Symposium on Best Practice

1
Resource Conversion

• William Lewis
• CSU Fresno

2
Preliminaries

• Eventually any resource will become obsolete
• ? Resource conversion is inevitable
• One should plan from the start for eventual
  conversion
• Encode your resource such that it
• is Migratable
• is Reusable
• will Endure

3
Best Practice?

• Simons (this symposium) argues there are 3
  relevant formats for encoding data
• Working form
• Presentation form
• Archival form
• (1) is tied to particular software.
• (2) is generally generated from (1), but itself
  is often semantically sparse.

4
Best Practice!

• Encoding resource in archival form (3) insures
  that
• the resource is reusable, facilitating
  interoperability
• the resource can be migrated to other formats
  (including presentation formats)
• the resource endures

5
Data Survivability

• Converting to archival XML form provides for data
  reuse and insures survivability

Working Form
Archival Form
Conversion Process
HTML
PDF
Other XML form
6
Data Conversion (text)

• Highly dependent on flexibility of working form
  and related software
• Converting from proprietary, binary format most
  difficult to be avoided
• Converting from plain text output easiest might
  be avoided due to potential data loss
• Converting from enriched text form (Unicode
  compliant) or XML coded data is best, but may not
  always be possible

7
Data Conversion
Working Form
Archival Form
Conversion Process
8
Data Conversion
Working Form
Inter-mediary Form (Enriched text)
Archival Form
CP
9
Intermediary Conversion

• Use
• Print Function or Save as
• Print Function or other file convert
• Data Query (direct to XML?)
• As is

• Resources in
• Word Processor
• Spreadsheet
• Proprietary Flat File DB
• Relational DB
• XML or enriched text (inc. Shoebox)

10
Intermediary Conversion

• Important
• Insure that conversion to Intermediary Form
  suffers no data loss, or that the data loss
  suffered is minimal
• Danger in Save As (and Print to file), in that
  data loss is possible

11
Final Conversion

• Intermediary to Archival Form (Best Practice
  XML)
• Font/Character transforms
• Macros or methods for enriching and aligning data
  elements
• Tables or glossaries defining how content and
  form should be interpreted

12
Data Conversion Case Study

• Converting Hopi Dictionary (Hill et al 1998) from
  working form (legacy format)
• Purpose
• Build software to extract relevant data from
  working form
• Generate reusable archival format
• For dissemination on the Web
• For use by others
• To preserve data should DB software become
  unusable

13
Hopi Dictionary

• Example entry from Hopi Dictionary

14
Hopi Dictionary Conversion

• Until now
• Generated text file from DB
• Manually converted IPA fonts in MSWord
• Generated PDFs for dissemination

15
Hopi Dictionary Conversion

• New Process
• Convert DB format to enriched text
• Software transforms for fonts from text format
  (Unicode compliant IPA)
• Identify the grammatical concepts used in
  entries, linked to GOLD (Farrar Langendoen,
  this symposium)
• Generate XML structured using modified EMELD
  IGT format (Bow, Hughes Bird 2003)

16
Archival Hopi Dictionary Record
17
Recipe for Resource Conversion

• Choose data format that is easily archived
• Where the software provides for data migration,
  or,
• The data format itself is easily converted
• Use existing software to bring you as close to
  Archival Form as possible (Intermediary Form)
• Clearly identify
• Content and structural semantics (terms)
• Fonts used (and transforms)
• Data alignment
• Construct transforms/macros/software to convert
  to Archival form