Future-Proofing Z39.50 6 December 2000

1
Future-Proofing Z39.506 December 2000

• John A. Kunze
• UC San Francisco and
• National Library of Medicine

2
Z39.50 History a Personal Look

• Inception and OSI postpartum
• Need for non-bibliographic IR
• Emergence of lean protocols and bulky markup

3
Z39.50 Group Dynamics

• Collective lt sum of parts?
• The bike shed discussion
• The restaurant debate
• Consensus-based planning woes elsewhere

4
The Rise of Worse-is-Better

• Right way the MIT school (LISP)
• Worse-is-Better way the New Jersey school
  (UNIX/C)
• Worse gives ultimate computer virus (UNIX/C)
• Right gives endless design process and difficult
  implementation

5
Back to Z39.50 Design

• OSI terminology still haunts us
• ASN.1 is for Albatross, BER is for Bummer
• Dealing with library-centric reputation
• Implementation burden of MARC

6
Comparison with Search Engines

• Result set management not sacred
• Statefulness not sacred
• Piggybacked search
• Scan, Sort, Explain
• Init, Close, AR Control

7
Z39.50 at a Crossroads

• IR over HTTP
• www.google.com/search?qpubgrub
• ?qpubgrubhlensafeoffstart10
• NLM Entrez interface
• www.ncbi.nlm.nih.gov/entrez
• The XML factor