Microsoft Word - UCCEISMIC_DesignPrinciples_final_draft.doc


	
   1	
  

Open	
  Principles,	
  Open	
  Data:	
  The	
  Design	
  Principles	
  and	
  Architecture	
  of	
  

the	
  UC	
  CEISMIC	
  Canterbury	
  Earthquakes	
  Digital	
  Archive	
  

James	
  Smithies,	
  Paul	
  Millar,	
  Chris	
  Thomson	
  

Digital	
  Humanists	
  are	
  developing	
  a	
  tradition	
  of	
  disaster	
  archiving.	
  The	
  trend	
  began	
  with	
  the	
  

Centre	
  for	
  History	
  and	
  New	
  Media’s	
  (CHNM)	
  9-­‐11	
  archive,1	
  started	
  after	
  the	
  attacks	
  on	
  the	
  

World	
  Trade	
  Centre	
  in	
  New	
  York	
  in	
  2001,	
  which	
  eventually	
  crowd-­‐sourced	
  150,000	
  items	
  

and	
  has	
  been	
  itself	
  archived	
  by	
  the	
  Library	
  of	
  Congress.	
  CHNM,	
  in	
  partnership	
  with	
  the	
  

University	
  of	
  New	
  Orleans,	
  later	
  developed	
  the	
  Hurricane	
  Memory	
  Bank	
  (2005),	
  in	
  response	
  

to	
  hurricanes	
  Katrina	
  and	
  Rita.2	
  Digital	
  humanists	
  have	
  also	
  been	
  involved	
  in	
  responses	
  to	
  

the	
  devastating	
  Tōhoku	
  earthquake	
  and	
  tsunami	
  of	
  2011.3	
  This	
  article	
  outlines	
  the	
  approach	
  

taken	
  by	
  digital	
  humanists	
  at	
  the	
  University	
  of	
  Canterbury,	
  New	
  Zealand,	
  to	
  the	
  series	
  of	
  

earthquakes	
  that	
  seriously	
  damaged	
  the	
  city	
  of	
  Christchurch	
  from	
  2010	
  –	
  2011.	
  In	
  many	
  

ways,	
  these	
  initiatives	
  represent	
  a	
  continuation	
  in	
  the	
  digital	
  sphere	
  of	
  activities	
  many	
  

centuries	
  old.	
  Disasters	
  are,	
  after	
  all,	
  primarily	
  human	
  events:	
  nature	
  may	
  cause	
  them,	
  but	
  it	
  

is	
  the	
  human	
  impact	
  that	
  demands	
  a	
  response.	
  Many	
  of	
  our	
  cultural	
  memories	
  of	
  great	
  

disasters	
  were	
  created	
  by	
  humanists:	
  oral	
  stories	
  and	
  subsequent	
  myths	
  about	
  the	
  

destruction	
  of	
  Atlantis	
  in	
  1620	
  BC,	
  extensive	
  literary	
  and	
  historical	
  references	
  to	
  the	
  Lisbon	
  

earthquake	
  of	
  1755	
  and	
  the	
  eruption	
  of	
  Krakatoa	
  in	
  1883,	
  artistic	
  impressions	
  of	
  the	
  

destruction	
  of	
  the	
  pink	
  and	
  white	
  terraces	
  in	
  the	
  North	
  Island	
  of	
  New	
  Zealand	
  in	
  1886.	
  

Although	
  any	
  disaster	
  response	
  must	
  be	
  regulated	
  by	
  the	
  need	
  to	
  prioritize	
  life	
  and	
  

property,	
  history	
  shows	
  us	
  that	
  humanists	
  will	
  normally	
  be	
  somewhere	
  in	
  the	
  mix,	
  collecting,	
  

preserving	
  and	
  commenting	
  on	
  the	
  event	
  for	
  present	
  and	
  future	
  generations.	
  The	
  

appearance	
  of	
  digital	
  tools	
  has	
  provided	
  us	
  with	
  new	
  avenues	
  for	
  our	
  activities	
  in	
  disaster	
  

and	
  post-­‐disaster	
  contexts,	
  but	
  has	
  done	
  little	
  to	
  alter	
  the	
  innate	
  drive	
  to	
  collect,	
  analyze,	
  



	
   2	
  

create,	
  and	
  explain.	
  This	
  article	
  outlines	
  the	
  design	
  principles	
  and	
  architecture	
  of	
  the	
  UC	
  

CEISMIC	
  Canterbury	
  Earthquakes	
  Digital	
  Archive	
  in	
  an	
  attempt	
  to	
  record	
  an	
  approach	
  to	
  DH	
  

disaster	
  response	
  that	
  might	
  benefit	
  future	
  efforts.	
  	
  	
  	
  	
  	
  	
  

1. Background	
  

At	
  12:51	
  on	
  Tuesday	
  22	
  February	
  2011	
  a	
  magnitude	
  M6.3	
  earthquake	
  struck	
  the	
  city	
  of	
  

Christchurch,	
  on	
  the	
  east	
  coast	
  of	
  New	
  Zealand’s	
  South	
  Island.	
  Local	
  geography	
  and	
  soil	
  

structure,	
  combined	
  with	
  a	
  series	
  of	
  faults	
  under	
  the	
  Canterbury	
  Plains,	
  produced	
  significant	
  

amounts	
  of	
  damage.	
  Unlike	
  the	
  7.1	
  earthquake	
  experienced	
  on	
  September	
  4th	
  of	
  the	
  

previous	
  year,	
  this	
  event	
  produced	
  remarkably	
  high	
  rates	
  of	
  ground	
  movement	
  and	
  resulted	
  

in	
  the	
  loss	
  of	
  185	
  lives.4	
  At	
  the	
  time	
  of	
  writing	
  over	
  13005	
  of	
  the	
  30006	
  buildings	
  in	
  the	
  

Central	
  Business	
  District	
  (CBD)	
  have	
  been	
  demolished,	
  in	
  a	
  process	
  that	
  will	
  continue	
  well	
  

past	
  the	
  two	
  year	
  anniversary	
  of	
  the	
  event.	
  The	
  government	
  has	
  noted	
  that	
  “much	
  of	
  the	
  

historic	
  fabric	
  [of	
  Christchurch]	
  has	
  been	
  lost,	
  as	
  have	
  key	
  facilities	
  such	
  as	
  the	
  Convention	
  

Centre,	
  a	
  significant	
  proportion	
  of	
  the	
  hotel	
  capacity,	
  and	
  sports	
  and	
  recreation	
  facilities”.7	
  

There	
  was	
  severe	
  damage	
  to	
  lifelines	
  infrastructure	
  across	
  the	
  city	
  of	
  450,000	
  people,	
  

“…including	
  road,	
  the	
  water	
  and	
  wastewater	
  networks	
  and	
  the	
  electric	
  systems”.8	
  

Eventually,	
  entire	
  suburbs	
  had	
  to	
  be	
  abandoned	
  due	
  to	
  liquefaction,	
  subsidence,	
  rock	
  fall,	
  

and	
  a	
  host	
  of	
  other	
  geo-­‐structural	
  issues.	
  The	
  city	
  rebuild	
  is	
  expected	
  to	
  cost	
  NZ$30	
  billion	
  

and	
  take	
  10	
  –	
  15	
  years	
  to	
  complete.	
  By	
  August	
  2012	
  over	
  143,000	
  insurance	
  claims	
  have	
  

been	
  made	
  with	
  the	
  New	
  Zealand’s	
  Earthquake	
  Commission.	
  9	
  	
  As	
  of	
  March	
  2012,	
  the	
  

disaster	
  ranked	
  as	
  the	
  third	
  costliest	
  insurance	
  event	
  in	
  history.10	
  Four	
  earthquakes	
  of	
  

magnitude	
  6	
  or	
  greater	
  and	
  over	
  11,000	
  aftershocks	
  have	
  been	
  recorded	
  in	
  the	
  area	
  since	
  

September	
  4th	
  2010.11	
  Residents	
  continue	
  to	
  struggle	
  with	
  significant	
  disruption	
  to	
  their	
  daily	
  

lives.	
  



	
   3	
  

2. Disaster	
  Management	
  and	
  Digital	
  Archives	
  

Post-­‐disaster	
  recovery	
  plans	
  focus	
  –	
  as	
  they	
  must	
  –	
  on	
  saving	
  lives,	
  identifying	
  victims,	
  

reconstituting	
  essential	
  services,	
  and	
  providing	
  information	
  to	
  residents	
  and	
  business	
  

owners.	
  They	
  are	
  characterized	
  by	
  overtly	
  vertical	
  command	
  and	
  control	
  management	
  

structures,	
  to	
  assist	
  with	
  the	
  coordination	
  of	
  emergency	
  teams,	
  government	
  agencies	
  and	
  

the	
  army.	
  Normally	
  only	
  essential	
  recovery	
  activities	
  are	
  resourced	
  in	
  an	
  attempt	
  to	
  

maximize	
  the	
  response	
  and	
  minimize	
  confusion.12	
  That	
  said,	
  disaster	
  management	
  

methodology	
  highlights	
  four	
  phases	
  that	
  guide	
  the	
  response	
  of	
  governments,	
  local	
  body	
  

authorities	
  and	
  Non-­‐governmental	
  Organizations	
  (NGOs)	
  to	
  disaster	
  situations:	
  mitigation,	
  

preparedness,	
  response,	
  and	
  recovery.	
  This	
  holistic	
  approach,	
  that	
  includes	
  preparation	
  

beforehand	
  and	
  recovery	
  afterwards,	
  is	
  particularly	
  well	
  suited	
  to	
  a	
  response	
  by	
  humanists,	
  

because	
  it	
  acknowledges	
  the	
  need	
  for	
  long-­‐term	
  participation	
  and	
  accepts	
  the	
  need	
  for	
  

cultural	
  and	
  educational	
  input	
  in	
  order	
  to	
  lessen	
  both	
  the	
  immediate	
  and	
  long-­‐term	
  impacts	
  

of	
  disasters.13	
  Although	
  every	
  country	
  differs	
  in	
  its	
  disaster	
  response	
  capabilities	
  and	
  policies	
  

(and	
  all	
  will	
  prioritize	
  the	
  preservation	
  of	
  life	
  and	
  property	
  in	
  the	
  immediate	
  aftermath	
  of	
  an	
  

event),	
  most	
  government	
  agencies	
  in	
  the	
  OECD	
  are	
  aware	
  of	
  the	
  need	
  for	
  broad-­‐based	
  

information	
  gathering,	
  education,	
  and	
  the	
  development	
  of	
  social	
  and	
  cultural	
  capital:	
  

Memory,	
  experience,	
  and	
  knowledge	
  are	
  critical	
  to	
  the	
  development	
  of	
  effective	
  response	
  

mechanisms.	
  Knowledge	
  of	
  past	
  events	
  can	
  condition	
  how	
  contemporary	
  society	
  not	
  only	
  

conceptualizes	
  the	
  risk	
  connected	
  with	
  particular	
  events	
  but	
  also	
  anticipates	
  the	
  impacts	
  of	
  

future	
  catastrophes.14	
  

This	
  principle	
  is	
  enshrined	
  in	
  UNESCO’s	
  Text	
  of	
  the	
  Convention	
  for	
  the	
  Safeguarding	
  of	
  

Intangible	
  Cultural	
  Heritage (2003)15	
  and	
  the	
  United	
  Nations’	
  Hyogo	
  Framework	
  for	
  Action	
  

(2005	
  –	
  2015),	
  which	
  includes	
  a	
  priority	
  action	
  to	
  “[u]se	
  knowledge,	
  innovation	
  and	
  



	
   4	
  

education	
  to	
  build	
  a	
  culture	
  of	
  safety	
  and	
  resilience	
  at	
  all	
  levels”.16	
  Similarly,	
  the	
  2013	
  

Vancouver	
  Declaration	
  explicitly	
  “urged	
  the	
  [UNESCO]	
  secretariat”	
  to	
  	
  

create	
  an	
  emergency	
  programme	
  aiming	
  at	
  preservation	
  of	
  documentary	
  material	
  

endangered	
  by	
  natural	
  disasters	
  or	
  armed	
  conflicts,	
  as	
  well	
  as	
  a	
  programme	
  for	
  the	
  recovery	
  

of	
  analogue	
  and	
  digital	
  heritage	
  that	
  is	
  under	
  threat	
  of	
  becoming,	
  or	
  is	
  already	
  inaccessible	
  

because	
  of	
  obsolete	
  hardware	
  or	
  software.17	
  

University	
  of	
  Canterbury18	
  researchers	
  were	
  heavily	
  involved	
  in	
  the	
  early	
  disaster	
  

management	
  response	
  to	
  the	
  February	
  earthquake,	
  coordinating	
  information	
  about	
  the	
  

status	
  of	
  essential	
  infrastructure	
  and	
  services,	
  providing	
  high	
  performance	
  computing	
  

storage	
  and	
  services,	
  and	
  educating	
  the	
  public	
  through	
  the	
  media.	
  Few	
  roles,	
  though,	
  were	
  

available	
  to	
  humanities	
  academics	
  in	
  the	
  weeks	
  after	
  the	
  February	
  2011	
  earthquake.	
  The	
  

University	
  of	
  Canterbury	
  closed	
  for	
  several	
  weeks,	
  to	
  allow	
  staff	
  and	
  students	
  to	
  contact	
  

loved	
  ones	
  and	
  attend	
  to	
  damaged	
  properties.	
  Thousands	
  of	
  people	
  left	
  the	
  city	
  to	
  avoid	
  

aftershocks.	
  A	
  state	
  of	
  emergency	
  stayed	
  in	
  place	
  until	
  April	
  30th,19	
  making	
  it	
  impossible	
  to	
  

attend	
  to	
  business	
  as	
  usual.	
  Damage	
  to	
  buildings	
  and	
  general	
  infrastructure,	
  and	
  severe	
  

dislocation	
  of	
  regional	
  infrastructure,	
  created	
  significant	
  operational	
  problems	
  for	
  many	
  

months.	
  The	
  loss	
  of	
  life	
  and	
  presence	
  of	
  international	
  search	
  and	
  rescue,	
  the	
  military,	
  and	
  

increased	
  numbers	
  of	
  police	
  contributed	
  to	
  a	
  general	
  sense	
  of	
  emergency	
  and	
  disorder.	
  

When	
  the	
  university	
  did	
  open,	
  during	
  March,	
  staff	
  conducted	
  some	
  classes	
  in	
  tents	
  as	
  

contractors	
  began	
  building	
  prefabricated	
  classrooms	
  on	
  vacant	
  ground;	
  the	
  situation	
  was	
  

anything	
  but	
  conducive	
  to	
  the	
  development	
  of	
  new	
  digital	
  resources.	
  It	
  wasn’t	
  until	
  May,	
  2	
  

months	
  after	
  the	
  first	
  major	
  event,	
  that	
  Paul	
  Millar	
  from	
  the	
  Department	
  of	
  English	
  was	
  able	
  

to	
  start	
  thinking	
  about	
  possible	
  responses.	
  In	
  an	
  indication	
  of	
  the	
  problems	
  university	
  staff	
  

were	
  facing	
  at	
  the	
  time,	
  Millar	
  discussed	
  possible	
  options	
  with	
  the	
  author	
  of	
  this	
  paper	
  while	
  

he	
  was	
  temporarily	
  located	
  in	
  Wellington,	
  because	
  his	
  house	
  was	
  unfit	
  for	
  habitation.	
  	
  



	
   5	
  

Despite	
  the	
  difficulties,	
  and	
  as	
  Millar	
  was	
  aware,	
  a	
  robust	
  response	
  to	
  disaster	
  situations	
  

by	
  humanists	
  soon	
  after	
  the	
  event	
  itself	
  is	
  important,	
  not	
  only	
  for	
  disaster	
  management	
  

decision	
  support20	
  but	
  also	
  in	
  broader	
  cultural	
  terms,	
  to	
  avoid	
  the	
  “digital	
  death”21	
  of	
  crucial	
  

artifacts	
  produced	
  as	
  a	
  direct	
  result	
  of	
  events.	
  Evidence	
  suggests	
  that	
  the	
  formation	
  of	
  

autobiographical	
  and	
  collective	
  memory	
  in	
  the	
  aftermath	
  of	
  significant	
  trauma	
  is	
  a	
  complex	
  

affair.	
  Psychologists22	
  and	
  sociologists23	
  alike	
  point	
  to	
  the	
  role	
  emotion	
  plays	
  in	
  the	
  

construction	
  of	
  individual	
  and	
  collective	
  memories	
  of	
  events,	
  so	
  it	
  is	
  important	
  that	
  large	
  

bodies	
  of	
  primary	
  material	
  are	
  identified	
  and	
  safely	
  stored	
  for	
  later	
  analysis	
  (both	
  by	
  

professional	
  researchers	
  and	
  individuals)	
  when	
  the	
  immediate	
  trauma	
  has	
  passed.	
  By	
  doing	
  

so,	
  individual	
  and	
  collective	
  cultural	
  memory	
  of	
  the	
  event	
  can	
  be	
  continually	
  revisited	
  and	
  

refined	
  and	
  a	
  longitudinal	
  understanding	
  developed.	
  Although	
  cultural	
  resources	
  tend	
  to	
  be	
  

“ignored	
  and	
  neglected”24	
  in	
  the	
  immediate	
  post-­‐disaster	
  phase,	
  the	
  experience	
  of	
  the	
  

September	
  11,	
  Hurricane	
  Memory	
  Bank,	
  UC	
  CEISMIC,	
  and	
  Tōhoku	
  archives	
  suggest	
  there	
  are	
  

compelling	
  arguments	
  for	
  the	
  speedy	
  deployment	
  of	
  cultural	
  heritage-­‐related	
  assets.	
  Studies	
  

suggest	
  a	
  focus	
  on	
  ‘social	
  capital’	
  (which	
  includes	
  a	
  cultural	
  component)	
  can	
  speed	
  recovery,	
  

enhance	
  cohesiveness,	
  and	
  contribute	
  to	
  post-­‐disaster	
  resilience.25	
  Patrick	
  Meier	
  has	
  

pointed	
  out	
  the	
  specific	
  role	
  collections	
  of	
  big	
  data	
  can	
  have	
  in	
  these	
  processes.26	
  It	
  is	
  

important	
  to	
  recognize	
  the	
  “multidimensionality”27	
  of	
  disasters,	
  and	
  the	
  impact	
  a	
  loss	
  of	
  

cultural	
  heritage	
  can	
  have	
  on	
  communities.	
  In	
  some	
  cases,	
  “rescuing	
  culture	
  is	
  essential	
  for	
  

the	
  mental	
  survival	
  of	
  people	
  in	
  emergency	
  situations,	
  and	
  can	
  contribute	
  to	
  their	
  overall	
  

resilience	
  and	
  empowerment	
  when	
  overcoming	
  catastrophe”.28	
  	
  

The	
  problem	
  is,	
  of	
  course,	
  that	
  easily	
  deployable	
  digital	
  archives	
  suitable	
  for	
  the	
  complex	
  

task	
  of	
  post-­‐disaster	
  collection	
  do	
  not	
  exist.	
  While	
  applications	
  like	
  Omeka,	
  Islandora,	
  

Ushahidi,	
  and	
  Fedora	
  Commons	
  provide	
  excellent	
  starting	
  points,	
  developing	
  a	
  ‘shrink-­‐

wrapped’	
  solution	
  tailored	
  specifically	
  to	
  post-­‐disaster	
  situations	
  would	
  be	
  a	
  non-­‐trivial	
  task.	
  

The	
  ideal	
  situation	
  would	
  be	
  one	
  in	
  which	
  humanists,	
  or	
  perhaps	
  government	
  employees	
  



	
   6	
  

within	
  a	
  central	
  cultural	
  heritage	
  agency,	
  could	
  visit	
  an	
  online	
  service	
  provider	
  (or	
  download	
  

an	
  easily	
  deployable	
  virtual	
  machine	
  to	
  deploy	
  on	
  their	
  own	
  infrastructure)	
  after	
  a	
  disaster	
  

and	
  provision	
  a	
  robust,	
  preservation-­‐quality,	
  archive	
  system	
  capable	
  of	
  ingesting	
  large	
  

quantities	
  of	
  digital	
  content	
  (either	
  crowd-­‐sourced	
  or	
  through	
  an	
  administrative	
  interface)	
  

according	
  to	
  configurable	
  standards-­‐based	
  ontologies.	
  The	
  infrastructure	
  would	
  be	
  on-­‐

demand	
  Infrastructure	
  as	
  a	
  Service	
  (IAAS)	
  and	
  fully	
  scalable.	
  The	
  software	
  would	
  be	
  provided	
  

as	
  a	
  service	
  as	
  well	
  (SAAS),	
  perhaps	
  with	
  a	
  modular	
  architecture	
  to	
  allow	
  administrators	
  to	
  

deploy	
  services	
  as	
  required	
  (including	
  integration	
  with	
  and	
  archiving	
  of	
  social	
  media	
  services,	
  

and	
  provision	
  of	
  downstream	
  big	
  data	
  analysis).	
  It	
  would	
  be	
  capable	
  of	
  metadata	
  

aggregation	
  to	
  allow	
  it	
  to	
  act	
  as	
  the	
  central	
  node	
  in	
  a	
  heterogeneous	
  federated	
  archive,	
  and	
  

have	
  an	
  easy	
  to	
  use	
  interface	
  and	
  ‘baked	
  in’	
  Terms	
  and	
  Conditions	
  and	
  copyright	
  tools,	
  to	
  

ensure	
  broad-­‐based	
  usage	
  and	
  legal	
  probity.	
  Large-­‐scale	
  data	
  export	
  functionality	
  would	
  

allow	
  for	
  migration	
  of	
  content	
  to	
  long-­‐term	
  preservation	
  systems	
  and	
  dark	
  archives.	
  In	
  lieu	
  

of	
  such	
  a	
  service	
  systems	
  need	
  to	
  be	
  put	
  together	
  very	
  quickly	
  by	
  humanities	
  or	
  cultural-­‐

heritage	
  teams,	
  in	
  the	
  difficult	
  circumstances	
  of	
  post-­‐disaster	
  management	
  and	
  with	
  

minimal	
  funding.	
  Sometimes,	
  as	
  was	
  the	
  case	
  with	
  the	
  September	
  11	
  archive,	
  the	
  process	
  

works	
  very	
  well	
  and	
  great	
  benefits	
  accrue	
  from	
  small	
  outlay;	
  at	
  other	
  times,	
  as	
  with	
  the	
  

Hurricane	
  Memory	
  Bank,	
  a	
  similar	
  team	
  can	
  go	
  through	
  the	
  same	
  process	
  but	
  with	
  

significantly	
  reduced	
  results.29	
  There	
  is	
  currently	
  no	
  example	
  of	
  a	
  broad	
  archive	
  aggregating	
  

post-­‐disaster	
  cultural	
  heritage,	
  scientific,	
  geographic	
  and	
  social	
  data	
  in	
  a	
  manner	
  conducive	
  

to	
  long-­‐term	
  preservation,	
  general	
  research	
  by	
  the	
  public	
  and	
  academics,	
  and	
  

computationally	
  intensive	
  research.	
  

3. Governance	
  

One	
  issue	
  with	
  the	
  development	
  of	
  a	
  generic	
  cultural-­‐heritage	
  disaster	
  archive	
  system	
  would	
  

be	
  how	
  to	
  deploy	
  it	
  in	
  a	
  variety	
  of	
  operational,	
  cultural,	
  and	
  legal	
  contexts.	
  While	
  New	
  



	
   7	
  

Zealand’s	
  culture	
  and	
  local	
  and	
  central	
  government	
  structure	
  bears	
  many	
  similarities	
  to	
  

other	
  OECD	
  nations	
  there	
  are	
  significant	
  differences	
  too,	
  especially	
  in	
  its	
  lack	
  of	
  state	
  or	
  

federal	
  structures,	
  which	
  enabled	
  UC	
  CEISMIC	
  to	
  develop	
  a	
  broad-­‐ranging	
  Consortium	
  of	
  

both	
  local	
  and	
  central	
  government	
  agencies.	
  Differences	
  with	
  developing	
  countries	
  would	
  

presumably	
  be	
  even	
  more	
  profound.	
  Detailed	
  discussion	
  of	
  these	
  issues	
  is	
  outside	
  the	
  scope	
  

of	
  this	
  article,	
  however.	
  The	
  important	
  thing	
  to	
  note	
  is	
  merely	
  that	
  the	
  UC	
  CEISMIC	
  archive	
  

was	
  designed	
  and	
  deployed	
  with	
  a	
  particular	
  operational	
  context	
  in	
  mind.	
  New	
  Zealand’s	
  

small	
  cultural	
  network	
  meant	
  that	
  it	
  soon	
  became	
  clear	
  that	
  a	
  variety	
  of	
  groups	
  were	
  

considering	
  or	
  actively	
  engaged	
  in	
  developing	
  archive	
  systems	
  to	
  collect	
  quake-­‐related	
  

content.	
  A	
  meeting	
  was	
  held	
  at	
  Lincoln	
  University	
  amongst	
  interested	
  parties	
  and	
  a	
  decision	
  

was	
  made	
  to	
  form	
  a	
  Consortium,	
  which	
  eventually	
  comprised	
  Archives	
  New	
  Zealand,	
  

Christchurch	
  City	
  Libraries,	
  the	
  Canterbury	
  Museum,	
  the	
  Canterbury	
  Earthquake	
  Recovery	
  

Authority,	
  the	
  Ministry	
  for	
  Culture	
  and	
  Heritage,	
  the	
  National	
  Library,	
  New	
  Zealand	
  Film	
  

Archive,	
  NZ	
  on	
  Screen,	
  the	
  Ngai	
  Tahu	
  Research	
  Centre	
  and	
  Te	
  Papa	
  Tongarewa:	
  The	
  Museum	
  

of	
  New	
  Zealand.	
  The	
  new	
  University	
  of	
  Canterbury	
  Digital	
  Humanities	
  Programme	
  led	
  the	
  

Consortium,	
  which	
  came	
  to	
  be	
  known	
  as	
  the	
  UC	
  CEISMIC	
  Consortium.30	
  Responsibility	
  for	
  

the	
  development	
  of	
  the	
  Consortium	
  and	
  leadership	
  of	
  the	
  project	
  as	
  a	
  whole	
  rested	
  with	
  

Paul	
  Millar	
  as	
  Director,	
  and	
  responsibility	
  for	
  project	
  management	
  and	
  technical	
  

development	
  of	
  the	
  federated	
  archive	
  (and	
  additional	
  University	
  of	
  Canterbury	
  research	
  

repository)	
  rested	
  with	
  the	
  author.	
  The	
  University	
  of	
  Canterbury	
  (UC)	
  provided	
  funds	
  for	
  the	
  

first	
  two	
  years	
  of	
  development	
  and	
  operations,	
  and	
  Consortium	
  members	
  offered	
  both	
  

technical	
  and	
  archival	
  resources,	
  and	
  input	
  from	
  Chief	
  Executives,	
  senior	
  managers,	
  and	
  

general	
  staff.	
  The	
  project	
  benefitted	
  from	
  a	
  remarkable	
  degree	
  of	
  inter-­‐agency	
  cooperation	
  

and	
  goodwill	
  that	
  might	
  not	
  have	
  been	
  possible	
  in	
  other	
  countries.	
  Although	
  there	
  turned	
  

out	
  to	
  be	
  little	
  practical	
  need	
  for	
  it,	
  the	
  project	
  was	
  underpinned	
  by	
  a	
  broad	
  adherence	
  to	
  a	
  

concept	
  of	
  mutual	
  aid.	
  It	
  was	
  agreed	
  that	
  Consortium	
  members	
  would	
  help	
  each	
  other	
  



	
   8	
  

where	
  possible,	
  even	
  if	
  that	
  meant	
  supporting	
  or	
  improving	
  a	
  ‘competing’	
  archive.	
  Similarly,	
  

if	
  smaller	
  nodes	
  began	
  failing	
  in	
  future	
  years,	
  the	
  broader	
  Consortium	
  would	
  try	
  to	
  step	
  in	
  

and	
  help,	
  or	
  migrate	
  their	
  content	
  to	
  more	
  robust	
  infrastructures.	
  The	
  goal	
  was	
  to	
  create	
  a	
  

radical	
  model,	
  where	
  the	
  whole	
  was	
  always	
  held	
  up	
  as	
  greater	
  than	
  its	
  parts.	
  Paperwork	
  and	
  

official	
  documentation	
  was	
  kept	
  to	
  a	
  minimum,	
  and	
  any	
  documents	
  that	
  were	
  produced	
  

adhered	
  to	
  a	
  ‘less	
  is	
  more’	
  principle.	
  The	
  founding	
  document	
  of	
  the	
  Consortium	
  was	
  a	
  three	
  

page	
  Memorandum	
  of	
  Understanding,	
  which	
  allowed	
  for	
  any	
  member	
  to	
  leave	
  the	
  

Consortium	
  with	
  two	
  weeks	
  notice,	
  signed	
  by	
  senior	
  representatives	
  of	
  all	
  organizations.	
  

Policies	
  and	
  processes	
  for	
  ingestion	
  into	
  the	
  archive	
  are	
  determined	
  by	
  the	
  policies	
  and	
  

processes	
  of	
  contributing	
  members,	
  with	
  the	
  Consortium	
  providing	
  advice	
  to	
  the	
  UC	
  team	
  in	
  

establishing	
  their	
  new	
  processes.	
  In	
  general,	
  aside	
  from	
  community	
  archives	
  where	
  a	
  lower	
  

standard	
  is	
  accepted,	
  the	
  contributing	
  archives	
  can	
  be	
  said	
  to	
  adhere	
  to	
  best	
  practice.31	
  In	
  

many	
  ways	
  the	
  UC	
  CEISMIC	
  federation	
  presents	
  a	
  classic	
  example	
  of	
  the	
  use	
  of	
  information	
  

federalism	
  to	
  manage	
  “information	
  and	
  [establish]	
  standards	
  for	
  cultures	
  that	
  celebrate	
  

empowerment	
  and	
  widespread	
  participation”.32	
  

4. Design	
  Principles	
  

The	
  first	
  act	
  in	
  the	
  technical	
  development	
  of	
  the	
  archive	
  was	
  the	
  organization	
  of	
  an	
  

information	
  architecture	
  workshop	
  involving	
  technical	
  personnel	
  from	
  a	
  variety	
  of	
  

Consortium	
  partners.	
  The	
  workshop	
  was	
  held	
  at	
  NV	
  Interactive,	
  a	
  web	
  development	
  

company	
  already	
  contracted	
  by	
  the	
  Ministry	
  for	
  Culture	
  and	
  Heritage	
  to	
  build	
  their	
  

‘QuakeStories’	
  archive,	
  33	
  and	
  later	
  contracted	
  by	
  UC	
  to	
  build	
  the	
  main	
  web	
  portal	
  for	
  the	
  

federated	
  archive:	
  http://www.ceismic.org.nz.	
  At	
  this	
  workshop,	
  general	
  principles	
  were	
  

agreed	
  about	
  how	
  the	
  Consortium	
  archives	
  would	
  work	
  together	
  and	
  it	
  was	
  agreed	
  that,	
  

rather	
  than	
  pooling	
  existing	
  resources	
  to	
  create	
  a	
  single	
  archive,	
  the	
  team	
  would	
  work	
  to	
  a	
  

“distributed	
  custody	
  model”,34	
  storing	
  content	
  in	
  a	
  variety	
  of	
  existing	
  and	
  planned	
  



	
   9	
  

repositories,	
  and	
  contributing	
  content	
  to	
  a	
  federated	
  archive	
  via	
  metadata	
  aggregation.	
  

Although	
  some	
  principles	
  were	
  pinned	
  down	
  later,	
  and	
  no	
  formal	
  list	
  was	
  ever	
  produced	
  and	
  

agreed	
  to,	
  the	
  following	
  design	
  principles	
  were	
  discussed:	
  

1. Open	
  Access:	
  The	
  concept	
  of	
  a	
  federated	
  archive	
  would	
  not	
  have	
  gained	
  approval	
  

without	
  this.	
  No	
  Consortium	
  partners	
  were	
  willing	
  to	
  contribute	
  their	
  existing	
  

content	
  to	
  a	
  gated	
  archive,	
  and	
  it	
  was	
  felt	
  that	
  individuals	
  and	
  organizations	
  would	
  

be	
  unlikely	
  to	
  contribute	
  additional	
  content	
  in	
  those	
  circumstances	
  either.	
  

2. Open	
  Source:	
  Some	
  of	
  the	
  government	
  agencies	
  were	
  already	
  using	
  proprietary	
  

software	
  and	
  would	
  continue	
  to	
  do	
  so,	
  but	
  the	
  workshop	
  evinced	
  broad	
  agreement	
  

that	
  open	
  source	
  components	
  should	
  be	
  used	
  wherever	
  possible,	
  in	
  order	
  to	
  foster	
  

sharing	
  and	
  reuse.	
  

3. Multi-­‐channel:	
  ceismic.org.nz	
  would	
  be	
  the	
  ‘front	
  door’,	
  but	
  it	
  was	
  agreed	
  that	
  the	
  

federation	
  would	
  aim	
  for	
  a	
  radically	
  multi-­‐channel	
  approach.	
  The	
  metaphor	
  of	
  an	
  

‘ecosystem’	
  was	
  used	
  to	
  describe	
  a	
  belief	
  that	
  all	
  nodes	
  in	
  the	
  federation	
  were	
  to	
  be	
  

of	
  equal	
  importance.	
  Small	
  community	
  archives,	
  after	
  all,	
  could	
  well	
  contain	
  more	
  

valuable	
  content	
  than	
  large	
  national	
  ones.	
  The	
  key	
  was	
  to	
  facilitate	
  and	
  foster	
  a	
  

broad,	
  healthy	
  federation,	
  which	
  was	
  capable	
  of	
  supporting	
  large	
  and	
  small	
  

partners.	
  	
  

4. Asymmetry:	
  Because	
  of	
  the	
  support	
  behind	
  the	
  UC	
  team	
  it	
  was	
  understood	
  that	
  the	
  

proposed	
  QuakeStudies	
  repository	
  was	
  likely	
  to	
  become	
  the	
  largest	
  node	
  in	
  the	
  

federation.	
  Other	
  contributing	
  organizations	
  were	
  constrained	
  by	
  their	
  normal	
  

business	
  as	
  usual	
  operations,	
  and	
  although	
  the	
  New	
  Zealand	
  government	
  had	
  

directed	
  them	
  to	
  prioritize	
  quake-­‐related	
  activities	
  as	
  part	
  of	
  their	
  strategic	
  plans	
  

(and	
  the	
  Christchurch	
  City	
  Libraries	
  team	
  in	
  particular	
  had	
  made	
  valiant	
  efforts	
  to	
  

get	
  an	
  archive	
  up	
  and	
  running	
  very	
  quickly	
  after	
  the	
  earthquakes),	
  it	
  was	
  clear	
  the	
  



	
   10	
  

UC	
  team	
  were	
  the	
  only	
  ones	
  in	
  a	
  position	
  to	
  focus	
  their	
  efforts	
  on	
  earthquake-­‐

related	
  content	
  ingestion	
  for	
  years	
  at	
  a	
  time.	
  

5. Heterogeneity:	
  Archival	
  ‘nodes’	
  mushroom	
  in	
  post-­‐disaster	
  situations,	
  due	
  to	
  the	
  

ubiquity	
  of	
  web	
  technologies	
  and	
  the	
  ease	
  with	
  which	
  simple	
  archives	
  can	
  be	
  

established	
  with	
  products	
  like	
  Wordpress.	
  Rather	
  than	
  impose	
  uniform	
  standards	
  

and	
  technologies	
  that	
  would	
  have	
  stifled	
  the	
  development	
  of	
  new	
  archives	
  and	
  left	
  

many	
  small	
  community	
  archives	
  outside	
  the	
  federation,	
  it	
  was	
  necessary	
  to	
  embrace	
  

heterogeneity	
  and	
  design	
  a	
  solution	
  that	
  could	
  cope	
  with	
  a	
  broad	
  range	
  of	
  

technologies.	
  

6. Extensibility:	
  The	
  development	
  of	
  an	
  open	
  dataset	
  allowed	
  for	
  the	
  possibility	
  of	
  

myriad	
  new	
  sites	
  and	
  applications	
  as	
  the	
  archival	
  ecosystem	
  developed.	
  This	
  was	
  

embraced	
  and	
  undertakings	
  were	
  made	
  to	
  encourage	
  the	
  development	
  of	
  widgets,	
  

mobile	
  applications	
  and	
  satellite	
  sites	
  in	
  order	
  to	
  broaden	
  the	
  reach	
  of	
  the	
  

Consortium	
  as	
  far	
  as	
  possible.	
  

7. Leveraging	
  existing	
  assets:	
  It	
  was	
  made	
  clear	
  that	
  UC	
  CEISMIC	
  was	
  a	
  national	
  project	
  

rather	
  than	
  a	
  regional	
  or	
  University	
  one.	
  The	
  post-­‐disaster	
  situation,	
  involving	
  

significant	
  loss	
  of	
  life	
  and	
  a	
  devastated	
  city,	
  meant	
  that	
  there	
  was	
  no	
  room	
  for	
  

partisan	
  politics.	
  From	
  the	
  outset	
  it	
  was	
  understood	
  that	
  the	
  programme	
  would	
  

leverage	
  existing	
  national	
  digital	
  infrastructure	
  as	
  much	
  as	
  possible.	
  There	
  was	
  no	
  

reason	
  to	
  spend	
  money	
  duplicating	
  existing	
  solutions	
  or	
  services	
  when	
  the	
  

earthquake	
  had	
  apparently	
  already	
  put	
  a	
  $10-­‐15	
  billion	
  hole	
  in	
  the	
  economy.	
  

8. Data	
  consistency:	
  The	
  development	
  of	
  federated	
  archives	
  requires	
  attention	
  to	
  data	
  

consistency,	
  to	
  aid	
  metadata	
  aggregation	
  and	
  facilitate	
  longitudinal	
  and	
  

computationally	
  intensive	
  research.	
  Metadata	
  consistency	
  was	
  relatively	
  low	
  across	
  

the	
  existing	
  archives,	
  despite	
  basic	
  adherence	
  to	
  Dublin	
  Core	
  essentials.	
  The	
  UC	
  



	
   11	
  

team	
  undertook	
  to	
  work	
  with	
  member	
  archives	
  to	
  improve	
  their	
  metadata	
  if	
  

necessary.	
  	
  

9. Data	
  Openness:	
  For	
  Phase	
  1	
  of	
  the	
  UC	
  project	
  (the	
  initial	
  build	
  and	
  deployment	
  of	
  

ceismic.org.nz	
  and	
  quakestudies.canterbury.ac.nz)	
  online	
  services	
  like	
  Facebook	
  and	
  

Twitter	
  wouldn’t	
  be	
  archived.	
  These	
  services	
  were	
  very	
  useful	
  in	
  the	
  post-­‐disaster	
  

context,	
  but	
  pose	
  difficulties	
  for	
  long-­‐term	
  preservation.	
  It	
  was	
  felt	
  better	
  to	
  deal	
  

with	
  the	
  basics	
  first	
  and	
  consider	
  social	
  media	
  later.	
  

10. Geo-­‐referencing:	
  Damage	
  caused	
  by	
  earthquakes	
  tends	
  to	
  be	
  associated	
  mainly	
  with	
  

built	
  environments	
  such	
  as	
  buildings	
  and	
  houses.	
  Indeed,	
  a	
  significant	
  proportion	
  of	
  

the	
  damage	
  to	
  Christchurch	
  was	
  in	
  the	
  Central	
  Business	
  District	
  (CBD)	
  and	
  suburbs	
  

hit	
  by	
  severe	
  liquefaction.	
  Although	
  time-­‐consuming,	
  it	
  was	
  agreed	
  that	
  efforts	
  

should	
  be	
  made	
  to	
  geo-­‐reference	
  as	
  much	
  content	
  as	
  possible	
  to	
  enable	
  

implementation	
  of	
  map-­‐based	
  discovery	
  tools.	
  

11. Linked	
  Open	
  Data:	
  Wherever	
  possible	
  design	
  efforts	
  would	
  enable	
  participation	
  in	
  

the	
  world	
  of	
  Linked	
  Open	
  Data	
  (LOD).	
  Because	
  many	
  of	
  the	
  ‘nodes’	
  in	
  the	
  federation	
  

were	
  already	
  established,	
  and	
  not	
  capable	
  of	
  LOD,	
  most	
  of	
  the	
  efforts	
  in	
  this	
  regard	
  

were	
  directed	
  towards	
  the	
  University	
  of	
  Canterbury’s	
  new	
  QuakeStudies	
  repository.	
  

None	
  of	
  these	
  principles	
  were	
  particularly	
  challenging	
  to	
  workshop	
  participants.	
  Conversely,	
  

they	
  represented	
  a	
  set	
  of	
  principles	
  –	
  or	
  a	
  common	
  language	
  -­‐	
  held	
  in	
  common	
  across	
  the	
  IT,	
  

cultural	
  heritage,	
  and	
  digital	
  humanities	
  worlds	
  that	
  bonded	
  the	
  group.	
  The	
  biggest	
  concern	
  

of	
  many	
  participants	
  before	
  the	
  workshop	
  was	
  the	
  possibility	
  that	
  one	
  or	
  more	
  participants	
  

would	
  not	
  be	
  aware	
  of	
  these	
  common	
  expectations	
  and	
  would	
  demand,	
  for	
  instance,	
  a	
  

gated	
  archive	
  using	
  proprietary	
  technologies	
  that	
  would	
  undermine	
  the	
  smaller,	
  more	
  

vulnerable	
  archives	
  in	
  the	
  Consortium.	
  Some	
  participants	
  worried	
  that	
  the	
  University,	
  in	
  

particular,	
  would	
  take	
  a	
  closed	
  approach	
  to	
  data	
  acquisition	
  and	
  sharing;	
  the	
  communication	
  

of	
  common	
  standards,	
  and	
  digital	
  humanities	
  principles	
  of	
  openness	
  and	
  sharing	
  went	
  a	
  long	
  



	
   12	
  

way	
  to	
  allaying	
  fears	
  and	
  allowing	
  development	
  to	
  proceed.	
  As	
  Linda	
  Barwick	
  has	
  noted,	
  

distributed	
  systems	
  like	
  UC	
  CEISMIC	
  “can	
  only	
  work	
  into	
  the	
  longer	
  term	
  if	
  they	
  are	
  built	
  on	
  

shared	
  standards,	
  formats,	
  and	
  procedures,	
  designed	
  for	
  long-­‐term	
  viability”.35	
  	
  

12. 	
  Architecture	
  

UC	
  CEISMIC	
  relies	
  on	
  two	
  main	
  assets:	
  the	
  UC	
  CEISMIC	
  Federation,	
  which	
  provides	
  a	
  

website,	
  metadata	
  aggregation	
  services,	
  and	
  federated	
  archive	
  comprised	
  of	
  over	
  10	
  

‘nodes’,	
  and	
  UC	
  QuakeStudies,	
  a	
  bespoke	
  repository	
  tailored	
  to	
  the	
  collection	
  of	
  cultural	
  

heritage	
  content	
  and	
  research	
  data.	
  Either	
  of	
  these	
  assets	
  could	
  be	
  deployed	
  individually,	
  

but	
  together	
  they	
  offer	
  a	
  wide-­‐ranging	
  solution	
  to	
  myriad	
  issues.	
  The	
  products	
  were	
  built	
  

concurrently	
  over	
  a	
  period	
  of	
  18	
  months	
  by	
  NV	
  Interactive	
  and	
  CWA	
  Media	
  /	
  Learning	
  Media	
  

Limited	
  respectively.	
  The	
  University	
  of	
  Canterbury	
  Digital	
  Humanities	
  Programme	
  led	
  the	
  

design	
  and	
  build	
  of	
  both	
  systems,	
  from	
  procurement	
  to	
  requirements	
  definition,	
  solution	
  

design,	
  development,	
  testing,	
  and	
  deployment.	
  Overall	
  project	
  management	
  lay	
  with	
  UC,	
  

who	
  also	
  held	
  the	
  responsibility	
  for	
  final	
  decisions	
  on	
  technical	
  matters.	
  That	
  said,	
  it	
  should	
  

be	
  made	
  clear	
  that	
  external	
  vendors	
  and	
  professional	
  software	
  developers	
  built	
  both	
  assets.	
  

QuakeStudies	
  in	
  particular	
  was	
  treated	
  as	
  a	
  major	
  enterprise	
  build,	
  involving	
  a	
  vendor-­‐side	
  

project	
  manager,	
  solution	
  architect,	
  front	
  and	
  backend	
  developers,	
  testers,	
  graphic	
  

designers	
  and	
  system	
  administrators.	
  	
  



	
   13	
  

i. UC	
  CEISMIC	
  Federation	
  

	
  

Figure	
  1:	
  UC	
  CEISMIC	
  Federation	
  Architecture	
  

The	
  UC	
  CEISMIC	
  federation	
  is	
  comprised	
  of	
  three	
  separate	
  layers.	
  At	
  base,	
  a	
  Memorandum	
  

of	
  Understanding	
  bonds	
  it,	
  with	
  signatories	
  agreeing	
  to	
  “ensure	
  that	
  digital	
  content	
  deemed	
  

appropriate	
  by	
  them	
  and	
  the	
  [Programme]	
  Board	
  is	
  made	
  available	
  to	
  users	
  of	
  [UC	
  

CEISMIC]”.36	
  Effectively,	
  this	
  means	
  that	
  Consortium	
  members	
  undertake	
  to	
  make	
  all	
  their	
  

digital	
  holdings	
  associated	
  with	
  the	
  Canterbury	
  earthquakes	
  available	
  to	
  the	
  broader	
  

federation	
  via	
  ceismic.org.nz	
  or	
  any	
  other	
  Consortium-­‐related	
  sites	
  that	
  might	
  appear.	
  For	
  

most	
  of	
  the	
  members	
  this	
  wasn’t	
  an	
  issue:	
  they	
  are	
  mandated	
  by	
  government	
  to	
  collect	
  

digital	
  material	
  related	
  to	
  significant	
  New	
  Zealand	
  events,	
  and	
  make	
  them	
  publically	
  

available.	
  Some	
  organizations	
  had	
  less	
  developed	
  digital	
  infrastructure,	
  or	
  no	
  digital	
  

infrastructure	
  at	
  all,	
  and	
  would	
  either	
  share	
  other	
  members’	
  infrastructure	
  or	
  develop	
  their	
  

own	
  (some	
  agencies	
  had	
  plans	
  for	
  implementing	
  digital	
  archives	
  but	
  had	
  not	
  yet	
  gone	
  live	
  

with	
  them).	
  At	
  go-­‐live	
  3	
  providers	
  were	
  able	
  to	
  deliver	
  content,	
  with	
  others	
  coming	
  online	
  

progressively	
  as	
  resources	
  allowed.	
  The	
  great	
  benefit	
  of	
  this	
  approach,	
  of	
  course,	
  is	
  that	
  



	
   14	
  

nothing	
  else	
  needs	
  to	
  be	
  done	
  after	
  the	
  archives	
  are	
  ‘plugged	
  into’	
  the	
  broader	
  federation:	
  

business	
  as	
  usual	
  practices	
  mean	
  that	
  contributions	
  to	
  UC	
  CEISMIC	
  will	
  grow	
  automatically	
  

as	
  federation	
  harvesting	
  proceeds.	
  The	
  same	
  is	
  the	
  case	
  for	
  any	
  community	
  sites	
  added	
  to	
  

the	
  federation.	
  At	
  the	
  time	
  of	
  writing,	
  with	
  the	
  system	
  operationally	
  complete	
  and	
  an	
  

archive	
  ‘seed’	
  completed,	
  the	
  archive	
  includes	
  over	
  60,000	
  items	
  contributed	
  by	
  over	
  12	
  

providers.	
  Projections	
  indicate	
  the	
  archive	
  will	
  hold	
  upwards	
  of	
  100,000	
  items	
  by	
  the	
  end	
  of	
  

2013,	
  contributed	
  by	
  many	
  different	
  content	
  providers.	
  The	
  expectation	
  is	
  that	
  year	
  on	
  year	
  

increases,	
  over	
  the	
  course	
  of	
  the	
  10	
  –	
  15	
  year	
  rebuild	
  process,	
  will	
  result	
  in	
  a	
  significant	
  

cultural	
  asset.	
  

	
   The	
  mechanism	
  for	
  metadata	
  aggregation	
  across	
  the	
  federation	
  was	
  initially	
  a	
  

technical	
  concern,	
  and	
  would	
  be	
  a	
  significant	
  challenge	
  for	
  any	
  group	
  intending	
  to	
  

implement	
  the	
  UC	
  CEISMIC	
  model	
  outside	
  New	
  Zealand.37	
  As	
  indicated	
  below,	
  a	
  decision	
  had	
  

been	
  made	
  relatively	
  early	
  to	
  use	
  Fedora	
  Commons	
  as	
  the	
  backend	
  repository	
  for	
  the	
  UC	
  

QuakeStudies	
  repository,	
  in	
  part	
  because	
  it	
  has	
  native	
  ability	
  to	
  act	
  as	
  a	
  metadata	
  

aggregation	
  point:	
  QuakeStudies	
  would	
  aggregate	
  content	
  from	
  the	
  federation	
  members,	
  

and	
  ceismic.org.nz	
  search	
  would	
  be	
  powered	
  by	
  API	
  queries	
  from	
  its	
  backend.	
  This	
  sounds	
  

straightforward	
  enough,	
  but	
  it	
  would	
  have	
  involved	
  significant	
  overhead.	
  Federation	
  

members	
  have	
  a	
  broad	
  variety	
  of	
  metadata	
  standards	
  (ranging	
  from	
  national	
  library	
  and	
  

museum	
  quality	
  to	
  the	
  barest	
  Dublin	
  Core	
  fields	
  on	
  a	
  Wordpress	
  instance),	
  which	
  would	
  

need	
  to	
  be	
  massaged	
  and	
  mapped	
  to	
  a	
  common	
  standard.	
  Legal	
  and	
  policy	
  issues	
  would	
  also	
  

have	
  been	
  a	
  significant	
  issue,	
  as	
  the	
  small	
  UC	
  CEISMIC	
  programme	
  team	
  would	
  have	
  needed	
  

to	
  administer	
  content	
  agreements	
  and	
  assure	
  the	
  University	
  legal	
  team	
  that	
  the	
  Terms	
  and	
  

Conditions	
  were	
  both	
  robust	
  and	
  enforceable.	
  As	
  it	
  happens,	
  QuakeStudies	
  is	
  only	
  likely	
  to	
  

become	
  a	
  metadata	
  aggregation	
  point	
  for	
  archives	
  within	
  the	
  University	
  of	
  Canterbury,	
  or	
  

perhaps	
  for	
  other	
  New	
  Zealand	
  universities	
  who	
  would	
  like	
  their	
  content	
  surfaced	
  via	
  UC	
  

QuakeStudies	
  as	
  well	
  as	
  ceismic.org.nz.	
  Such	
  a	
  ‘mini-­‐federation’	
  would	
  offer	
  administrative	
  



	
   15	
  

advantages,	
  in	
  that	
  research-­‐heavy	
  data	
  could	
  be	
  corralled	
  into	
  a	
  single	
  repository	
  that	
  has	
  

easy	
  access	
  to	
  high	
  performance	
  computing	
  services	
  (see	
  below).	
  

	
   As	
  it	
  happens,	
  New	
  Zealand	
  has	
  an	
  existing	
  metadata	
  aggregation	
  service	
  based	
  in	
  

the	
  National	
  Library,	
  known	
  as	
  DigitalNZ.	
  Although	
  outside	
  the	
  earliest	
  technical	
  discussions,	
  

this	
  team	
  were	
  sent	
  the	
  Detailed	
  Requirements	
  for	
  the	
  QuakeStudies	
  archive	
  early	
  in	
  the	
  

design	
  process,	
  which	
  included	
  an	
  ontology	
  based	
  on	
  their	
  own	
  metadata	
  schema	
  (itself	
  

based	
  on	
  Dublin	
  Core),	
  to	
  enhance	
  interoperability.	
  It	
  quickly	
  became	
  apparent	
  that	
  not	
  only	
  

was	
  the	
  UC	
  CEISMIC	
  project	
  well	
  aligned	
  to	
  their	
  strategic	
  direction,	
  they	
  were	
  already	
  

aggregating	
  content	
  from	
  several	
  of	
  the	
  federation	
  members	
  and,	
  as	
  a	
  government	
  agency,	
  

had	
  robust	
  policies	
  and	
  procedures	
  that	
  would	
  enable	
  the	
  UC	
  team	
  to	
  completely	
  outsource	
  

metadata	
  aggregation	
  to	
  them.	
  In	
  some	
  senses	
  this	
  is	
  the	
  great	
  ‘cheat’	
  of	
  the	
  UC	
  CEISMIC	
  

programme:	
  a	
  core	
  architectural	
  component,	
  crucial	
  to	
  the	
  integrity	
  of	
  the	
  entire	
  system	
  and	
  

difficult	
  to	
  implement,	
  was	
  outsourced	
  to	
  a	
  government	
  unit.	
  And	
  yet,	
  it	
  was	
  only	
  the	
  

underlying	
  governing	
  principles	
  and	
  clear	
  articulation	
  of	
  the	
  proposed	
  technical	
  architecture	
  

that	
  made	
  this	
  possible.	
  Open	
  architectures,	
  collaboration,	
  leveraging	
  existing	
  assets,	
  

working	
  for	
  the	
  public	
  good	
  –	
  all	
  these	
  attitudes	
  combined	
  to	
  point	
  to	
  DigitalNZ.	
  Although	
  it	
  

might	
  seem	
  an	
  obvious	
  choice,	
  it	
  would	
  not	
  have	
  been	
  possible	
  if	
  the	
  project	
  had	
  been	
  

constituted	
  in	
  a	
  different	
  way,	
  or	
  been	
  based	
  on	
  different	
  attitudes:	
  constant	
  recourse	
  to	
  

principles	
  of	
  openness	
  and	
  sharing	
  offered	
  significant	
  benefits,	
  but	
  was	
  in	
  the	
  final	
  analysis	
  a	
  

choice	
  that	
  could	
  have	
  been	
  made	
  differently.	
  	
  

	
   DigitalNZ	
  is	
  a	
  beguilingly	
  simple	
  service.	
  Established	
  in	
  2008,	
  and	
  now	
  a	
  business	
  unit	
  

of	
  the	
  National	
  Library	
  of	
  New	
  Zealand,	
  the	
  service	
  uses	
  SOLR	
  to	
  aggregate	
  metadata	
  from	
  

hundreds	
  of	
  content	
  providers	
  across	
  New	
  Zealand	
  and	
  expose	
  the	
  metadata	
  through	
  a	
  

Ruby	
  on	
  Rails	
  CMS,	
  and	
  simple	
  API.	
  The	
  platform	
  is	
  capable	
  of	
  harvesting	
  from	
  a	
  variety	
  of	
  

formats	
  (OAI-­‐PMH,	
  API,	
  RSS),	
  which	
  are	
  then	
  mapped	
  to	
  the	
  DigitalNZ	
  metadata	
  schema.	
  At	
  



	
   16	
  

the	
  time	
  of	
  writing	
  the	
  service	
  aggregates	
  25	
  million	
  records	
  from	
  more	
  than	
  120	
  

providers.38	
  It	
  has	
  recently	
  been	
  used	
  as	
  the	
  search	
  engine	
  for	
  the	
  upgraded	
  National	
  Library	
  

website.39	
  The	
  key	
  point	
  in	
  terms	
  of	
  UC	
  CEISMIC,	
  of	
  course,	
  is	
  that	
  all	
  content	
  in	
  the	
  

federation	
  is	
  aggregated	
  upstream	
  in	
  DigitalNZ,40	
  and	
  available	
  at	
  both	
  digitalnz.org	
  and	
  via	
  

the	
  DigitalNZ	
  API,	
  providing	
  a	
  radically	
  open	
  data	
  model.	
  In	
  some	
  ways,	
  although	
  it	
  makes	
  

sense	
  to	
  offer	
  the	
  public	
  a	
  ‘front	
  window’	
  to	
  the	
  UC	
  CEISMIC	
  collection	
  at	
  ceismic.org.nz	
  for	
  

ease	
  of	
  use	
  and	
  to	
  ensure	
  the	
  programme	
  has	
  a	
  solid	
  web	
  presence,	
  there	
  is	
  no	
  technical	
  

reason	
  to	
  do	
  this	
  –	
  the	
  content	
  is	
  available	
  at	
  DigitalNZ	
  anyway.	
  It	
  should	
  also	
  be	
  noted	
  that,	
  

depending	
  on	
  the	
  individual	
  licensing	
  restrictions	
  placed	
  on	
  content,	
  users	
  are	
  free	
  to	
  

mashup,	
  remix,	
  and	
  reuse	
  UC	
  CEISMIC	
  content	
  using	
  the	
  DigitalNZ	
  API	
  as	
  well.	
  

	
   This	
  open	
  architecture	
  offers	
  significant	
  advantages	
  in	
  the	
  context	
  of	
  Web	
  2.0	
  and	
  

the	
  movement	
  towards	
  a	
  mobile	
  web.	
  In	
  some	
  senses,	
  the	
  UC	
  CEISMIC	
  archive	
  is	
  conceived	
  

as	
  post-­‐website	
  service.	
  The	
  dataset	
  itself,	
  and	
  the	
  API	
  that	
  exposes	
  it,	
  are	
  the	
  essential	
  

components;	
  multiple	
  access	
  points,	
  or	
  channels,	
  can	
  be	
  developed	
  as	
  need	
  or	
  interest	
  

arises.	
  ceismic.org.nz,	
  although	
  highly	
  functional,	
  is	
  merely	
  a	
  Microsoft	
  Umbraco	
  CMS	
  with	
  a	
  

search	
  function	
  that	
  queries	
  the	
  DigitalNZ	
  API;	
  a	
  light-­‐weight	
  front-­‐end	
  to	
  a	
  highly	
  

distributed	
  data	
  architecture.	
  In	
  some	
  ways,	
  the	
  success	
  of	
  the	
  programme	
  will	
  be	
  

determined	
  by	
  the	
  number	
  of	
  channels	
  that	
  are	
  built	
  to	
  expose	
  the	
  archival	
  content:	
  the	
  

more	
  that	
  exist,	
  the	
  more	
  uses	
  the	
  content	
  is	
  being	
  put	
  to.	
  The	
  first	
  mobile	
  application	
  –	
  a	
  

Windows	
  8	
  ‘Metro’	
  application	
  that	
  was	
  a	
  winner	
  in	
  the	
  Microsoft	
  New	
  Zealand	
  Humanising	
  

Data	
  competition41	
  -­‐	
  has	
  been	
  released	
  and	
  more	
  are	
  expected	
  to	
  follow.	
  To	
  a	
  similar	
  end	
  

Lincoln	
  University	
  Applied	
  Computing	
  have	
  produced	
  an	
  HTML	
  web	
  framework	
  that	
  includes	
  

basic	
  search	
  and	
  authentication	
  to	
  the	
  UC	
  CEISMIC	
  collection	
  in	
  the	
  DigitalNZ	
  API.	
  The	
  

framework	
  is	
  designed	
  to	
  facilitate	
  the	
  creation	
  of	
  ‘satellite	
  sites’	
  that	
  can	
  be	
  set	
  up	
  to	
  

showcase	
  particular	
  collections	
  in	
  the	
  archive.	
  Some	
  major	
  content	
  providers,	
  for	
  instance,	
  

might	
  want	
  to	
  showcase	
  their	
  earthquake-­‐related	
  content	
  and	
  will	
  be	
  able	
  to	
  do	
  so	
  using	
  the	
  



	
   17	
  

satellite	
  site	
  framework.	
  It	
  is	
  one	
  more	
  way	
  to	
  open	
  up	
  multiple	
  channels	
  to	
  the	
  content,	
  

making	
  it	
  more	
  accessible	
  and,	
  hopefully,	
  more	
  widely	
  used.	
  More	
  importantly,	
  it	
  provides	
  a	
  

way	
  to	
  encourage	
  the	
  development	
  of	
  “survivor	
  rites”,42	
  those	
  various	
  cultural	
  expressions	
  

that	
  bond	
  post-­‐disaster	
  communities	
  and	
  aid	
  in	
  the	
  recovery,	
  rebuild	
  and	
  memorialization	
  

process.	
  If	
  there	
  is	
  one	
  ‘design	
  principle’	
  that	
  disaster-­‐related	
  DH	
  projects	
  should	
  consider	
  

aiming	
  to	
  support,	
  it	
  is	
  surely	
  this.	
  

ii. UC	
  QuakeStudies	
  

	
  

Figure	
  2:	
  UC	
  QuakeStudies	
  Architecture	
  

The	
  UC	
  QuakeStudies	
  repository	
  can	
  also	
  be	
  described	
  in	
  terms	
  of	
  three	
  layers,	
  although	
  as	
  a	
  

single	
  system,	
  rather	
  than	
  a	
  broad	
  federation	
  of	
  systems.	
  Following	
  the	
  production	
  of	
  high	
  

level	
  and	
  detailed	
  requirements,	
  a	
  solution	
  options	
  process	
  chose	
  Drupal	
  and	
  Fedora	
  



	
   18	
  

Commons	
  as	
  the	
  preferred	
  components.	
  Omeka	
  was	
  considered	
  as	
  both	
  a	
  standalone	
  

solution	
  and	
  as	
  a	
  front-­‐end	
  to	
  Fedora	
  Commons,	
  but	
  Drupal	
  and	
  Fedora	
  Commons	
  were	
  

chosen	
  as	
  more	
  ubiquitous	
  technologies,	
  with	
  larger	
  open	
  source	
  communities	
  and	
  

therefore	
  more	
  opportunity	
  for	
  technical	
  support.	
  Fedora	
  Commons	
  fitted	
  particularly	
  well	
  

with	
  requirements	
  for	
  metadata	
  aggregation,	
  native	
  RDF	
  indexing	
  and	
  API,	
  and	
  the	
  ability	
  to	
  

use	
  no-­‐SQL	
  databases.	
  Islandora	
  was	
  initially	
  used,	
  based	
  on	
  the	
  assumption	
  that	
  it	
  would	
  

provide	
  a	
  good	
  base	
  for	
  further	
  development	
  and	
  a	
  desire	
  to	
  contribute	
  to	
  an	
  excellent	
  open	
  

source	
  project,	
  but	
  it	
  was	
  abandoned	
  relatively	
  early	
  on	
  in	
  favor	
  of	
  a	
  ‘clean’	
  development	
  

platform	
  of	
  Drupal	
  7	
  and	
  Fedora	
  Commons,	
  connected	
  via	
  a	
  heavily	
  customized	
  Drupal	
  7	
  

REST	
  API	
  module.43	
  This	
  created	
  significant	
  development	
  overhead	
  and	
  introduced	
  a	
  higher	
  

degree	
  of	
  risk	
  but,	
  in	
  probably	
  the	
  biggest	
  decision	
  made	
  in	
  the	
  build	
  as	
  a	
  whole,	
  it	
  was	
  

decided	
  that	
  the	
  benefits	
  of	
  being	
  able	
  to	
  use	
  Drupal	
  7	
  and	
  code	
  its	
  connection	
  to	
  Fedora	
  

Commons	
  afresh	
  out-­‐weighed	
  those	
  risks.	
  The	
  end	
  result	
  is	
  a	
  system	
  that	
  could	
  be	
  open-­‐

sourced	
  and	
  added	
  as	
  an	
  option	
  alongside	
  products	
  like	
  Omeka,	
  Ushahidi,	
  and	
  Islandora.	
  	
  

	
   UC	
  QuakeStudies	
  is	
  a	
  fairly	
  straight-­‐forward	
  web	
  application	
  and	
  archiving	
  system,	
  

but	
  its	
  architecture	
  needs	
  to	
  be	
  described	
  ‘in	
  the	
  round’.	
  The	
  programme	
  has	
  three	
  

environments	
  in	
  total:	
  development,	
  testing	
  and	
  production.	
  Production	
  sits	
  on	
  two	
  

virtualized	
  RHEL	
  servers	
  with	
  a	
  total	
  of	
  10	
  CPUs	
  and	
  16gb	
  of	
  RAM,	
  themselves	
  located	
  on	
  the	
  

university’s	
  main	
  SAN.44	
  Disk	
  size	
  is	
  in	
  the	
  terabyte	
  range	
  but	
  can	
  (and	
  will	
  need	
  to)	
  scale	
  to	
  

whatever	
  is	
  required.	
  Most	
  of	
  the	
  compute	
  power	
  is	
  deployed	
  on	
  the	
  Tomcat	
  server	
  running	
  

Fedora	
  Commons,	
  and	
  is	
  required	
  to	
  support	
  Java	
  processes	
  triggered	
  by	
  requests	
  from	
  the	
  

Drupal	
  frontend.	
  Drupal	
  application	
  caching	
  reduces	
  overhead	
  significantly,	
  but	
  the	
  

combination	
  of	
  Drupal	
  and	
  Fedora	
  Commons	
  is	
  a	
  heavyweight	
  solution;	
  cosmetic	
  

improvements	
  are	
  planned	
  for	
  the	
  UI,	
  but	
  it	
  is	
  unlikely	
  to	
  ever	
  provide	
  a	
  highly	
  responsive	
  

user	
  experience.	
  This	
  was	
  a	
  considered	
  decision	
  in	
  light	
  of	
  predicted	
  capacity	
  requirements:	
  

the	
  system	
  would	
  need	
  an	
  architectural	
  upgrade	
  to	
  achieve	
  it,	
  but	
  could	
  theoretically	
  scale	
  



	
   19	
  

to	
  hold	
  10	
  million	
  objects.	
  It	
  should	
  also	
  be	
  noted	
  that	
  the	
  ‘post-­‐website’	
  approach	
  to	
  the	
  

project	
  as	
  a	
  whole	
  meant	
  that	
  the	
  Fedora	
  Commons	
  Resource	
  Index	
  /	
  API	
  and	
  OAI-­‐PMH	
  feed	
  

are	
  the	
  most	
  important	
  components	
  in	
  the	
  application	
  stack:	
  as	
  long	
  as	
  these	
  are	
  live	
  and	
  

able	
  to	
  feed	
  DigitalNZ,	
  and	
  from	
  there	
  ceismic.org.nz,	
  mobile	
  apps	
  and	
  satellite	
  sites,	
  the	
  

broader	
  ‘ecosystem’	
  will	
  be	
  healthy.	
  The	
  QuakeStudies	
  front-­‐end	
  could	
  in	
  fact	
  be	
  removed	
  

entirely,	
  with	
  little	
  impact	
  on	
  broader	
  operations	
  beyond	
  a	
  loss	
  of	
  the	
  full-­‐text	
  and	
  advanced	
  

search	
  functions	
  provided	
  by	
  the	
  QuakeStudies	
  SOLR	
  component.	
  Although	
  not	
  entirely	
  

desirable,	
  this	
  provides	
  an	
  excellent	
  low-­‐cost	
  option	
  should	
  maintenance	
  costs	
  for	
  the	
  

Drupal	
  component	
  become	
  unsupportable.	
  It	
  may	
  be	
  that	
  in	
  twenty	
  years	
  time,	
  when	
  

content	
  ingestion	
  is	
  no	
  longer	
  a	
  priority,	
  all	
  UC	
  CEISMIC	
  content	
  is	
  surfaced	
  via	
  

ceismic.org.nz,	
  with	
  QuakeStudies	
  being	
  scaled	
  back	
  to	
  the	
  bare	
  Fedora	
  Commons	
  data	
  

store.	
  

Cloud	
  options	
  were	
  considered	
  for	
  infrastructure,	
  but	
  using	
  normal	
  operational	
  

infrastructure	
  offers	
  significant	
  benefits.	
  Firstly,	
  it	
  means	
  the	
  UC	
  CEISMIC	
  Programme	
  Office	
  

(the	
  team	
  that	
  runs	
  the	
  archive)	
  is	
  supported	
  by	
  normal	
  UC	
  IT	
  support	
  networks.	
  System	
  

maintenance	
  and	
  upgrades	
  are	
  all	
  submitted	
  to	
  the	
  central	
  University	
  Change	
  Advisory	
  

Board	
  (CAB),	
  and	
  the	
  development	
  expertise	
  paid	
  for	
  through	
  a	
  Service	
  Level	
  Agreement	
  

with	
  the	
  vendor	
  is	
  augmented	
  by	
  system	
  administration	
  and	
  network	
  expertise	
  on	
  campus.	
  

The	
  system	
  also	
  uses	
  the	
  standard	
  University	
  disaster	
  recovery	
  processes,	
  and	
  backups	
  to	
  

tape	
  occur	
  as	
  a	
  part	
  of	
  business	
  as	
  usual.	
  Because	
  the	
  infrastructure	
  is	
  virtualized,	
  additional	
  

storage	
  can	
  be	
  provisioned	
  through	
  a	
  simple	
  university	
  helpdesk	
  request	
  and	
  then	
  deployed	
  

using	
  minimal	
  vendor	
  support.	
  University	
  infrastructure	
  also	
  offers	
  easy	
  access	
  to	
  high	
  

performance	
  computing	
  services	
  at	
  the	
  UC	
  Blue	
  Fern	
  super-­‐computer45	
  and	
  New	
  Zealand’s	
  

national	
  grid	
  computing	
  network	
  (NeSI),46	
  via	
  the	
  KAREN	
  high-­‐speed	
  research	
  network.47	
  It	
  is	
  

worth	
  noting	
  that	
  this	
  choice	
  of	
  infrastructure	
  was	
  not	
  a	
  given,	
  but	
  the	
  result	
  of	
  a	
  carefully	
  

considered	
  solution	
  options	
  assessment	
  specifically	
  related	
  to	
  infrastructure.	
  Although	
  



	
   20	
  

development	
  of	
  the	
  system	
  itself	
  proceeded	
  using	
  an	
  Agile	
  methodology,	
  major	
  decisions	
  

like	
  this	
  involved	
  more	
  formal	
  ‘waterfall-­‐like’	
  methods.	
  It	
  is	
  only	
  now	
  that	
  the	
  system	
  is	
  live	
  

that	
  it	
  has	
  become	
  fully	
  clear	
  just	
  how	
  seriously	
  poor	
  decision-­‐making	
  at	
  crucial	
  moments	
  

would	
  have	
  negatively	
  impacted	
  the	
  project.	
  

In	
  terms	
  of	
  functionality,	
  the	
  Drupal	
  front-­‐end	
  is	
  fairly	
  basic.	
  Its	
  main	
  purpose,	
  

besides	
  offering	
  users	
  browse	
  and	
  search	
  access	
  to	
  the	
  collections	
  held	
  in	
  Fedora	
  Commons,	
  

is	
  as	
  a	
  tool	
  to	
  allow	
  UC	
  CEISMIC	
  administrators	
  to	
  archive	
  items.	
  The	
  ingestion	
  process	
  is	
  

governed	
  by	
  detailed	
  policies	
  and	
  processes	
  developed	
  in	
  consultation	
  with	
  the	
  UC	
  Human	
  

Ethics	
  Committee	
  and	
  the	
  UC	
  CEISMIC	
  Research	
  Committee	
  (composed	
  of	
  senior	
  academics	
  

from	
  all	
  areas	
  of	
  the	
  university,	
  along	
  with	
  additional	
  members	
  from	
  the	
  Otago	
  Medical	
  

School	
  and	
  Massey	
  University).48	
  Administrators	
  can	
  upload	
  items	
  individually,	
  or	
  using	
  a	
  

bulk	
  ingest	
  facility	
  capable	
  of	
  ingesting	
  up	
  to	
  300	
  items	
  an	
  hour	
  with	
  individualized	
  metadata	
  

(for	
  each	
  item	
  if	
  necessary)	
  included	
  on	
  a	
  .csv	
  file.	
  During	
  early	
  requirements	
  definition	
  a	
  lot	
  

of	
  emphasis	
  was	
  placed	
  on	
  the	
  inclusion	
  of	
  high-­‐quality	
  metadata	
  (it	
  remains	
  one	
  of	
  the	
  key	
  

performance	
  indicators	
  for	
  the	
  service),	
  which	
  requires	
  the	
  Programme	
  Office	
  team	
  to	
  spend	
  

considerable	
  time	
  scoping,	
  defining	
  and	
  improving	
  metadata	
  before	
  ingestion.	
  The	
  goal	
  was	
  

never,	
  after	
  all,	
  to	
  try	
  to	
  collect	
  everything	
  produced	
  as	
  a	
  result	
  of	
  the	
  earthquakes	
  but	
  to	
  

curate	
  a	
  large,	
  high-­‐quality	
  archive	
  that	
  would	
  be	
  useful	
  for	
  both	
  the	
  general	
  public,	
  but	
  also	
  

researchers.	
  Done	
  properly,	
  the	
  hope	
  was	
  that	
  the	
  QuakeStudies	
  repository	
  would	
  generate	
  

a	
  series	
  of	
  use	
  cases	
  requiring	
  high	
  performance	
  computing	
  time.	
  Several	
  such	
  use	
  cases	
  

have	
  already	
  been	
  identified.	
  

Metadata	
  requirements	
  were	
  a	
  key	
  focus	
  of	
  the	
  early	
  design	
  efforts,	
  led	
  by	
  the	
  lead	
  

architect,	
  Jason	
  Darwin.	
  A	
  range	
  of	
  options	
  were	
  considered,	
  including	
  Dublin	
  Core,	
  ICOM-­‐

CIDOC,	
  MARC,	
  METS	
  and	
  the	
  emerging	
  “international	
  standard	
  for	
  digital	
  archiving”,	
  49	
  OAIS,	
  

in	
  an	
  attempt	
  to	
  find	
  a	
  commonly	
  accepted	
  standard	
  that	
  could	
  both	
  provide	
  the	
  necessary	
  



	
   21	
  

descriptive	
  elements,	
  and	
  facilitate	
  data	
  sharing.	
  While	
  many	
  projects	
  have	
  difficulty	
  finding	
  

a	
  standard	
  suitable	
  for	
  their	
  specific	
  heritage	
  purposes,	
  50	
  the	
  situation	
  was	
  complicated	
  by	
  

the	
  post-­‐disaster	
  context,	
  which	
  required	
  event-­‐related	
  information	
  rarely	
  required	
  in	
  

business	
  as	
  usual	
  heritage	
  contexts	
  and	
  therefore	
  not	
  included	
  in	
  any	
  of	
  the	
  various	
  cultural	
  

heritage	
  standards.	
  FOAF51	
  was	
  seriously	
  considered	
  as	
  well,	
  but	
  abandoned	
  due	
  to	
  concerns	
  

about	
  ethics	
  and	
  privacy;	
  it	
  was	
  felt	
  that	
  it	
  would	
  be	
  more	
  sensible	
  to	
  develop	
  specific	
  social	
  

media	
  projects	
  that	
  could	
  use	
  FOAF	
  on	
  a	
  case-­‐by-­‐case	
  basis	
  than	
  implementing	
  it	
  as	
  a	
  

system-­‐wide	
  feature.	
  The	
  Fedora	
  Commons	
  Digital	
  Object	
  Model	
  was	
  extremely	
  useful	
  in	
  

this	
  regard,	
  because	
  it	
  offers	
  the	
  ability	
  to	
  connect	
  multiple	
  ontologies	
  (or	
  ‘datastreams’)	
  to	
  

a	
  single	
  digital	
  object.52	
  This	
  allowed	
  the	
  team	
  to	
  implement	
  a	
  single	
  ‘base’	
  ontology,	
  but	
  

offer	
  content	
  providers	
  the	
  opportunity	
  to	
  use	
  a	
  reference	
  standard	
  of	
  their	
  own	
  choice	
  to	
  

describe	
  their	
  collection.	
  With	
  this	
  realization,	
  it	
  became	
  apparent	
  that	
  there	
  was	
  no	
  

pressing	
  need	
  to	
  choose	
  only	
  one	
  of	
  the	
  metadata	
  standards	
  listed	
  above:	
  the	
  key	
  was	
  to	
  

develop	
  a	
  base	
  ontology	
  that	
  would	
  satisfy	
  the	
  immediate	
  requirements	
  of	
  post-­‐earthquake	
  

Christchurch.	
  From	
  there	
  it	
  was	
  a	
  relatively	
  simple	
  step	
  to	
  implement	
  a	
  combination	
  of	
  

Dublin	
  Core	
  (satisfying	
  the	
  basic	
  requirements	
  for	
  international	
  data	
  transfer)	
  and	
  

DigitalNZ’s	
  bespoke	
  standard	
  (itself	
  based	
  largely	
  on	
  Dublin	
  Core),	
  making	
  use	
  of	
  their	
  

aggregation	
  service	
  considerably	
  easier.	
  Additional	
  standards	
  would	
  be	
  attached	
  as	
  

subsidiary	
  datastreams	
  when	
  required.	
  It	
  was	
  this	
  decision,	
  when	
  communicated	
  to	
  

DigitalNZ,	
  which	
  led	
  to	
  the	
  use	
  of	
  DigitalNZ	
  as	
  the	
  aggregation	
  point	
  for	
  the	
  entire	
  UC	
  

CEISMIC	
  federation:	
  a	
  case	
  of	
  internal	
  project	
  decisions	
  aligning	
  well	
  to	
  external	
  service	
  

options.	
  

The	
  design	
  focus	
  placed	
  on	
  metadata	
  was	
  related	
  to	
  a	
  significant	
  long-­‐term	
  goal	
  to	
  

create	
  a	
  dataset	
  conducive	
  to	
  analysis	
  by	
  high	
  performance	
  computers	
  (HPCs).	
  While	
  in	
  

some	
  senses	
  a	
  lack	
  of	
  structured	
  data	
  would	
  have	
  offered	
  more	
  use	
  cases	
  for	
  HPCs,	
  which	
  

would	
  have	
  been	
  required	
  to	
  derive	
  structure	
  and	
  meaning	
  programmatically,	
  it	
  was	
  felt	
  



	
   22	
  

that	
  effort	
  should	
  be	
  directed	
  towards	
  providing	
  a	
  base	
  layer	
  of	
  human-­‐entered	
  metadata	
  to	
  

not	
  only	
  facilitate	
  basic	
  content	
  curation,	
  search,	
  and	
  usability	
  but	
  to	
  act	
  as	
  a	
  control	
  against	
  

future	
  implementations	
  of	
  crowd-­‐sourced	
  and	
  machine-­‐derived	
  metadata:	
  users	
  would	
  be	
  

able	
  to	
  toggle	
  crowd-­‐sourced	
  and	
  machine-­‐derived	
  metadata	
  on	
  or	
  off,	
  giving	
  them	
  the	
  

ability	
  to	
  interpret	
  the	
  content	
  through	
  three	
  different	
  ‘lenses’	
  and	
  three	
  different	
  levels	
  of	
  

reliability.	
  Another	
  requirement	
  called	
  for	
  the	
  implementation	
  of	
  the	
  Resource	
  Description	
  

Framework	
  (RDF)	
  to	
  provide	
  researchers	
  with	
  semantic	
  meaning;	
  this	
  was	
  also	
  implemented	
  

using	
  native	
  Fedora	
  Commons	
  functionality.	
  As	
  can	
  be	
  seen	
  in	
  Figure	
  1,	
  Fedora	
  Commons	
  

includes	
  a	
  REST	
  API	
  that	
  can	
  be	
  used	
  to	
  access	
  content	
  in	
  the	
  backend	
  directly,	
  by-­‐passing	
  

the	
  user-­‐interface.	
  A	
  web	
  service	
  known	
  as	
  Resource	
  Index	
  Search	
  “exposes	
  the	
  

relationships	
  described	
  in	
  the	
  Quake	
  Studies	
  Repository	
  data	
  model,	
  allowing	
  relationships	
  

between	
  the	
  QSR	
  classes	
  to	
  be	
  queried	
  and	
  navigated”;	
  it	
  can	
  be	
  programmatically	
  accessed	
  

via	
  HTTP	
  GET	
  or	
  POST.53	
  Tuples	
  and	
  Triples	
  can	
  be	
  returned	
  using	
  a	
  variety	
  of	
  query	
  

languages,	
  offering	
  the	
  opportunity	
  to	
  engage	
  in	
  rich	
  data	
  analysis	
  and	
  create	
  a	
  broad	
  range	
  

of	
  data	
  visualizations.	
  Although	
  it	
  will	
  be	
  desirable	
  to	
  use	
  computers	
  to	
  programmatically	
  

derive	
  additional	
  meaning	
  (especially	
  given	
  the	
  very	
  large	
  amount	
  of	
  content	
  expected	
  to	
  be	
  

stored	
  and	
  the	
  corresponding	
  difficulty	
  of	
  finding	
  the	
  resources	
  to	
  describe	
  it	
  manually),	
  the	
  

basis	
  for	
  solid	
  semantic	
  analysis	
  is	
  native	
  to	
  the	
  system.	
  

UC	
  QuakeStudies	
  went	
  live	
  as	
  a	
  Beta	
  service	
  on	
  26	
  September	
  2012.	
  Many	
  

requirements	
  (including	
  the	
  provision	
  of	
  a	
  user-­‐friendly	
  API)	
  were	
  either	
  scaled	
  back	
  or	
  

abandoned	
  due	
  to	
  pressure	
  of	
  time	
  and	
  funding.	
  	
  Because	
  of	
  this,	
  at	
  go-­‐live	
  the	
  list	
  of	
  

potential	
  development	
  jobs	
  was	
  long,	
  including	
  improving	
  the	
  user	
  interface,	
  adding	
  crowd-­‐

sourcing	
  functions,	
  using	
  high-­‐performance	
  computers	
  to	
  analyse	
  and	
  improve	
  collections	
  

and	
  metadata,	
  and	
  improving	
  access	
  to	
  the	
  API.	
  Drupal	
  plugins	
  are	
  planned	
  to	
  allow	
  more	
  

sophisticated	
  browsing	
  and	
  searching	
  of	
  archival	
  content	
  based	
  on	
  maps	
  and	
  timelines,	
  the	
  

administrative	
  workflow	
  could	
  be	
  improved,	
  and	
  functionality	
  to	
  allow	
  the	
  addition	
  of	
  



	
   23	
  

crowd-­‐sourced	
  and	
  computer-­‐generated	
  metadata	
  will	
  be	
  implemented	
  if	
  funding	
  allows.	
  

Improvement	
  continues	
  at	
  an	
  infrastructure	
  level,	
  too,	
  with	
  changes	
  being	
  made	
  to	
  the	
  way	
  

the	
  virtual	
  servers	
  connect	
  to	
  the	
  university	
  SAN	
  to	
  improve	
  performance	
  as	
  the	
  archive	
  

scales.	
  UC	
  QuakeStudies	
  is	
  a	
  significant	
  asset	
  to	
  maintain,	
  and	
  the	
  project	
  team’s	
  espousal	
  of	
  

continuous	
  improvement	
  over	
  several	
  decades	
  brings	
  with	
  it	
  considerable	
  overhead,	
  but	
  this	
  

is	
  simply	
  another	
  aspect	
  of	
  the	
  model	
  offered	
  to	
  the	
  digital	
  humanities	
  community.	
  Although	
  

a	
  lot	
  of	
  work,	
  it	
  is	
  possible	
  for	
  a	
  small	
  DH	
  team	
  to	
  design,	
  build,	
  maintain,	
  and	
  improve	
  a	
  

significant	
  enterprise	
  asset	
  (in	
  this	
  case	
  marked	
  as	
  such	
  at	
  a	
  University-­‐wide	
  level).	
  	
  

13. 	
  UC	
  CEISMIC	
  Digital	
  Archive:	
  Current	
  State	
  

Despite	
  its	
  success,	
  the	
  UC	
  CEISMIC	
  Digital	
  Archive	
  is	
  best	
  considered	
  a	
  working	
  ‘proof	
  of	
  

concept’.	
  The	
  key	
  contribution	
  of	
  the	
  project	
  to	
  the	
  digital	
  humanities	
  community	
  is	
  not	
  in	
  

the	
  specific	
  tools	
  used	
  (Drupal,	
  Fedora	
  Commons	
  etc),	
  but	
  the	
  commitment	
  to	
  national	
  

federation	
  as	
  a	
  core	
  principle,	
  and	
  the	
  combination	
  of	
  that	
  with	
  long-­‐term	
  commitment,	
  

community-­‐focused	
  attitudes,	
  and	
  open	
  access	
  principles.	
  Although	
  the	
  technical	
  solution	
  is	
  

elegant	
  enough,	
  and	
  covers	
  a	
  very	
  broad	
  variety	
  of	
  use	
  cases,	
  further	
  development	
  is	
  

required	
  to	
  refine	
  it	
  and	
  (as	
  with	
  most	
  IT	
  projects)	
  the	
  solution	
  would	
  benefit	
  greatly	
  from	
  

additional	
  funding.	
  Much	
  of	
  the	
  value	
  of	
  UC	
  CEISMIC	
  lies	
  in	
  its	
  status	
  as	
  a	
  useful	
  model	
  for	
  

others	
  to	
  consider,	
  and	
  the	
  attitudes,	
  governance	
  mechanisms,	
  processes	
  and	
  policies	
  that	
  

underpin	
  the	
  Consortium	
  and	
  control	
  content	
  curation,	
  ingestion	
  and	
  sharing.	
  Perhaps	
  the	
  

most	
  significant	
  element	
  in	
  the	
  project	
  is	
  the	
  sheer	
  scale	
  envisaged	
  by	
  Project	
  Director	
  Paul	
  

Millar	
  from	
  inception.	
  Unlike	
  other	
  digital	
  humanities	
  disaster	
  archives,	
  that	
  have	
  been	
  

envisaged	
  as	
  scalable	
  but	
  in	
  important	
  ways	
  limited	
  undertakings,	
  UC	
  CEISMIC	
  was	
  

conceived	
  from	
  the	
  outset	
  as	
  a	
  vast	
  all-­‐encompassing	
  archive	
  of	
  national	
  and	
  even	
  

international	
  scope	
  that	
  is	
  intended	
  to	
  keep	
  collecting	
  for	
  as	
  long	
  as	
  funding	
  allows.	
  In	
  some	
  

ways,	
  the	
  federated	
  architecture	
  reflects	
  a	
  need	
  to	
  accommodate	
  this	
  vision;	
  the	
  suitability	
  



	
   24	
  

of	
  this	
  approach	
  for	
  digital	
  preservation	
  in	
  a	
  post-­‐disaster	
  content	
  is	
  merely	
  one	
  happy	
  result	
  

of	
  it.	
  

	
   Lucky	
  coincidences	
  aside,	
  it	
  should	
  be	
  noted	
  that	
  if	
  the	
  goal	
  was	
  only	
  to	
  collect	
  

content	
  in	
  the	
  immediate	
  aftermath	
  of	
  the	
  earthquakes	
  the	
  ‘UC	
  CEISMIC	
  solution’	
  would	
  

have	
  been	
  unwieldy	
  and	
  too	
  slow	
  to	
  implement.	
  The	
  design	
  approach	
  was	
  based	
  on	
  an	
  

assumption	
  that	
  the	
  recovery	
  of	
  the	
  Canterbury	
  region	
  will	
  take	
  decades	
  rather	
  than	
  months	
  

or	
  even	
  years,	
  and	
  that	
  the	
  UC	
  CEISMIC	
  team	
  will	
  need	
  to	
  continue	
  collecting	
  for	
  decades	
  to	
  

come.54	
  While	
  the	
  broad	
  goal	
  of	
  creating	
  a	
  federated	
  resource	
  would	
  suit	
  most	
  situations,	
  

the	
  development	
  of	
  a	
  relatively	
  large	
  bespoke	
  repository	
  to	
  augment	
  such	
  a	
  federation	
  is	
  a	
  

significant	
  undertaking	
  that	
  will	
  produce	
  most	
  rewards	
  over	
  the	
  long	
  term.	
  The	
  expectation	
  

is	
  that	
  hundreds	
  of	
  thousands	
  of	
  items	
  will	
  be	
  ingested	
  into	
  the	
  archive	
  year	
  on	
  year,	
  

creating	
  a	
  dataset	
  capable	
  of	
  providing	
  valuable	
  insights	
  into	
  the	
  nature	
  of	
  disaster	
  risk,	
  

resilience,	
  and	
  renewal.	
  In	
  this	
  model,	
  which	
  could	
  perhaps	
  be	
  more	
  effectively	
  

implemented	
  by	
  governments	
  or	
  communities	
  as	
  part	
  of	
  preparedness	
  programmes	
  before	
  

disasters	
  occur,	
  the	
  goal	
  is	
  to	
  establish	
  a	
  robust	
  and	
  very	
  wide-­‐ranging	
  ‘net’	
  capable	
  of	
  

catching	
  and	
  preserving	
  as	
  much	
  digital	
  content	
  as	
  possible	
  over	
  as	
  long	
  a	
  time	
  period	
  as	
  

possible.	
  	
  

	
   One	
  of	
  the	
  great	
  barriers	
  to	
  success	
  is	
  the	
  identification	
  and	
  storage	
  of	
  what	
  

UNESCO	
  began	
  referring	
  to	
  in	
  1952	
  as	
  “intangible	
  heritage”55,	
  largely	
  due	
  to	
  difficulties	
  

associated	
  with	
  archiving	
  social	
  media	
  services.	
  Although	
  the	
  term	
  initially	
  referred	
  to	
  

folklore,	
  some	
  commentators	
  have	
  pointed	
  out	
  that	
  a	
  focus	
  on	
  ‘intangible	
  heritage’	
  opens	
  

up	
  interesting	
  opportunities	
  in	
  the	
  digital	
  era.	
  Silberman	
  and	
  Purser	
  suggest	
  that	
  in	
  the	
  

future	
  

[t]he	
  task	
  of	
  heritage	
  professionals	
  will	
  be	
  rather	
  to	
  enable	
  contemporary	
  communities	
  to	
  

digitally	
  (re)produce	
  historical	
  environments,	
  collective	
  narratives	
  and	
  geographical	
  



	
   25	
  

visualizations	
  that	
  cluster	
  individual	
  perspectives	
  into	
  shared	
  forms	
  and	
  processes	
  of	
  

remembering.	
  These	
  interactions	
  are	
  reminiscent	
  of	
  the	
  conversations	
  that	
  once	
  occurred	
  

much	
  more	
  frequently	
  at	
  corner	
  bars,	
  in	
  town	
  squares	
  and	
  by	
  evening	
  campfires	
  (cf.	
  Putnam	
  

2001)	
  as	
  a	
  vital	
  part	
  of	
  the	
  exercise	
  of	
  cultural	
  diversity	
  that	
  is	
  now	
  seen	
  as	
  a	
  central	
  

component	
  of	
  world	
  heritage	
  (UNESCO	
  2005).56	
  

Capturing	
  this	
  kind	
  of	
  “performative	
  memory”57	
  is	
  made	
  very	
  difficult	
  by	
  the	
  widespread	
  use	
  

of	
  social	
  media	
  services	
  like	
  Facebook	
  and	
  Twitter.	
  These	
  services	
  are	
  easy	
  to	
  use	
  and	
  

Canterbury	
  people	
  flocked	
  to	
  them	
  after	
  the	
  earthquakes,	
  posting	
  extensive	
  comments	
  on	
  

Facebook	
  and	
  using	
  the	
  #eqnz	
  hashtag	
  to	
  comment	
  on	
  and	
  organize	
  themselves.	
  The	
  

University	
  of	
  Canterbury	
  Student	
  Volunteer	
  Army	
  was	
  almost	
  solely	
  organized	
  around	
  

Facebook,58	
  generating	
  over	
  27,000	
  ‘Likes’,	
  and	
  BeckerFraserPhotos	
  (the	
  photographer	
  of	
  

record)	
  used	
  Facebook	
  to	
  publish	
  and	
  develop	
  a	
  community	
  around	
  their	
  photos,	
  generating	
  

over	
  14,000	
  ‘Likes’.	
  Ideally,	
  content	
  from	
  services	
  like	
  Facebook	
  and	
  Twitter	
  would	
  be	
  

integrated	
  into	
  the	
  UC	
  CEISMIC	
  archive,	
  but	
  it	
  is	
  very	
  difficult	
  to	
  navigate	
  licensing	
  issues	
  and	
  

organize	
  local	
  storage	
  of	
  the	
  content	
  outside	
  those	
  eco-­‐systems.	
  The	
  current	
  approach	
  is	
  

threefold:	
  rely	
  on	
  New	
  Zealand’s	
  National	
  Digital	
  Heritage	
  Archive	
  (NDHA)	
  domain	
  

harvesting	
  process	
  (which	
  takes	
  a	
  complete	
  harvest	
  of	
  the	
  .co.nz	
  web	
  domain	
  and	
  targets	
  

selected	
  important	
  sites	
  for	
  special	
  treatment)	
  and	
  hope	
  resources	
  appear	
  that	
  will	
  allow	
  

integration	
  of	
  that	
  content	
  into	
  the	
  UC	
  CEISMIC	
  archive	
  in	
  years	
  to	
  come;	
  identify	
  

organizations	
  and	
  teams	
  with	
  significant	
  local	
  datasets	
  of	
  content	
  that	
  might	
  be	
  able	
  to	
  be	
  

made	
  available	
  via	
  UC	
  CEISMIC	
  if	
  progress	
  can	
  be	
  made	
  with	
  the	
  relevant	
  companies;	
  hope	
  

that	
  the	
  companies	
  themselves	
  are	
  developing	
  long-­‐term	
  archiving	
  solutions,	
  as	
  is	
  the	
  case	
  

with	
  Twitter	
  and	
  the	
  Library	
  of	
  Congress;	
  and	
  make	
  every	
  effort	
  to	
  contact	
  the	
  relevant	
  

social	
  media	
  companies	
  to	
  request	
  partnering	
  arrangements.	
  Taming	
  these	
  data	
  

‘decahoses’,59	
  and	
  integrating	
  them	
  into	
  a	
  federated	
  archive	
  to	
  facilitate	
  research,	
  analysis	
  

and	
  public	
  memory,	
  is	
  a	
  difficult	
  task.	
  Cornelius	
  Puschmann	
  and	
  Jean	
  Burgess	
  are	
  correct	
  to	
  



	
   26	
  

note	
  that,	
  despite	
  what	
  many	
  people	
  might	
  think,	
  the	
  “owners”	
  of	
  social	
  media	
  data	
  are	
  the	
  

platform	
  providers,	
  not	
  the	
  users.	
  “[W]hile	
  the	
  data	
  in	
  social	
  media	
  platforms	
  is	
  sought	
  after	
  

by	
  companies,	
  governments	
  and	
  scientists,	
  the	
  users	
  who	
  produce	
  it	
  have	
  the	
  least	
  degree	
  

of	
  control	
  over	
  “their”	
  data”.	
  60	
  	
  	
  

While	
  this	
  might	
  not	
  cause	
  many	
  issues	
  for	
  day-­‐to-­‐day	
  use,	
  it	
  has	
  significant	
  

implications	
  for	
  cultural	
  heritage	
  and	
  research	
  purposes	
  because	
  it	
  makes	
  it	
  impossible	
  for	
  

social	
  media	
  users	
  to	
  effectively	
  gift	
  their	
  content	
  to	
  an	
  archive.	
  When	
  people	
  do	
  make	
  an	
  

attempt	
  it	
  soon	
  becomes	
  clear	
  that	
  the	
  only	
  option	
  is	
  to	
  manually	
  download	
  and	
  re-­‐describe	
  

images,	
  videos,	
  comments	
  and	
  suchlike,	
  with	
  an	
  attendant	
  loss	
  of	
  essential	
  context.	
  

Although	
  the	
  services	
  are	
  fundamentally	
  useful	
  in	
  post-­‐disaster	
  contexts,	
  the	
  “rhetoric	
  of	
  

democratization”61	
  that	
  drives	
  them	
  often	
  falls	
  flat	
  when	
  it	
  comes	
  to	
  preserving	
  their	
  

content	
  for	
  posterity	
  and	
  research	
  because	
  the	
  services	
  are	
  oriented	
  towards	
  “findability	
  

rather	
  than	
  preservation”.62	
  This	
  issue	
  is	
  of	
  particular	
  concern	
  if	
  one	
  accepts	
  Scott	
  Lash’s	
  

contention	
  that	
  our	
  lives	
  have	
  come	
  to	
  be	
  not	
  only	
  mediated	
  by	
  information	
  and	
  technology,	
  

but	
  also	
  constituted	
  by	
  them.	
  If	
  this	
  is	
  the	
  case	
  archiving	
  the	
  digital	
  outputs	
  that	
  resulted	
  

from	
  the	
  earthquakes	
  takes	
  on	
  a	
  significant	
  moral	
  imperative.63	
  Despite	
  succeeding	
  in	
  

developing	
  a	
  flexible	
  national	
  system	
  capable	
  of	
  archiving	
  a	
  significant	
  snapshot	
  of	
  the	
  digital	
  

content	
  associated	
  with	
  the	
  Canterbury	
  earthquakes,	
  a	
  vast	
  amount	
  of	
  social	
  media	
  content	
  

is	
  currently	
  not	
  accounted	
  for.	
  It	
  sits	
  in	
  trust	
  with	
  commercial	
  entities	
  whose	
  business	
  drivers	
  

are	
  by	
  no	
  means	
  certain	
  to	
  ensure	
  it	
  will	
  be	
  preserved	
  for	
  the	
  long-­‐term.	
  The	
  problem	
  isn’t	
  

solely	
  related	
  to	
  the	
  commercial	
  nature	
  of	
  many	
  web	
  services,	
  of	
  course.	
  Rather,	
  it	
  is	
  part	
  of	
  

the	
  broader	
  problem	
  of	
  digital	
  preservation.	
  Instant	
  Messages	
  (IMs),	
  emails,	
  Internet	
  Relay	
  

Chats	
  (IRC)	
  and	
  other	
  detritus	
  of	
  the	
  so-­‐called	
  “deep	
  web”64	
  are	
  also	
  likely	
  to	
  be	
  available	
  to	
  

UC	
  CEISMIC	
  for	
  archiving,	
  but	
  archiving	
  it,	
  displaying	
  it	
  to	
  users,	
  and	
  making	
  it	
  findable	
  poses	
  

a	
  different	
  set	
  of	
  problems	
  again.	
  Novel	
  approaches	
  are	
  required	
  to	
  allow	
  archives	
  to	
  store	
  

and	
  describe	
  a	
  wide	
  variety	
  of	
  digital	
  formats,	
  that	
  –	
  in	
  the	
  context	
  of	
  the	
  Canterbury	
  



	
   27	
  

earthquakes	
  –	
  appears	
  to	
  include	
  3D	
  panoramas,	
  Flash,	
  VRML	
  and	
  QTVR	
  objects,	
  CAD	
  files,	
  

GIS	
  mapping	
  data,	
  and	
  LIDAR	
  imagery.	
  As	
  with	
  social	
  media,	
  “[t]he	
  potential	
  of	
  virtual	
  

heritage	
  has	
  been	
  hindered	
  as	
  much	
  by	
  people	
  as	
  by	
  diverse	
  technologies,	
  poor	
  provenance,	
  

and	
  changing	
  systems.”65	
   	
  	
  

14. 	
  Conclusion	
  

Although	
  outlier	
  content	
  such	
  as	
  social	
  media	
  and	
  less	
  ubiquitous	
  file	
  types	
  have	
  yet	
  to	
  be	
  

archived,	
  the	
  UC	
  CEISMIC	
  system	
  holds	
  great	
  promise	
  as	
  a	
  model	
  for	
  post-­‐disaster	
  digital	
  

archiving.	
  The	
  combination	
  of	
  national	
  federated	
  archive	
  and	
  a	
  bespoke	
  archive	
  designed	
  for	
  

the	
  ingestion	
  of	
  research-­‐oriented	
  content	
  has	
  proved	
  to	
  be	
  a	
  powerful	
  combination,	
  and	
  

cemented	
  a	
  broad	
  community	
  of	
  content	
  providers	
  reaching	
  from	
  local	
  community	
  sites	
  to	
  

the	
  largest	
  national	
  archives.	
  The	
  success	
  of	
  the	
  broader	
  programme	
  has	
  resulted	
  directly	
  

from	
  its	
  insistence	
  on	
  the	
  use	
  of	
  open	
  source	
  tools	
  and	
  the	
  implementation	
  of	
  open	
  access	
  

policies.	
  



	
   28	
  

	
  

Funding	
  	
  
This	
  phase	
  of	
  the	
  UC	
  CEISMIC	
  project	
  was	
  supported	
  by	
  funding	
  from	
  the	
  University	
  of	
  
Canterbury,	
  SysDoc,	
  the	
  Canterbury	
  Community	
  Trust,	
  the	
  Christchurch	
  Press,	
  the	
  (London)	
  
Christchurch	
  Earthquake	
  Appeal,	
  InternetNZ,	
  and	
  the	
  Malaysian	
  Chamber	
  of	
  Commerce.	
  



	
   29	
  

	
  

References	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  
1	
  http://911digitalarchive.org/.	
  
2	
  http://hurricanearchive.org/.	
  
3	
  http://www.jdarchive.org/;	
  “Digital	
  Humanities	
  Efforts	
  Range	
  from	
  Database	
  Design	
  to	
  
New	
  Creations”.	
  Harvard	
  Magazine.	
  May-­‐Jun	
  2012.	
  Accessed	
  January	
  14,	
  2013.	
  	
  

4	
  Eileen	
  McSaveney.	
  'Historic	
  earthquakes	
  -­‐	
  The	
  2011	
  Christchurch	
  earthquake',	
  Te	
  Ara	
  -­‐	
  the	
  
Encyclopedia	
  of	
  New	
  Zealand,	
  updated	
  13-­‐Jul-­‐12.	
  	
  
<	
  http://www.TeAra.govt.nz/en/historic-­‐earthquakes/page-­‐13>.	
  
5	
  Charlie	
  Gates.	
  “Our	
  Disappearing	
  City	
  Centre.”	
  Stuff.co.nz,	
  September	
  29,	
  2012,	
  sec.	
  
Christchurch	
  Earthquake	
  2011.	
  http://www.stuff.co.nz/the-­‐press/news/christchurch-­‐
earthquake-­‐2011/7744317/Our-­‐disappearing-­‐city-­‐centre.	
  

6	
  Weng	
  Y.	
  Kam,	
  Stefano	
  Pampanin,	
  and	
  Ken	
  Elwood.	
  “Seismic	
  Performance	
  of	
  Reinforced	
  
Concrete	
  Buildings	
  in	
  the	
  22	
  February	
  Christchurch	
  (Lyttelton)	
  Earthquake.”	
  Bulletin	
  of	
  the	
  
New	
  Zealand	
  Society	
  for	
  Earthquake	
  Engineering	
  44,	
  no.	
  2	
  (December	
  2011):	
  243.	
  

7	
  NZ	
  Government.	
  “Minute	
  of	
  Decision:	
  Christchurch	
  CBD	
  Recovery.”	
  NZ	
  Government,	
  April	
  
12,	
  2012:	
  4.	
  

8	
  Sonia	
  Giovinazzi	
  and	
  Thomas	
  Wilson.	
  “‘Recovery	
  of	
  Lifelines’	
  Following	
  the	
  22nd	
  February	
  
2011	
  Christchurch	
  Earthquake:	
  Successes	
  and	
  Issues.”	
  2012	
  NZSEE	
  Conference	
  Proceedings	
  
(2012).	
  

9	
  Building	
  Research	
  Advisory	
  New	
  Zealand.	
  “BRANZ	
  Bulletin	
  551.”	
  BRANZ,	
  August	
  2012.	
  
10	
  Jarrod	
  Greig.	
  “Christchurch	
  Quake	
  Third	
  Most	
  Expensive	
  Disaster	
  Ever	
  -­‐	
  Insurer.”	
  New	
  
Zealand	
  Herald,	
  March	
  29,	
  2012,	
  sec.	
  National.	
  
http://www.nzherald.co.nz/nz/news/article.cfm?c_id=1&objectid=10795342.	
  

11	
  Anna	
  Turner.	
  “11,000	
  quakes	
  since	
  Sept	
  2010.”	
  Stuff.co.nz,	
  January	
  17,	
  2013.	
  
http://www.stuff.co.nz/the-­‐press/news/christchurch-­‐earthquake-­‐2011/8190585/11-­‐000-­‐
quakes-­‐since-­‐Sept-­‐2010.	
  

12	
  Jack	
  Pinkowski.	
  Disaster	
  Management	
  Handbook.	
  Hoboken:	
  CRC,	
  2008.	
  
13	
  Coppola,	
  Damon	
  P.	
  Introduction	
  to	
  International	
  Disaster	
  Management.	
  2nd	
  ed.	
  
Burlington:	
  Butterworth-­‐Heinemann,	
  2010:	
  9.	
  

14	
  Georgina	
  H.	
  Endfield,	
  Sarah	
  J.	
  Davies,	
  Isabel	
  Fernández	
  Tejedo,	
  Sarah	
  E.	
  Metcalfe,	
  and	
  
Sarah	
  L.	
  O’Hara.	
  “Documenting	
  Disaster:	
  Archival	
  Investigations	
  of	
  Climate,	
  Crisis,	
  and	
  
Catastrophe	
  in	
  Colonial	
  Mexico.”	
  In	
  Natural	
  Disasters,	
  Cultural	
  Responses :	
  Case	
  Studies	
  
Toward	
  a	
  Global	
  Environmental	
  History,	
  305	
  –	
  325.	
  Lanham:	
  Lexington	
  Books,	
  2009:	
  305.	
  

15	
  UNESCO.	
  Text	
  of	
  the	
  Convention	
  for	
  the	
  Safeguarding	
  of	
  Intangible	
  Cultural	
  Heritage.	
  
Paris:	
  UNESCO,	
  October	
  17,	
  2003.	
  

16	
  United	
  Nations	
  International	
  Strategy	
  for	
  Disaster	
  Reduction.	
  “Hyogo	
  Framework	
  for	
  
Action	
  (HFA):	
  Building	
  the	
  Resilience	
  of	
  Nations	
  and	
  Communities	
  to	
  Disasters.”	
  United	
  
Nations,	
  2005.	
  

17	
  UNESCO/UBC.	
  Vancouver	
  Declaration.	
  The	
  Memory	
  of	
  the	
  World	
  in	
  the	
  Digital	
  Age:	
  
Digitization	
  and	
  Preservation.	
  Vancouver:	
  UNESCO/UBC,	
  2013.	
  

18	
  The	
  University	
  of	
  Canterbury	
  is	
  the	
  main	
  university	
  serving	
  the	
  broader	
  Christchurch	
  
region.	
  Other	
  significant	
  tertiary	
  institutions	
  in	
  the	
  area	
  include	
  Lincoln	
  University	
  and	
  the	
  
Christchurch	
  Polytechnic	
  Institute	
  of	
  Technology	
  (CPIT).	
  

19	
  State	
  of	
  emergency	
  lifted	
  in	
  Christchurch".	
  3	
  News.	
  1	
  May	
  2011.	
  Retrieved	
  6	
  May	
  2011.	
  
http://www.3news.co.nz/State-­‐of-­‐emergency-­‐lifted-­‐in-­‐
Christchurch/tabid/423/articleID/209247/Default.aspx.	
  



	
   30	
  

	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  
20	
  Rolland,	
  Erik,	
  Raymond	
  A.	
  Patterson,	
  Keith	
  Ward,	
  and	
  Bajis	
  Dodin.	
  “Decision	
  support	
  for	
  
disaster	
  management.”	
  Operations	
  Management	
  Research	
  3,	
  no.	
  1–2	
  (March	
  2010):	
  68–79.	
  	
  

21	
  Stacey	
  Pitsillides,	
  Janis	
  Jeffries	
  and	
  Martin	
  Conreen,	
  'Museum	
  of	
  the	
  self	
  and	
  digital	
  death:	
  
an	
  emerging	
  curatorial	
  dilemma	
  for	
  digital	
  heritage'.	
  Heritage	
  and	
  Social	
  Media:	
  
Understanding	
  Heritage	
  in	
  a	
  Participatory	
  Culture.	
  1st	
  ed.	
  New	
  York,	
  NY:	
  Routledge,	
  2012:	
  
58.	
  

22	
  Brown,	
  Norman	
  R.,	
  Peter	
  J.	
  Lee,	
  Mirna	
  Krslak,	
  Frederick	
  G.	
  Conrad,	
  Tia	
  G.	
  B.	
  Hansen,	
  Jelena	
  
Havelka,	
  and	
  John	
  R.	
  Reddon.	
  “Living	
  in	
  History:	
  How	
  War,	
  Terrorism,	
  and	
  Natural	
  Disaster	
  
Affect	
  the	
  Organization	
  of	
  Autobiographical	
  Memory.”	
  Psychological	
  Science	
  (Wiley-­‐
Blackwell)	
  20,	
  no.	
  4	
  (April	
  2009):	
  399–405.	
  

23	
  Eviatar	
  Zerubavel.	
  “Social	
  Memories:	
  Steps	
  to	
  a	
  Sociology	
  of	
  the	
  Past.”	
  Qualitative	
  
Sociology	
  19,	
  no.	
  3	
  (September	
  1,	
  1996):	
  283–299.	
  	
  

24	
  Dirk	
  H.R.	
  Spennemann,	
  “Cultural	
  Heritage	
  Conservation	
  During	
  Emergency	
  Management:	
  
Luxury	
  or	
  Necessity?”	
  International	
  Journal	
  of	
  Public	
  Administration	
  22,	
  no.	
  5	
  (January	
  1,	
  
1999):	
  746.	
  

25	
  Daniel	
  P.	
  Aldrich.	
  Building	
  Resilience :	
  Social	
  Capital	
  in	
  Post-­‐disaster	
  Recovery.	
  Chicago,	
  IL:	
  
University	
  of	
  Chicago	
  Press,	
  2012.	
  Aldrich	
  defines	
  resilience	
  as	
  (7)	
  “a	
  neighborhood’s	
  
capacity	
  to	
  weather	
  crises	
  such	
  as	
  disasters	
  and	
  engage	
  in	
  effective	
  and	
  efficient	
  recovery	
  
through	
  coordinated	
  efforts	
  and	
  cooperative	
  activities”.	
  

26	
  Patrick	
  Meier.	
  “How	
  to	
  Create	
  Resilience	
  Through	
  Big	
  Data.”	
  iRevolution,	
  January	
  11,	
  2013.	
  	
  
27	
  Anthony	
  Oliver-­‐Smith,	
  'The	
  Centrality	
  of	
  Culture	
  in	
  Post-­‐Disaster	
  Reconstruction'.	
  
Goldewijk,	
  Berma	
  Klein,	
  Georg	
  Frerks,	
  and	
  Els	
  van	
  der	
  Plas,	
  eds.	
  Cultural	
  Emergency	
  in	
  
Conflict	
  and	
  Disaster.	
  Rotterdam:	
  NAi	
  Publishers,	
  2011:	
  224.	
  

28	
  Iwana	
  Chronis,	
  Louk	
  de	
  la	
  Rive	
  Box,	
  Eleonore	
  de	
  Merode,	
  'Cultural	
  Emergency	
  Response:	
  
At	
  the	
  Crossroads	
  of	
  Heritage	
  and	
  Humanitarianism'.	
  Goldewijk,	
  Berma	
  Klein,	
  Georg	
  Frerks,	
  
and	
  Els	
  van	
  der	
  Plas,	
  eds.	
  Cultural	
  Emergency	
  in	
  Conflict	
  and	
  Disaster.	
  Rotterdam:	
  NAi	
  
Publishers,	
  2011:	
  348.	
  See	
  also	
  Coppola:	
  407	
  –	
  408.	
  

29	
  Sheila	
  A.	
  Brennan,	
  and	
  T.	
  Mills	
  Kelly.	
  “Why	
  Collecting	
  History	
  Online	
  Is	
  Web	
  1.5.”	
  Roy	
  
Rosenzweig	
  Center	
  for	
  History	
  and	
  New	
  Media.	
  ND.	
  Accessed	
  January	
  14,	
  2013.	
  	
  

30	
  ‘CEISMIC’	
  initially	
  stood	
  for	
  ‘Canterbury	
  Earthquakes	
  Images,	
  Stories	
  and	
  Media	
  Integrated	
  
Collection’,	
  but	
  the	
  full	
  title	
  was	
  soon	
  dropped	
  in	
  favour	
  of	
  the	
  more	
  easily	
  digestible	
  ‘UC	
  
CEISMIC	
  Canterbury	
  Earthquakes	
  Digital	
  Archive’.	
  

31	
  Michael	
  Forstrom,	
  Nancy	
  Kuhl,	
  Susan	
  Thomas,	
  Jeremy	
  Leighton	
  John,	
  Megan	
  Barnard,	
  
Gabriela	
  Redwine,	
  Kate	
  Donovan,	
  Erika	
  Farr,	
  Will	
  Hansen,	
  and	
  Seth	
  Shaw.	
  Born	
  Digital:	
  
Guidance	
  for	
  Donors,	
  Dealers,	
  and	
  Archival	
  Repositories.	
  Media	
  Commons	
  Press,	
  2013.	
  	
  

32	
  Ingrid	
  Mason,	
  'Cultural	
  Information	
  Standards	
  -­‐	
  Political	
  Territory	
  and	
  Rich	
  Rewards',	
  Ed.	
  
Fiona	
  Cameron,	
  Sarah	
  Kenderdine.	
  Theorizing	
  Digital	
  Cultural	
  Heritage:	
  a	
  Critical	
  Discourse.	
  
Media	
  in	
  Transition.	
  Cambridge,	
  Mass:	
  MIT	
  Press,	
  2007:	
  231.	
  

33	
  http://quakestories.govt.nz.	
  QuakeStories,	
  along	
  with	
  the	
  Christchurch	
  City	
  Libraries	
  Kete	
  
archive	
  
(http://ketechristchurch.peoplesnetworknz.info/canterbury_earthquakes_2010_2011),	
  
was	
  one	
  of	
  the	
  first	
  post-­‐quake	
  archiving	
  initiatives.	
  

34	
  Gillian	
  Oliver,	
  Brenda	
  Chawner,	
  and	
  Hai	
  Ping	
  Liu.	
  “Implementing	
  Digital	
  Archives:	
  Issues	
  of	
  
Trust.”	
  Archival	
  Science	
  11,	
  no.	
  3–4	
  (November	
  1,	
  2011):	
  313.	
  

35	
  Barwick,	
  Linda.	
  “Turning	
  It	
  All	
  Upside	
  Down	
  .	
  .	
  .	
  Imagining	
  a	
  Distributed	
  Digital	
  Audiovisual	
  
Archive.”	
  Literary	
  and	
  Linguistic	
  Computing	
  19,	
  no.	
  3	
  (September	
  1,	
  2004):	
  254.	
  

36	
  UC	
  CEISMIC	
  Programme	
  Board.	
  Memorandum	
  of	
  Understanding.	
  Christchurch,	
  09	
  
September	
  2011:	
  1.	
  
37	
  At	
  the	
  time	
  of	
  writing	
  it	
  seems	
  likely	
  that	
  DigitalNZ	
  will	
  open	
  source	
  their	
  metadata	
  
aggregation	
  toolset,	
  making	
  it	
  possible	
  for	
  the	
  entire	
  UC	
  CEISMIC	
  system	
  to	
  be	
  replicated.	
  



	
   31	
  

	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  
38	
  http://www.digitalnz.org/about.	
  Accessed	
  14	
  February	
  2013.	
  
39	
  http://natlib.govt.nz/.	
  
40	
  http://www.digitalnz.org/records?text=%22CEISMIC%22.	
  
41	
  http://www.nvinteractive.co.nz/news/nv-­‐announced-­‐as-­‐winner-­‐in-­‐microsofts-­‐
humanising-­‐data-­‐competition.	
  Accessed	
  14	
  February	
  2013.	
  

42	
  Susanna	
  M.	
  Hoffman.	
  'The	
  Worst	
  of	
  Times,	
  the	
  Best	
  of	
  Times:	
  Toward	
  a	
  Model	
  of	
  Cultural	
  
Response	
  to	
  Disaster'.	
  The	
  Angry	
  Earth:	
  Disaster	
  in	
  Anthropological	
  Perspective.	
  New	
  York:	
  
Routledge,	
  1999:	
  143.	
  

43	
  fedora_rest	
  by	
  Don	
  Gourley	
  (https://github.com/dongourley/fedora_rest).	
  The	
  intention	
  is	
  
to	
  open	
  source	
  the	
  modifications	
  when	
  time	
  and	
  resource	
  allows.	
  

44	
  Consideration	
  was	
  given	
  to	
  using	
  open	
  source	
  CENTOS	
  servers,	
  but	
  as	
  the	
  University	
  uses	
  
RHEL	
  it	
  was	
  felt	
  sensible	
  to	
  stay	
  with	
  the	
  supported	
  product.	
  Because	
  of	
  the	
  close	
  
relationship	
  between	
  RHEL	
  and	
  CENTOS	
  the	
  feeling	
  is	
  that	
  the	
  current	
  application	
  stack	
  
could	
  be	
  open	
  sourced	
  using	
  CENTOS	
  relatively	
  easily.	
  
45	
  http://www.bluefern.canterbury.ac.nz/.	
  At	
  the	
  time	
  of	
  writing	
  Blue	
  Fern®	
  features	
  IBM	
  
Blue	
  Gene/L,	
  and	
  IBM	
  Blue	
  Gene/P	
  and	
  an	
  IBM	
  POWER7	
  cluster.	
  

46	
  http://www.nesi.org.nz/.	
  
47	
  http://www.karen.net.nz/.	
  
48	
  At	
  the	
  time	
  of	
  writing	
  the	
  UC	
  CEISMIC	
  Research	
  Committee	
  was	
  chaired	
  by	
  Prof.	
  and	
  Dean	
  
of	
  Postgraduate	
  Studies	
  Lucy	
  Johnston,	
  and	
  included	
  representatives	
  from	
  the	
  University	
  of	
  
Canterbury	
  (including	
  the	
  Human	
  Ethics	
  Committee),	
  Massey	
  University,	
  and	
  Otago	
  
University.	
  

49	
  Mason,	
  'Cultural	
  Information	
  Standards’:	
  226.	
  
50	
  Alonzo	
  C.	
  Addison,	
  'The	
  vanishing	
  virtual:	
  Safeguarding	
  heritage’s	
  endangered	
  
digital	
  record".	
  Kalay,	
  Yehuda.	
  New	
  Heritage	
  New	
  Media	
  and	
  Cultural	
  Heritage.	
  Hoboken:	
  
Taylor	
  &	
  Francis,	
  2007:	
  35.	
  

51	
  http://www.foaf-­‐project.org/.	
  
52	
  Fedora	
  Commons.	
  The	
  Fedora	
  Digital	
  Object	
  Model.	
  http://fedora-­‐
commons.org/documentation/3.0b1/userdocs/digitalobjects/objectModel.html.	
  	
  

53	
  Jason	
  Darwin,	
  ‘QuakeStudies	
  Repository:	
  API	
  Documentation’.	
  
54	
  This	
  assumption	
  is	
  backed	
  up	
  by	
  research.	
  See	
  Hoffman,	
  'The	
  Worst	
  of	
  Times,	
  the	
  Best	
  of	
  
Times’:	
  149.	
  

55	
  Barbara	
  Kirshenblatt-­‐Gimblett.	
  “Intangible	
  Heritage	
  as	
  Metacultural	
  Production.”	
  Museum	
  
International	
  56,	
  no.	
  1/2	
  (May	
  2004):	
  54.	
  
56	
  Neil	
  Silberman	
  and	
  Margaret	
  Purser,	
  'Collective	
  Memory	
  as	
  Affirmation:	
  People-­‐centered	
  

cultural	
  Heritage	
  in	
  a	
  digital	
  age'.	
  Heritage	
  and	
  Social	
  Media:	
  Understanding	
  Heritage	
  in	
  a	
  
Participatory	
  Culture.	
  1st	
  ed.	
  New	
  York,	
  NY:	
  Routledge,	
  2012:	
  14.	
  
57	
  Ibid.	
  
58	
  https://www.facebook.com/StudentVolunteerArmy.	
  
59	
  Patrick	
  Meier.	
  “Social	
  Media:	
  Pulse	
  of	
  the	
  Planet?”	
  Blog.	
  iRevolution,	
  February	
  2,	
  2013.	
  	
  
60	
  Cornelius	
  Puschmann	
  and	
  Jean	
  Burgess.	
  “The	
  Politics	
  of	
  Twitter	
  Data	
  (January	
  23,	
  2013).	
  
HIIG	
  Discussion	
  Paper	
  Series	
  No.	
  2013-­‐01.	
  	
  

61	
  Beer,	
  David.	
  “Power	
  Through	
  the	
  Algorithm?	
  Participatory	
  Web	
  Cultures	
  and	
  the	
  
Technological	
  Unconscious.”	
  New	
  Media	
  &	
  Society	
  11,	
  no.	
  6	
  (September	
  1,	
  2009):	
  986.	
  

62	
  Sheenagh	
  Pietrobruno.	
  “YouTube	
  and	
  the	
  Social	
  Archiving	
  of	
  Intangible	
  Heritage.”	
  New	
  
Media	
  &	
  Society	
  (January	
  13,	
  2013):	
  6.	
  

63	
  Cited	
  in	
  David	
  Beer.	
  “Power	
  Through	
  the	
  Algorithm?	
  Participatory	
  Web	
  Cultures	
  and	
  the	
  
Technological	
  Unconscious.”	
  New	
  Media	
  &	
  Society	
  11,	
  no.	
  6	
  (September	
  1,	
  2009):	
  987.	
  

64	
  Michael	
  K.	
  Bergman,	
  “White	
  Paper:	
  The	
  Deep	
  Web:	
  Surfacing	
  Hidden	
  Value.”	
  The	
  Journal	
  
of	
  Electronic	
  Publishing	
  7,	
  no.	
  1	
  (August	
  2001).	
  	
  



	
   32	
  

	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  
65	
  Alonzo	
  C.	
  Addison,	
  'The	
  vanishing	
  virtual:	
  Safeguarding	
  heritage’s	
  endangered	
  
digital	
  record".	
  Kalay,	
  Yehuda.	
  New	
  Heritage	
  New	
  Media	
  and	
  Cultural	
  Heritage.	
  Hoboken:	
  
Taylor	
  &	
  Francis,	
  2007:	
  37.