Thomas G' Habingthabinguiuc'edu

1
Simple Provider Tools for MS Windows
Platforms19 September 2001 -- OAI Provider
Workshop,
University of Illinois at Urbana-Champaign

• Thomas G. Habing(thabing_at_uiuc.edu)
• University of Illinois at Urbana-Champaign

2
WHAT

• Two Simple OAI Provider Implementations
• A - Metadata and protocol data stored in RDBMS
• B - Metadata stored as XML files on filesystem
  and protocol data stored in RDBMS
• MS Windows Platforms (NT, 2000, 9x?)
• IIS, ASP, VBScript or JScript
• ODBC-compliant data sources with stored
  procedures (Access, SQL Server, Oracle, etc.)
• MSXML Parser (DOM, XSLT, XML Schema)

3
Architecture

• Common to Implementations A and B
• Every record is in one set
• Sets are non-overlapping and hierarchical
• Protocol-specific administrative data, such as
  OAI identifiers, datestamps, sets, metadata
  formats, and repository identity, are stored in a
  database, and accessed via stored procedures.
• The minimal administrative data required are
• OAI Persistent Identifiers
• Metadata Datestamps

4
Architecture

• For Implementation A
• Metadata records are stored in the database
• Stored procedures are used to access the metadata
• This makes the database schema somewhat
  independent of the OAI metadata formats
• The stored procedure can do the mapping to DC or
  qualified DC
• The web server scripts transform the stored
  procedure results into the appropriate XML

5
Architecture

• For Implementation B
• Metadata records are stored as XML files on the
  filesystem (or the Internet)
• The database stores the OAI administrative data
  plus pointers to the XML files
• The database also has pointers to XSLT files
  which are used to transform the source XML
  metadata files to DC, qualified DC, or other
  formats
• This allows the source XML metadata to be of
  almost any format, such as EAD, so long as there
  is an XSLT mapping to DC available.

6
Architecture

• For Implementation B (cont.)
• All XML metadata files belonging to a given set
  share the same XML Schema and XSLT stylesheets
• A crawler is used to walk through all the XML
  metadata files in the filesystem and update the
  administrative data in the database.

7
Database for Implementation A
MetaFormat
Objects
Sets
URL to Digital Object
URL to XML Schema
RepositoryIdentity
Internet
Metadata
8
Database for Implementation B
MetaFormat
Objects
Sets
URL to Digital Object
URL to XML Schema
URL to XSLT
URL to XMLMetadata File
InternetorFilesystem
RepositoryIdentity
9
Demonstrations

• Implementation A
• Implementation B
• Metadata Case Study

Downloads

• Implementation A
• Implementation B

10
Conclusion

• We hope we have illustrated some simple ways to
  make OAI work for your metadata
• Metadata, both source and delivered, can be in
  various formats
• Must maintain at least minimal administrative
  data
• Metadata Datestamps
• Persistent OAI Identifiers