<?xml version="1.0" encoding="utf-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: The Story of Meehan&#8217;s Tale-Spin</title>
	<atom:link href="http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/feed/" rel="self" type="application/rss+xml" />
	<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/</link>
	<description>A group blog about computer narrative, games, poetry, and art.</description>
	<lastBuildDate>Thu, 30 Sep 2010 22:43:57 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.1.2</generator>
	<item>
		<title>By: Grand Text Auto &#187; Façade, Petz, and The Expressivator</title>
		<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/comment-page-1/#comment-159122</link>
		<dc:creator>Grand Text Auto &#187; Façade, Petz, and The Expressivator</dc:creator>
		<pubDate>Mon, 17 Dec 2007 01:05:17 +0000</pubDate>
		<guid isPermaLink="false">http://grandtextauto.org/?p=1281#comment-159122</guid>
		<description>[...] about a number of influential digital fiction systems, including James Meehan&#8217;s Tale-Spin (1 2), Scott Turner&#8217;s Minstrel (1 2), and Michael Lebowitz&#8217;s Universe (1). Now I&#8217;m [...]</description>
		<content:encoded><![CDATA[<p>[...] about a number of influential digital fiction systems, including James Meehan&#8217;s Tale-Spin (1 2), Scott Turner&#8217;s Minstrel (1 2), and Michael Lebowitz&#8217;s Universe (1). Now I&#8217;m [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Grand Text Auto &#187; Michael Lebowitz on Universe</title>
		<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/comment-page-1/#comment-143608</link>
		<dc:creator>Grand Text Auto &#187; Michael Lebowitz on Universe</dc:creator>
		<pubDate>Mon, 19 Nov 2007 17:01:17 +0000</pubDate>
		<guid isPermaLink="false">http://grandtextauto.org/?p=1281#comment-143608</guid>
		<description>[...] systems. I&#8217;ve been posting what I can share publicly, leading to posts about Tale-Spin (1 2) and Minstrel (1 2). Now I&#8217;m pleased to add a post with more information about Michael [...]</description>
		<content:encoded><![CDATA[<p>[...] systems. I&#8217;ve been posting what I can share publicly, leading to posts about Tale-Spin (1 2) and Minstrel (1 2). Now I&#8217;m pleased to add a post with more information about Michael [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Grand Text Auto &#187; The Context of Minstrel&#8217;s Creation</title>
		<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/comment-page-1/#comment-135565</link>
		<dc:creator>Grand Text Auto &#187; The Context of Minstrel&#8217;s Creation</dc:creator>
		<pubDate>Wed, 07 Nov 2007 14:17:19 +0000</pubDate>
		<guid isPermaLink="false">http://grandtextauto.org/?p=1281#comment-135565</guid>
		<description>[...] oughts on Grand Text Auto &#8212; as James Meehan had last year let us publish some of his memories of the creation of Tale-Spin. This came to fruition yesterday, when  [...]</description>
		<content:encoded><![CDATA[<p>[...] oughts on Grand Text Auto &#8212; as James Meehan had last year let us publish some of his memories of the creation of Tale-Spin. This came to fruition yesterday, when  [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Grand Text Auto &#187; Tale Spin at &#8220;Smart Machines&#8221;</title>
		<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/comment-page-1/#comment-135051</link>
		<dc:creator>Grand Text Auto &#187; Tale Spin at &#8220;Smart Machines&#8221;</dc:creator>
		<pubDate>Mon, 29 Oct 2007 02:15:02 +0000</pubDate>
		<guid isPermaLink="false">http://grandtextauto.org/?p=1281#comment-135051</guid>
		<description>[...] tp://grandtextauto.org/archives/smartMachinesCover_full.jpg&quot;&gt; 	From my earlier post on James Meehan&#8217;s Tale-Spin (now with a new  [...]</description>
		<content:encoded><![CDATA[<p>[...] tp://grandtextauto.org/archives/smartMachinesCover_full.jpg&#8221;&gt; 	From my earlier post on James Meehan&#8217;s Tale-Spin (now with a new  [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Scott Turner</title>
		<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/comment-page-1/#comment-134930</link>
		<dc:creator>Scott Turner</dc:creator>
		<pubDate>Fri, 26 Oct 2007 12:00:02 +0000</pubDate>
		<guid isPermaLink="false">http://grandtextauto.org/?p=1281#comment-134930</guid>
		<description>Early in my graduate student days, when I&#039;d just begun to work on Minstrel, I ran into James Meehan at a conference somewhere.  IIRC correctly, James was at DEC (?) at the time, and it was not an AI conference, so I was somewhat surprised to meet him in that context.  I blurted out that I was working on storytelling, and had just finished reading his dissertation (on a third-hand Xerox).  This was almost ten years after his dissertation had been published (to universal disregard).

James looked at me gravely for a few seconds and said &quot;Either I was ten years ahead of my time or you&#039;re ten years behind.&quot;

It took me ten years to fully understand his sentiment :-).

(In all seriousness, I had a nice conversation with him about the limits of Talespin and what the important issues would be for the next generation of storytelling programs.)</description>
		<content:encoded><![CDATA[<p>Early in my graduate student days, when I&#8217;d just begun to work on Minstrel, I ran into James Meehan at a conference somewhere.  IIRC correctly, James was at DEC (?) at the time, and it was not an AI conference, so I was somewhat surprised to meet him in that context.  I blurted out that I was working on storytelling, and had just finished reading his dissertation (on a third-hand Xerox).  This was almost ten years after his dissertation had been published (to universal disregard).</p>
<p>James looked at me gravely for a few seconds and said &#8220;Either I was ten years ahead of my time or you&#8217;re ten years behind.&#8221;</p>
<p>It took me ten years to fully understand his sentiment :-).</p>
<p>(In all seriousness, I had a nice conversation with him about the limits of Talespin and what the important issues would be for the next generation of storytelling programs.)</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Grand Text Auto &#187; Ergodic Histories in the Cybertext Database</title>
		<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/comment-page-1/#comment-133822</link>
		<dc:creator>Grand Text Auto &#187; Ergodic Histories in the Cybertext Database</dc:creator>
		<pubDate>Tue, 09 Oct 2007 06:26:13 +0000</pubDate>
		<guid isPermaLink="false">http://grandtextauto.org/?p=1281#comment-133822</guid>
		<description>[...] 21; see our earlier discussion (ignoring, if you will, my ill-informed characterization of Tale-Spin). 	 	 	                         	  	         		 		 	    Leave  [...]</description>
		<content:encoded><![CDATA[<p>[...] 21; see our earlier discussion (ignoring, if you will, my ill-informed characterization of Tale-Spin). 	</p>
<p>         		 		 	    Leave  [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: michael</title>
		<link>http://grandtextauto.org/2006/09/13/the-story-of-meehans-tale-spin/comment-page-1/#comment-96987</link>
		<dc:creator>michael</dc:creator>
		<pubDate>Wed, 20 Sep 2006 22:20:59 +0000</pubDate>
		<guid isPermaLink="false">http://grandtextauto.org/?p=1281#comment-96987</guid>
		<description>Great post Noah! Loved reading the behind-the-scenes history of Tale-Spin. I didn&#039;t know Meehan is at Google - now that I&#039;m close I&#039;ll have to look him up. 

With my interactive narrative students, Tale-Spin is where I start when covering AI-based story generation. I&#039;ve even had students modify the micro version of Tale-Spin to tell stories in a different domain. 

In contemporary work on drama management, Tale-Spin teaches a valuable lesson on the limits of emergent storytelling. If characters independently pursue their private goals, what you end up with is rarely a story, but rather a problem-solving trace that lacks the global structure that would make the trace a narrative. All subsequent work on AI-based story generation included explicit modeling of author-level knowledge, in addition to character-level modeling. In interactive drama, this plays out as the distinction between the drama manager and autonomous characters. But, even though Tale-Spin taught us the limits of character-only &quot;emergent&quot; story generation, people keep on reviving the idea in the context of story-based games and interactive drama. 

A re-implemented version of Tale-Spin was used to generate stories that served as input to the &lt;a href=&quot;http://mcox.org/Meta-AQUA/&quot; rel=&quot;nofollow&quot;&gt;Meta-Aqua story understanding system&lt;/a&gt;. The re-implemented Tale-Spin spun stories about a dysfunctional suburban family whom Elvis had come to live with. A typical story appears below. By the way, the language generator spit out template text for every action and state explicitly represented during the story generation process. The story understander&#039;s job is take this overdetailed explication, pull out the high points (the significant thread) and infer connections between the significant events. 

&lt;blockquote&gt;
One day ...ELVIS was JONESING. Elvis pushed cupboard-door away from the cupboard1. The cupboard1 was open. He took the pipe2 from the cupboard1. He had the pipe2. The cupboard1 didn&#039;t have the pipe2. He pushed cupboard-door to the cupboard1. The cupboard1 wasn&#039;t open. He took the ganja1 from the rug1. He had the ganja1. The rug1 didn&#039;t have the ganja1. The phone1 was ringing. Dad picked up phone-receiver1. The phone1 wasn&#039;t ringing. He had phone-receiver1. He let go of phone-receiver1. He didn&#039;t have phone-receiver1. Elvis poured the ganja1 into the pipe2. The pipe2 was filled with the ganja1. He took the lighter1 from the table2. He had the lighter1. The table2 didn&#039;t have the lighter1. He pushed the lighter1. The lighter1 was on. Police-and-dogs arrived. Officer1 went to outside. The police-dog1 went to outside. He pushed door-bell-switch1. The door-bell1 was ringing. He didn&#039;t push door-bell-switch1. The door-bell1 wasn&#039;t ringing. He went to the kitchen. The police-dog1 went to the kitchen. The police-dog1 went to Elvis. The police-dog1 sniffed Elvis. The police-dog1 barked at Elvis. The police-dog1 was barking. He went to Elvis. He took the ganja1 from Elvis. He had the ganja1. Elvis didn&#039;t have the ganja1. Officer1 arrested Elvis. He controlled Elvis. He arrested Elvis because he wanted to control Elvis. He went to outside. Elvis went to outside. The police-dog1 went to outside. The police-dog1 barked at him because the police-dog1 detected the ganja1. He had the lighter1. The ganja1 couldn&#039;t be burning. Officer1 controlled Elvis. Elvis couldn&#039;t get near the pipe1. He couldn&#039;t get the pipe1. The ganja1 couldn&#039;t fill the pipe1. He was still jonesing.
&lt;/blockquote&gt;

Michael (Michael Cox, the author of Meta-Aqua) told me about an interesting mis-spun tale generated by his system. The police came into the home demanding to know where the ganja was hidden. One of the actions available to the police was to beat people up as a form of coercion in order to gain desired information. But the coercion plan had no preconditions regarding believing that someone actually possessed the desired information. In this particular run, nobody in the family knew where the ganja was hidden, resulting in the police systematically beating up every family member.</description>
		<content:encoded><![CDATA[<p>Great post Noah! Loved reading the behind-the-scenes history of Tale-Spin. I didn&#8217;t know Meehan is at Google &#8211; now that I&#8217;m close I&#8217;ll have to look him up. </p>
<p>With my interactive narrative students, Tale-Spin is where I start when covering AI-based story generation. I&#8217;ve even had students modify the micro version of Tale-Spin to tell stories in a different domain. </p>
<p>In contemporary work on drama management, Tale-Spin teaches a valuable lesson on the limits of emergent storytelling. If characters independently pursue their private goals, what you end up with is rarely a story, but rather a problem-solving trace that lacks the global structure that would make the trace a narrative. All subsequent work on AI-based story generation included explicit modeling of author-level knowledge, in addition to character-level modeling. In interactive drama, this plays out as the distinction between the drama manager and autonomous characters. But, even though Tale-Spin taught us the limits of character-only &#8220;emergent&#8221; story generation, people keep on reviving the idea in the context of story-based games and interactive drama. </p>
<p>A re-implemented version of Tale-Spin was used to generate stories that served as input to the <a href="http://mcox.org/Meta-AQUA/" rel="nofollow">Meta-Aqua story understanding system</a>. The re-implemented Tale-Spin spun stories about a dysfunctional suburban family whom Elvis had come to live with. A typical story appears below. By the way, the language generator spit out template text for every action and state explicitly represented during the story generation process. The story understander&#8217;s job is take this overdetailed explication, pull out the high points (the significant thread) and infer connections between the significant events. </p>
<blockquote><p>
One day &#8230;ELVIS was JONESING. Elvis pushed cupboard-door away from the cupboard1. The cupboard1 was open. He took the pipe2 from the cupboard1. He had the pipe2. The cupboard1 didn&#8217;t have the pipe2. He pushed cupboard-door to the cupboard1. The cupboard1 wasn&#8217;t open. He took the ganja1 from the rug1. He had the ganja1. The rug1 didn&#8217;t have the ganja1. The phone1 was ringing. Dad picked up phone-receiver1. The phone1 wasn&#8217;t ringing. He had phone-receiver1. He let go of phone-receiver1. He didn&#8217;t have phone-receiver1. Elvis poured the ganja1 into the pipe2. The pipe2 was filled with the ganja1. He took the lighter1 from the table2. He had the lighter1. The table2 didn&#8217;t have the lighter1. He pushed the lighter1. The lighter1 was on. Police-and-dogs arrived. Officer1 went to outside. The police-dog1 went to outside. He pushed door-bell-switch1. The door-bell1 was ringing. He didn&#8217;t push door-bell-switch1. The door-bell1 wasn&#8217;t ringing. He went to the kitchen. The police-dog1 went to the kitchen. The police-dog1 went to Elvis. The police-dog1 sniffed Elvis. The police-dog1 barked at Elvis. The police-dog1 was barking. He went to Elvis. He took the ganja1 from Elvis. He had the ganja1. Elvis didn&#8217;t have the ganja1. Officer1 arrested Elvis. He controlled Elvis. He arrested Elvis because he wanted to control Elvis. He went to outside. Elvis went to outside. The police-dog1 went to outside. The police-dog1 barked at him because the police-dog1 detected the ganja1. He had the lighter1. The ganja1 couldn&#8217;t be burning. Officer1 controlled Elvis. Elvis couldn&#8217;t get near the pipe1. He couldn&#8217;t get the pipe1. The ganja1 couldn&#8217;t fill the pipe1. He was still jonesing.
</p></blockquote>
<p>Michael (Michael Cox, the author of Meta-Aqua) told me about an interesting mis-spun tale generated by his system. The police came into the home demanding to know where the ganja was hidden. One of the actions available to the police was to beat people up as a form of coercion in order to gain desired information. But the coercion plan had no preconditions regarding believing that someone actually possessed the desired information. In this particular run, nobody in the family knew where the ganja was hidden, resulting in the police systematically beating up every family member.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

