/wiki/Google Earth fact hoard
  PIM/RememberTheMilk

via @sashimanu-san