Storage of digital objects

1
Storage of digital objects

• Adolf Knoll
• National Library of the Czech Republic
• adolf.knoll_at_nkp.cz

2
Basic principles

• store digital objects in such a way that you can
  always
• find them know where they are
• identify them know what they contain
• open and access them
• manipulate with them
• provide access to them to users

3
Search of realiability

• image and other data files
• metadata platform and definitions
• structural schemes
• storage media
• Everything is in motion we should evaluate its
  speed and prefer more durable solutions.

4
Image and data files

• they should have unique names
• they sholuld be attached to identification
  metadata
• they should be archived in highly used ISO
  formats (TIFF, JPEG)
• we should be aware of the fact that the source
  archival image can be reused
• e.g. Czech NL now optimizing access image sets

5
Metadata

• openly readable platforms (SGML Family)
• content markup and possibility to reuse them
• content description rules with enough granularity
• reliable storage of metadata
• awareness that one day they might be migrated
• e.g, Czech NL migrates now medata sets for ca.
  1.3 million digitized page (in continuing
  production)
• make them independent of access tools

6
Structuring and storing

• Structure metadata and link them to image
  entities
• Store the structured units together
• If creating database access, pay a lot of
  attention to preservation of the database
• have always a reference off-line archives for
  digitized documents

7
(No Transcript)
8
Media

• media are not critical for long-term storage
• the dependence of digital documents on software
  and hardware environments is critical
• media can be refreshed
• always have more independently stored copies

9
Stability of media vs. obsolescence

• It seems that the most reliable medium is CD
• obsolete, but well standardized
• it contains redundant data for corrections
• far from being dead it is well suitable also
  for video thanks to new compression standards
• cheap
• the CD archives should be measured and monitored
• do not OVERBURN!

10
Other media

• Most data is stored on magnetic tapes
• problems with writing software after years, no
  data redundancy
• fragility tapes should be rewound
• mass robotic storages facilities/libraries
• self-checking, rewriting, but
• Hard disks backups more access media
• DVD, flash cards,

11
An example compact disc
scratch from manipulation with reader
printing matrix made this scratch
metallic coating insuficciency
12
Origin of damages of CD

• manipulation scratches 43
• dust 28
• damages incurred in production 15
• Table of Contents errors 2
• Other 15
• Not all of them are critical for readability, but
  they warn us.

13
Quality vs. commerce

• Industrial production works for current users
  acceptable quality of media
• For long-term storage needed more reliable
  quality
• Only co-operation with producers is the solution
• Especially compact discs and magnetic tapes
  (UNESCO involvement)

14
Proprietary vs. Open source

• Open source we possess resources and can change
  the product usually required by academic sphere
• However Microsoft, Adobe, Ex-Libris no chance
• Society has a consumer character also in more
  important fields than libraries
• I.e., work with small and medium enterprises,
  local companies, try to develop yourselves in
  harmony and grow both