The eXtensible Past

1
The eXtensible Past

• XML As a Means for Easy Access to Historical
  Research Data and a Strategy for Digital
  Preservation

2
NHDA

• Registered research
• Deposited Datasets
• Archiving, Dissemination Publishing
• No restrictions
• Restricted access

3
Reasons for starting X-past

• Problems with ASCII preservation format
• HTML-datasets
• Need for decentralized repositories
• Making data searchable on the Web
• Data-publishing increases demand for data.
• Improve reusability of datasets

4
X-past Project principles

• To investigate XML as a possible new strategy for
  the long-term preservation of research data
• Exploring the possibilities of XML and OAI (Open
  Archives Initiative) for providing better access
  to and sharing of digital data collections by
  researchers

5
XML for long-term preservation

• XML is human readable
• XML is platform independent
• XML integrates Data, Metadata and Structure
• XML a hype?

6
ASCII fixed format representation and codebook.

• DATA
• 26502 Martensz Matheeus Kruidenier Antwerpen
  19-05-1586 B 35
• CODEBOOK
• Name Explanation Sort position
• PersID personal identification
  number integer 1-5
• Family name Surname of person text 6-25

7
Record in XML-representation

• ltrecordgt
• ltrowgt
• ltpersIDgt26502lt/persIDgt  
  ltfamily_namegtMartenszlt/family_namegt
  ltfirst_namegtMatheeuslt/first_namegt
  ltprofessiongtKruidenierlt/professiongt
• ltorigingtAntwerpenlt/origingt
• ltdate_of_entrygt19-05-1586lt/date_of_entrygt
• ltentry_numbergtBlt/entry_numbergt
• ltentry_pagegt35lt/entry_pagegt
• lt/rowgt
• lt/recordgt

8
Providing better access

• Data searchable on the web
• Customizable datapresentation using XSLT
• Downloads available in xml original format
• Integration with other Metadata repositories
  possible OAI

9
OAI Overview
10
What have we built?

• Repository Server
• Management Application
• Access Portal

CMS
Repository Server
Manager
Default Portal
Other Application
11
System Architecture

• Modular
• Extensible

12
Automatic Conversion

• Archivist uploads dataset
• Intermediate format (CSV)
• Original Format (DBASE, ASCII)
• More formats to come
• System automatically converts format to XML
• Creates searchable index on material

13
Current Issues

• Carried out prototype validation with different
  user groups
• Difference between metadata search and dataset
  search is not clear
• Downloading XML files can take a while
• Navigation of search results must be improved

14
XML-issues

• Only applicable for text-oriented datasets
• Limited human-readability for binary content
  (images)
• Problems arise with more complex datasets
• Loss of functionality
• Loss of context

15
Loss of context

• Data is converted from original format to XML
• Information can be available in other parts of
  the dataset than just the raw data
• Field names / labels
• Queries generating reports

16
Conclusions

• XML is human readable (text, databases
  spreadsheets)
• XML-schema like DDI makes integration of Data,
  Metadata Codebooks possible
• XSLT makes rapid Customized Data-publishing
  possible.
• Reusability improves with XML.

17
Conclusions II

• OAI makes distributed Repository Access possible
• X-past has a flexible modular Architecture

18
Follow-up Project XARA

• Should adress following issues
• Loss of functionality
• related tables xlink?
• Stored procedures
• Loss of context
• Developing customized websites of datasets.
• Using DDI-schema for integrated metadata-data
  presentation.