System Overview

\n

WAIS was developed by Brewster Kahle and Roy Fielding in 1991 as a client/server information retrieval system designed to search large text databases across networks. Before HTML became dominant, WAIS relied on the ANSI Z39.50 standard to enable uniform searching across heterogeneous databases.

Unlike modern web search, WAIS returned scored results based on exact term frequency, inverse document frequency, and boolean operators. It supported multiple query types, including title search, keyword search, date range, and proximity searches.

waiss://merlin.lib.uci.edu/ query -t TI -k "network protocols" -n 10
Searching catalog: US-NCST
[204] Score: 87.3 TCP/IP Architecture (1992)
[203] Score: 74.1 X.25 vs Frame Relay (1991)
[202] Score: 61.8 Internetworking Guidelines (1990)
... 7 more results below score 50.0
waiss://merlin.lib.uci.edu/ _

Query Syntax & Operators

WAIS queries used structured prefixes to target specific document fields. The protocol supported boolean logic, proximity matching, and exact phrase searches.

PrefixTypeDescription
TITitle SearchMatches document titles (highest weight)
KWKeyword SearchFull-text search across body content
DADate SearchFilters by publication or modification year
EXExact PhraseRequires contiguous term matching
OPBoolean/LogicSU (substring), NE (near), AND/OR/NOT

Network Architecture

WAIS operated on a strict client/server model using TCP port 210. The client formatted requests according to Z39.50, while the server maintained an inverted index database. Responses included document scores, metadata, and binary payloads (often ASCII text or early PostScript).

Protocol
ANSI Z39.50
Default Port
210 / TCP
Index Type
Inverted Term Frequency
Encoding
ASCII / US-NCST Binary
Auth Method
Host-based / RWHOIS
Peak Usage
1993–1994