W3C home > Mailing lists > Public > public-vocabs@w3.org > March 2012

espn/google sports vocabs, first pass and questions

From: Paul Kelly <paul@xmlteam.com>
Date: Mon, 12 Mar 2012 23:04:10 -0400
Message-Id: <D0B164B7-CF03-48A4-97A0-8C98A9025B8D@xmlteam.com>
To: public-vocabs@w3.org
OK, first off thanks to the ESPN and Google crew for putting this out. This is an important contribution to something that I think a lot of people are interested in. Here are some questions and comments from a SportsML perspective. I realize we're talking about web pages here, not xml representations of sports events. So please forgive me if I'm misunderstanding the intent behind some of these properties.

I'll refer to pages numbers in the pdf here:


1) SportsAssociation (p.2)

A sub-class of Organization i.e Thing > Organization > SportsAssociation. So Major League Baseball (MLB) is SportsAssociation/Baseball. But so is a teams's division, according to the documentation -- "For example: NY Yankees belongs to the East Division (in the American League, MLB))." (p.7)

SportsML calls the latter an affiliation. It's not uncommon for a team to move between divisions or even conferences but still belong to the main organization, the league.

2) Sports extensions (p.2)

For hockey (SportsAssociation/Hockey) the assumption is ice-hockey. Field Hockey would be explicitly designated in future?

3) SportsTeam (p.7)

The extensions are all the usual team sports. But what about tennis doubles and golf match play? Also Davis and Ryder Cups. Should these be added for those circumstances?

4) SportsAthlete (p.9)

I'm interested in the hierarchy expressed in the example, for Husain Abdullah (p.10), who plays football for the Minnesota Vikings, and whose position is safety. The top level div is as follows:

<div itemscope="itemscope" itemtype="http://schema.org/SportsAthlete/Football/Safety">

Paul Wilton has asked about SportsAthlete as a sub-class of Person. I'll just add that people do switch between sports (eg. Michael Jordan) and position (too many to name). It's also true that many athletes play more than one position throughout a season or even an event. That's OK, I would only treat this is a temporal property. But I just wonder about the specificity of that formation in the sample. I notice it continues with the page items:

<span itemscope="itemscope" itemtype="http://schema.org/SportsAthlete/Football/Safety"><link itemprop="url" href="http://espn.go.com/nfl/player/_/id/11910/husain-abdullah" /></span> (p.13)

5) SportStat (p.24)

I saw "http://schema.org/SportStat/TotalTackles" in the Abdullah sample (p.11) and wondered why this wasn't qualified by sport. But I see later in the doc that baseball-specific stats are offered. Will tackles be football-specific or are they more general and equivalent to soccer tackles?

6) SportsMatch (p.14)

Here is the structure: Event > SportsEvent > SportsMatch

What is the distinction between these three? I understand there's a difference between the event is the event as a whole including ticket sales, ceremonies, peripheral events like parades, etc. Match is the actual contest?

Is broadcast info a property of event or match?

7) SportsSeries (p.24)

Series is the same as SportsML's "season-type". The vocab is pre-season, regular-season, post-season and tournament. A few questions here:

a) presumably the vocab can be extended to include all-star matches, exhibition (friendlies) and other possible types
b) how would one designate a best-of-seven playoff "series" in MLB, NBA or NHL?
c) for something like the US Open for golf or tennis, is that the tournament? Is there a way to specify the entire season?
d) what would a "single-table" series like EPL (and other domestic European leagues for many sports) be called, tournament or regular-season?
e) is there a need to specify the stages of a tournament like FIFA World Cup? eg. qualifier, groups, elimination.

8) Status vocab (p.14)

This vocab seems fairly generic and applicable beyond just sport. Could it be generalized? FWIW, here is the full SportsML enumeration:

			<enumeration value="pre-event">
			<enumeration value="mid-event">
			<enumeration value="post-event">
			<enumeration value="postponed">
			<enumeration value="suspended">
			<enumeration value="halted">
			<enumeration value="forfeited">
			<enumeration value="rescheduled">
			<enumeration value="delayed">
			<enumeration value="canceled">
			<enumeration value="intermission">
			<enumeration value="if-necessary">
			<enumeration value="discarded">

The IPTC also has EventsML which l imagine has something similar.

9) Period (p.20)

SportsML uses sub-score. Period seems to me to shift toward presentation than to logic or semantics. Was this discussed?

10) Rank (p.27) and SportsStandings (p.26)

Does Rank denote seeding in standings as well as win/loss?  If for standings "All the teams ordered according to their ranking", how are the rank values marked up?

Paul Kelly
Director, Software Development, XML Team Solutions Corp.
Chair, Sports Content Working Party, IPTC
Received on Tuesday, 13 March 2012 03:04:34 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:22 UTC