Sportwire II: Requirements

Capabilities and Deliverables

Gary Lawrence Murphy

Project Architect
Teledynamics Communications Inc

7 Forest Place
          Sauble Beach 
          Ontario 
          Canada
        

$Revision: 1.3 $

Copyright

Sportwire is copyright by the Canadian Broadcasting Corporation. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.1 or any later version published by the Free Software Foundation. This software is provided without warrantee and no promises are implied; use at your own risk. The Sportwire project is hosted at Sourceforge.net and community participation is welcome and encouraged.

$Date: 2001/11/18 16:17:02 $

Abstract

Sportwire is an archiving and query system for XML documents distributed by sports newswire services. Sportwire is specifically designed for the ESPN Sportsticker, but should be adaptable to any XML-based news feed which provides an event-centric description of sports news. The publishing system provides the tools to parse and store multiple DTD-described feeds and web-side tools for presenting extracted reports.

For more information on the ESPN Sportsticker, please refer to http://sportsticker.com.


Table of Contents

prelude
Reporting Requirements
Phase I Sportsticker Reports
Phase II AFP Olympics Reports
Phase III Sportsticker XML Reports
Administration Reports
Constraints
Dataserver Requirements
Constraints
Webserver Requirements
Constraints
Data Acquisition Requirements
Constraints

prelude

When Sportwire I was released in 1997, there was no XML standard for news, and precious few tools for object relational mapping. That system imposed an ad-hoc XML on the binary Sportsticker feed, with a hierarchy of Java objects to parse and store incoming documents based on the hierarchy of bc-* Sportsticker slug names.

The new system seeks to fix several difficulties with the original program. Our first priority is in improving reliability, overcoming game-number and schedule errors in the Sportsticker feed. Our second priority is to prepare for the 2002 Sportsticker XML feed. A close third priority is to free Sportwire from the Sportsticker formats, and to base feed parsing on the XML DocType regardless of ESPN-specific slug names.

Glossary

Sports Feed Terminology

ESPN

The Sports Network provides its Sportsticker service in a custome XML schema, delivered via FTP or through a TCP socket connection. The XML uses a unique set of DTDs for each sport, where each message type has a unique DTD (even if it is virtually identical to the same class of messages in another sport). ESPN is part of MSN.

AFP

Agence France Presse has been providing detailed feeds for the Winter Olympics since 1948; the feed is available in French and English. Since NewsML does not (yet) define a sports schema, AFP Olympics coverage is delivered as a namespace within a NewsML envelope.