<?xml version="1.0" encoding="ISO-8859-1"?>

<annotation-spec>
 <doc>
 This annotation spectification was developed to label incremental
 understanding in the fruitcarts corpus.  Lexical entries at the
 transcript level should be normalized to fit the domain specific
 vocabulary. This way "apple" is labelled as "tomato" since in this
 domain there is only tomatos, and speaker intents to refer to a tomato
 instead. Equally, a Move action can be realized with many different
 words such as "put", "move", "take", etc. Each of these words are all
 mapped to the Move action if that is the speaker's intentions.

 Labels in most cases should be alinged with some word, clause, or
 utterance in the transcript layer. It may be the case that some
 words required multiple labels, since Anvil does not allow
 overlapping tags, we have created a Merged checkbox which indicates
 that the current label should be merged with the previous unmarked
 label. The last simultaneous label marks the end of the set of
 overlapping labels. 

Version 20: Added plain triangle, plain square, and screen to Unique Id's. 
Version 21: Added autonomous boolean value to the actor layer. 
Version 22: Added exchange objects domain action. This allows to
 change the location (possibly implicit) of two themes. 
Version 23: Added Repeat action. This is a version of "reapply" which
requests the actor to apply a previous action to a new object, color
or location. 

</doc> 

<head>

<valuetype-def>

  <valueset name="AnchorTypes">
    <doc>
    </doc>


    <value-el >Indefinite
     <doc> "a"
     </doc></value-el>

    <value-el >Definite
     <doc> "the"
     </doc></value-el>
 
    <value-el >Pronoun
     <doc> 
     </doc></value-el>
    <value-el >Demonstrative
     <doc>"That"
     </doc></value-el>
    <value-el >Name
     <doc>Region's names
     </doc></value-el>

    <value-el >Definite Anaphoric

     <doc>In cases where "the object", or "the triangle" and such
	 objects have already been described in full, we have a definite
	 description that still refers to a previous description
	 (anaphoric).  </doc></value-el>

    <value-el >Clause Modifier

     <doc> This introduces an achor for a relative clause. This is
     usually a "that". However this word can be ommitted in speech, in
     those cases the first word from the modifier clause.
     </doc></value-el>

    <value-el >Other
     <doc>Please enter in comments. 
     </doc></value-el>
  </valueset>


  <valueset name="Roles">
    <doc>
    </doc>
    <value-el >Part of
     <doc>  Use when there is an object part of
     another object, for example "corner of triangle", also when speaker
     refers to triangle and/or pole (part) of a flag.  
     </doc></value-el>


<!-- CLAUSE RELATED -->
    <value-el >Ending condition
     <doc> Example: "rotate it so that the star is on the right". There are clauses that determine a semantic role for an action. In this case, the final angle for the rotate action. 
	 "rotate to where the right angle is on top". In here "so that" or "to where"
	 is the anchor for the upcoming clause annotated on the object layer (anchor type: clause modifier). 
	 
     </doc></value-el>

    <value-el >Clause Modifier
     <doc>  This role identifies two things, the qualified phrase
    (i.e. "*the* box" where *the* is the anchor) and the qualifying
    clause (i.e. "the box *that* you put ..." where *that* is the value). 
     </doc></value-el>

    <value-el >Content
     <doc> This role identifies the content of the relative clause
     (i.e. the action of the relative clause) and links it to the
     relative clause anchor (i.e.  "the box *that* you put" ). It is also used to mark
Continue Content actions (i.e. the action that is requested to be continued). 
     </doc></value-el>


<!-- OBJECT RELATED -->
    <value-el >Object Type
     <doc> This role is used to identify the type of object being
     described. In the sentence "it is a diamond" we annotate an id-role under 
	 diamond where it links "diamond" as the object type of "it". 
     </doc></value-el>
    <value-el >Object Referent
     <doc> This id-role links the unique-id, value, on the atomic layer with
     a referential expression, anchor, on the object layer.
     </doc></value-el>
    <value-el >Size
     <doc>This identifies the Size
     </doc></value-el>
    <value-el >Decoration 
     <doc>This identifies the type of Decoration 
     </doc></value-el>
<!-- LOCATION RELATED -->
    <value-el >Ground

     <doc>
This id-role links an location entry, anchor, to its ground on the object layer, value. In the sentence "in central park", "in" has an entry "inside" on location layer. This is the anchor of id-role ground. "central park" has an entry on the object layer as "name". This is the value of the id-role ground. 

As another example, if a pen is under the table, the ground for "under" is the table. The anchor for ground relation is "under", and "the table" is the value of the ground role. 

    </doc></value-el>


<!-- ACTION RELATED -->

    <value-el >Agent
     <doc> This role identifies the agent (value), or executor of an
     action (anchor)
     </doc></value-el>

    <value-el >Theme
     <doc> This role is used to identify the action's direct object. The action is the anchor and the object, on the object layer, is the value. It should also be used to link an action as direct object of Continue or Stop. 
     </doc></value-el>
    <value-el >Location
     <doc>This identifies various locations when describing an action (i.e., move), object location, and decoration location. In the case of a Move, the anchor of the id-role location is the Move action. The value is the location entry. In the case of an object location, the anchor is the entry on the object layer, and the value is on the location layer. Similarly for a decoration location description. 
     </doc></value-el>
    <value-el >Distance
     <doc>This identifies the Distance
     </doc></value-el>
    <value-el >Heading
     <doc>This identifies the Heading (i.e. clockwise, counterclockwise)
     </doc></value-el>
    <value-el >Source
     <doc>This identifies the source from which an object is taken. 
     </doc></value-el>
    <value-el >Color
     <doc>This identifies the Color
     </doc></value-el>
    <value-el >Other
     <doc>please describe in comments. 
     </doc></value-el>
  </valueset>




  <valueset name="DomainActions">
    <doc>
    There are four basic domain specific actions that most actor
    speakers consistently performed. These are: Select, Move, Rotate and
    Paint. These are domain specific actions and can be grouped as a
    Request speech act.  Speakers also utter discourse related speech
    acts that manage the progress, refinement, correction and ending
    of ongoing directives. Other actions include mouse pointing, mouse
    release object, mouse motion, etc.


    Domain specific actions should point to the atomic layer where all
	uniquely id's are located.

    </doc>
    <value-el color="light blue">Select
     <doc>
E.g. "see the object xx" is marked as a Select. 
     Select is used for bringing an object into discourse. "take object xx.." is annotated until it is disambiguated (between move or select for instance). 



     </doc></value-el>
    <value-el color="cyan">Move
     
	 <doc>
	 A move action implies a mouse motion where the object is already
	 grabbed.  Therefore on the Actor layer there should be a grabbed
	 action inmediately before a Move action. On the speaker layer, there is no 
need for this requirement. 
	 
     </doc></value-el>

    <value-el color="cyan">Exchange
	 <doc>
	   Exchange action takes two objects and flips their
	   location. Location may be unspecified.
     </doc></value-el>

    <value-el color="cyan">Repeat
	 <doc>
	   Repeat action takes as content the previous action(s) which are
	   to be repeated. New arguments (new theme or location) should
	   still be annotated. Arguments that are the same from previous
	   actions should not. 
     </doc></value-el>

    <value-el color="pink">Rotate
     <doc>

     Rotate (object, location, distance, heading).  Rotates an object
     by a certain distance (abstraction for angular distance "rotate
     it 30 degrees"), heading (for example "rotate the square
     clockwise"), location entry in the location layer is used to
     express  spatial relation ("a little bit more").

     </doc></value-el>
    <value-el color="magenta">Paint
     <doc>
     Paint (object, color). Paints an object with a certain color.
     </doc></value-el>

    <value-el color="light red">Stop
      <doc>Speaker wants to stop the action underway. 
      </doc></value-el>
    <value-el color="light red">Continue
      <doc>Speaker wants to continue the action underway. 
      </doc></value-el>

    <value-el color="light yellow">Point
      <doc> Mouse is over an object. However it has not grab it. 
      </doc></value-el>

    <value-el color="light yellow">Mouse motion
      <doc>Mouse is moving without object. 
      </doc></value-el>

    <value-el color="light blue">Grab
      <doc>Mouse has selected an object and is holding to it but it is
      not moving. E.g. "pick up object xx". There is always a grab action before 
	  any move action. 
      </doc></value-el>

    <value-el color="blue">Release
      <doc>Mouse releases an object which was dragging or grabbing. 
      </doc></value-el>


    <value-el color="yellow">Unspecified
      <doc> Unspecified or uninterpreatable action. Annotator should explain in comments. 
      </doc></value-el>
  </valueset>




  <valueset name="SpeechActs">

   <doc>This is a list of speech acts (SA) performed by the speaker. Whenever a speech act is labelled (as soon as it is unambiguously recognized), the label should be extended until the end of the speech act. This way we know when 
the SA finishes. 

  </doc>

    <value-el color="green">Request
      <doc>
       Speaker can request the actor to perform any of the domain
      specific actions. Request content links to an entry in the
      domain action layer.  
      </doc></value-el>

    
    <value-el color="green">Accept
      <doc> Speaker can accept/confirm an action performed by the
    actor. Accept should link to an entry in the actor track which
      must lay in the past (Before the Accept).
      </doc></value-el>

    <value-el color="orange">Reject

      <doc>Speaker may reject a actor action or their own action (as "in
      forget that"). Content of a Reject action should point to
      actions either at the speaker or actor track. 

      </doc></value-el>

      <value-el color="green">Self-Correct

      <doc> A Self-Correct action points to the speaker track. The
content should point to the id-role which corrects the speaker's
requested domain action.  </doc>

</value-el>

<!-- action in the actor track. If it is  a self-correct, -->
<!-- where speaker is correcting him/herself, then action will point to an -->
<!-- speaker action in the speaker track. -->

    <value-el color="light green">Refine 

<doc> A refine action's content is pointing to the id-role entry. The id-role entry in turn points to the previous action that the id-role is elaborating. For instance, we have two utterances: "move the square to the right" and "a little bit more". The first utterance is a move action. The second one has a location id-role and a refine action. 
The refine action points to the elaborating id-role ("a little bit more"). 


This action refines arguments of previously given actions. Single
      utterances such as "a little bit lower" would be labelled as
      Refine actions. Refine content should point to the action of
      whoever had the last turn in the dialogue. This means that such
      action, either speaker or actor layer, is being elaborated.


Refine label cover sentences which present ellipsis phenomena. Usually
these sentence will further restrict arguments given in previous
utterances or fill previously empty arguments.

Notice that the new utterance, either a "a little bit lower" or "keep
going" are not contradictory with previous actions. If a contradiction
were the case, the new utterance would be a Correct or Self-Correct
action. This would distinguis Refine from Correct. 

      </doc></value-el>
    <value-el color="gray">Apology
      <doc> Speaker apologizes. Does not need a content link. 
      </doc></value-el>
    <value-el color="gray">Abandon
      <doc> Utterance is abandoned. 
      </doc></value-el>
    <value-el color="dark gray">Monolog
      <doc> Utterance is not intended for actor interpretation.
       Example "Let's see".
      </doc></value-el>

    <value-el color="yellow">Unspecified
      <doc> Unspecified or uninterpreatable action. Annotator should explain in comments. 
      </doc></value-el>
  </valueset>




  <valueset name="Transactions">
    <doc> This track (layer) encompases all actions performed to
    modify one simple object in three ways an object was moved,
    rotated or painted. Transaction may be viewed as the highest level
    summary of what happened at the end of a set of actions which will
    include accept, reject, refines, etc. 
    </doc>
    <value-el color="cyan">Move
      <doc>Move transaction.
      </doc></value-el>
    <value-el color="pink">Rotate
      <doc>Rotate transaction.
      </doc></value-el>
    <value-el color="magenta">Paint
      <doc>Paint transaction.
      </doc></value-el>
  </valueset>



  <valueset name="ObjectTypes">
    <doc>Object Type: (unique-id, triangle, square, flag, banana, etc). 
    </doc>


    <value-el>Unique ID
      <doc> If object is refered with unique ID, then annotate its  
      corresponding value. 
      </doc></value-el>

<!--   generic objects  -->
    <value-el>Square
      <doc>
      Square. If unique-id object can not be inferred, then annotate 
      all attributes that describe the object as speaker provides information.
      </doc></value-el>
    <value-el>Triangle
      <doc> Big Triangle
      </doc></value-el>
    <value-el>Flag
      <doc>
      </doc></value-el>

    <value-el>Flag_Pole
      <doc>
      </doc></value-el>

<!-- DECORATION TYPE -->
    <value-el> Star
      <doc>
      </doc></value-el>
    <value-el> Heart
      <doc>
      </doc></value-el>
    <value-el> Diamond
      <doc>
      </doc></value-el>
    <value-el> Circle
      <doc>Also "dot"
      </doc></value-el>


<!-- DECORATION LOCATION -->
    <value-el> Sqr_Corner
      <doc>
      </doc></value-el>

    <value-el> Right_Angle_Corner
      <doc>
      </doc></value-el>
    <value-el> Acute_Angle_Corner
      <doc>
      </doc></value-el>
    <value-el> Side
      <doc>
      </doc></value-el>
    <value-el> Hypothenuse
      <doc>
      </doc></value-el>

<!--   fruits  -->
    <value-el>Avocado
      <doc>
      </doc></value-el>
    <value-el>Banana
      <doc>
      </doc></value-el>
    <value-el>Cucumber
      <doc>
      </doc></value-el>
    <value-el>Grapefruit
      <doc>
      Same as "lemon"
      </doc></value-el>
    <value-el>Tomato
      <doc>Same as "apple"
      </doc></value-el>
    <value-el>Other
      <doc>
      </doc></value-el>

  </valueset>




  <valueset name="Objects">
    <doc> These are uniquely identified objects (unique-ids). 
    </doc>

<!--  regions -->
    <value-el>Base
      <doc>
      </doc></value-el>

    <value-el>Color Buttons
      <doc>
      </doc></value-el>
    <value-el>Rotate Buttons
      <doc>
      </doc></value-el>
    <value-el>Central Park
      <doc>
      </doc></value-el>
    <value-el>Forest Hills
      <doc>
      </doc></value-el>
    <value-el>Morningside
      <doc>
      </doc></value-el>
    <value-el>Morningside Heights
      <doc>
      </doc></value-el>
    <value-el>Ocean View
      <doc>
      </doc></value-el>
    <value-el>Ocean View Terrace
      <doc>
      </doc></value-el>
    <value-el>Pine Tree Mountain
      <doc>
      </doc></value-el>
    <value-el>Snake River Delta
      <doc>
      </doc></value-el>
    <value-el>Screen
      <doc>
      </doc></value-el>


<!-- end regions -->

<!--  Flags in three regions -->

    <value-el>CentralPark_Flag
      <doc>
      </doc></value-el>

    <value-el>MorningSideHeights_Flag
      <doc>
      </doc></value-el>

    <value-el>PineTreeMountain_Flag
      <doc>
      </doc></value-el>


<!--   agent values  -->
    <value-el>Speaker
      <doc>Mark when speaker says "I"
      </doc></value-el>
    <value-el>Actor
      <doc>Mark when speaker says "you". 
      </doc></value-el>
    <value-el>Speaker-Actor
      <doc>Mark when speaker says "we" :-)
      </doc></value-el>


<!--   geometric figures -->
<!--       SQUARES    -->
    <value-el>_SqrDs
      <doc>
      Big Square with a Diamond on the Side 
      </doc></value-el>
    <value-el>_SqrDc
      <doc>
      Big Square with a Diamond on the Corner
      </doc></value-el>
    <value-el>sSqrDs
      <doc>
      </doc></value-el>
    <value-el>sSqrDc
      <doc>
      </doc></value-el>
    <value-el>_SqrHs
      <doc> Big Square with Heart on the Side
      </doc></value-el>
    <value-el>_SqrHc
      <doc>
      </doc></value-el>
    <value-el>sSqrHs
      <doc>
      </doc></value-el>
    <value-el>sSqrHc
      <doc>
      </doc></value-el>

    <value-el>_Sqr__
      <doc>Plain Big square
      </doc></value-el>
    <value-el>sSqr__
      <doc>Plain Small square
      </doc></value-el>

<!--       TRIANGLE    -->
    <value-el>_TriSs
      <doc> Big Triangle with Star on the Side (or hypothenuse)
      </doc></value-el>
    <value-el>sTriSs
      <doc>  Small Triangle  with a Star on the Side (or hypothenuse)
      </doc></value-el>
    <value-el>_TriSc
      <doc>
      </doc></value-el>
    <value-el>sTriSc
      <doc>
      </doc></value-el>
    <value-el>_TriCs
      <doc>
      </doc></value-el>
    <value-el>sTriCs
      <doc>
      </doc></value-el>
    <value-el>_TriCc
      <doc>
      </doc></value-el>
    <value-el>sTriCc
      <doc>
      </doc></value-el>
    <value-el>_Tri__
      <doc>Big Plain Triangle
      </doc></value-el>

    <value-el>sTri__
      <doc>Small Plain Traingle
      </doc></value-el>


    <value-el>Other
      <doc>
       Out of domain: when speaker says something that actor can't
      do. Please example in comments. 
      </doc></value-el>
  </valueset>


  <valueset name="ActorLocations">
    <doc> These are locations for the Actor layer. 
    </doc>

<!--  Flags in three regions -->

    <value-el>CentralPark_Flag
      <doc>
      </doc></value-el>

    <value-el>MorningSideHeights_Flag
      <doc>
      </doc></value-el>

    <value-el>PineTreeMountain_Flag
      <doc>
      </doc></value-el>

<!--  regions -->


    <value-el>Screen
      <doc>
      </doc></value-el>

    <value-el>Base
      <doc>
      </doc></value-el>

    <value-el>Color Buttons
      <doc>
      </doc></value-el>
    <value-el>Rotate Buttons
      <doc>
      </doc></value-el>
    <value-el>Central Park
      <doc>
      </doc></value-el>
    <value-el>Forest Hills
      <doc>
      </doc></value-el>
    <value-el>Morningside
      <doc>
      </doc></value-el>
    <value-el>Morningside Heights
      <doc>
      </doc></value-el>
    <value-el>Ocean View
      <doc>
      </doc></value-el>
    <value-el>Ocean View Terrace
      <doc>
      </doc></value-el>
    <value-el>Pine Tree Mountain
      <doc>
      </doc></value-el>
    <value-el>Snake River Delta
      <doc>
      </doc></value-el>
<!-- end regions -->

<!--   geometric figures -->
<!--       SQUARES    -->

    <value-el>_Sqr__
      <doc>Plain Big square
      </doc></value-el>
    <value-el>sSqr__
      <doc>Plain Small square
      </doc></value-el>

    <value-el>_SqrDs
      <doc>
      Big Square with a Diamond on the Side 
      </doc></value-el>
    <value-el>_SqrDc
      <doc>
      Big Square with a Diamond on the Corner
      </doc></value-el>
    <value-el>sSqrDs
      <doc>
      </doc></value-el>
    <value-el>sSqrDc
      <doc>
      </doc></value-el>
    <value-el>_SqrHs
      <doc> Big Square with Heart on the Side
      </doc></value-el>
    <value-el>_SqrHc
      <doc>
      </doc></value-el>
    <value-el>sSqrHs
      <doc>
      </doc></value-el>
    <value-el>sSqrHc
      <doc>
      </doc></value-el>

<!--       TRIANGLE    -->


    <value-el>_Tri__
      <doc>Big Plain Triangle
      </doc></value-el>

    <value-el>sTri__
      <doc>Small Plain Traingle
      </doc></value-el>

    <value-el>_TriSs
      <doc> Big Triangle with Star on the Side (or hypothenuse)
      </doc></value-el>
    <value-el>sTriSs
      <doc>  Small Triangle  with a Star on the Side (or hypothenuse)
      </doc></value-el>
    <value-el>_TriSc
      <doc>
      </doc></value-el>
    <value-el>sTriSc
      <doc>
      </doc></value-el>
    <value-el>_TriCs
      <doc>
      </doc></value-el>
    <value-el>sTriCs
      <doc>
      </doc></value-el>
    <value-el>_TriCc
      <doc>
      </doc></value-el>
    <value-el>sTriCc
      <doc>
      </doc></value-el>

<!--  End of geometric figures -->

  </valueset>

  <valueset name="ActorObjects">
    <doc> These are uniquely identified objects (unique id's). 
    </doc>


<!--   fruits  -->
    <value-el>Avocado
      <doc>
      </doc></value-el>
    <value-el>Banana
      <doc>
      </doc></value-el>
    <value-el>Cucumber
      <doc>
      </doc></value-el>
    <value-el>Grapefruit
      <doc>
      Same as "lemon"
      </doc></value-el>
    <value-el>Tomato
      <doc>Same as "apple"
      </doc></value-el>

<!--   geometric figures -->
<!--       SQUARES    -->
    <value-el>_SqrDs
      <doc>
      Big Square with a Diamond on the Side 
      </doc></value-el>
    <value-el>_SqrDc
      <doc>
      Big Square with a Diamond on the Corner
      </doc></value-el>
    <value-el>sSqrDs
      <doc>
      </doc></value-el>
    <value-el>sSqrDc
      <doc>
      </doc></value-el>
    <value-el>_SqrHs
      <doc> Big Square with Heart on the Side
      </doc></value-el>
    <value-el>_SqrHc
      <doc>
      </doc></value-el>
    <value-el>sSqrHs
      <doc>
      </doc></value-el>
    <value-el>sSqrHc
      <doc>
      </doc></value-el>

<!--       TRIANGLE    -->
    <value-el>_TriSs
      <doc> Big Triangle with Star on the Side (or hypothenuse)
      </doc></value-el>
    <value-el>sTriSs
      <doc>  Small Triangle  with a Star on the Side (or hypothenuse)
      </doc></value-el>
    <value-el>_TriSc
      <doc>
      </doc></value-el>
    <value-el>sTriSc
      <doc>
      </doc></value-el>
    <value-el>_TriCs
      <doc>
      </doc></value-el>
    <value-el>sTriCs
      <doc>
      </doc></value-el>
    <value-el>_TriCc
      <doc>
      </doc></value-el>
    <value-el>sTriCc
      <doc>
      </doc></value-el>

  </valueset>




  <valueset name="DecorationTypes">
    <doc> 
    </doc>
    <value-el>Star
      <doc>
      </doc></value-el>
    <value-el>Diamond
      <doc>
      </doc></value-el>
    <value-el>Heart
      <doc>
      </doc></value-el>
    <value-el>Other
      <doc>Please put in comments
      </doc></value-el>
  </valueset>



  <valueset name="DecorationLocations">
    <doc> Decoration Location: (side, corner)
    </doc>
    <value-el>Side
      <doc>
      </doc></value-el>
    <value-el>Hypothenuse
      <doc>
      </doc></value-el>
    <value-el>Corner
      <doc>
      </doc></value-el>
    <value-el>Other
      <doc>Please write in comments
      </doc></value-el>
  </valueset>



  <valueset name="VerticalDirection">

  <doc>
  This is the spatial relation list for the Vertical dimension.
  </doc>

     <value-el>Up
      <doc>
      </doc></value-el>
     <value-el>V_middle
      <doc>
      </doc></value-el>
     <value-el>Down
      <doc>
      </doc></value-el>
     <value-el>Other
      <doc>Please explain in comments. 
      </doc></value-el>
  </valueset>


  <valueset name="HorizontalDirection">
  <doc>
  This is the spatial relation list for the Horizontal dimension.
  </doc>
     <value-el>Left
      <doc>
      </doc></value-el>
     <value-el>H_middle
      <doc>
      </doc></value-el>
     <value-el>Right
      <doc>
      </doc></value-el>
     <value-el>Other
      <doc>Please explain in comments. 
      </doc></value-el>
  </valueset>



  <valueset name="TopologicalProperties">
  <doc>
  This is a list of topological properties that act as modifiers for the 
spatial relation; both horizontal and vertical dimension.
  </doc>
     <value-el>Inside
      <doc>
      </doc></value-el>
     <value-el>In-Adjacent
      <doc>
      </doc></value-el>
     <value-el>Overlap
      <doc>
      </doc></value-el>
     <value-el>Outside
      <doc>
      </doc></value-el>
     <value-el>Out-Adjacent
      <doc>Implies being outside of ground region/object. 
      </doc></value-el>
     <value-el>Other
      <doc>Please explain in comments. 
      </doc></value-el>
  </valueset>



  <valueset name="RelationModifiers">
    <doc> This is a list of modifiers for spatial relations in the fruitcarts domain.
    </doc>
  <value-el> aligned
      <doc>
      </doc></value-el>
  
  <value-el>Not
      <doc>
      </doc></value-el>
    <value-el>A little more
      <doc>Same as "A bit more".
      </doc></value-el>
    <value-el>A little less
      <doc>Same as "A bit less".
      </doc></value-el>
    <value-el>Exactly
      <doc>This means the smallest distance 
possible between ground and object. To be used for an object "on top of". Also for instructions
    "exactly to the right", "exactly to the left". 
      </doc></value-el>

    <value-el>Other
      <doc>

      </doc></value-el>

  </valueset>



  <valueset name="ActorHeadings">
    <doc>Heading is an attribute of Rotate. 
    </doc>
    <value-el>Right
      <doc>Same as "right". It is also used for clockwise rotation. 
      </doc></value-el>
    <value-el>Left
      <doc>Same as "left". Also used for counter clockwise rotation. 
      </doc></value-el>
  </valueset>



  <valueset name="Sizes">
    <doc>
    </doc>
    <value-el>Big
      <doc>
      </doc></value-el>

    <value-el>Small
      <doc>
      </doc></value-el>

  </valueset>

  <valueset name="Colors">
    <doc>If speaker says a color that is clearly one of the
    possibilities, such as magenta = purple, 
peach = orange, select one of the six colors. Otherwise
if color not listed, select "other", and add comments.
    </doc>
    <value-el>Black
      <doc>
      </doc></value-el>
    <value-el>Brown
      <doc>
      </doc></value-el>
    <value-el>Orange
      <doc>
      Same as "peach".
      </doc></value-el>
    <value-el>Blue
      <doc>
      </doc></value-el>
    <value-el>Pink
      <doc>
      </doc></value-el>

    <value-el>Yellow
      <doc>Not an option for a Paint action, but used to described
      objects. 
      </doc></value-el>
    <value-el>Magenta
      <doc>Same as "purple"
      </doc></value-el>
    <value-el>Other
      <doc>
      When the speaker referes a color that doesn't exist and they
    clearly didnt mean one of the six possible colors. Add comments to explain.
      </doc></value-el>
    <value-el color="dark yellow">Unspecified
      <doc>Color ellipsis. 
      </doc></value-el>
  </valueset>

 <valueset name="OverlapChoices">
  
  <doc> 



    </doc>
    <value-el>Overlapping
      <doc>
If Overlapping is chosen, this label is on top of a previously introduced labe\
l. However,
different from simultaneous, it does not share the start and end point
of the previous label. The time stamps for this label are indeed the begining
and end of this label only, and not the group of overlapping labels.
      </doc></value-el>
    <value-el>Simultaneous
      <doc>
    If Simultaneous option is checked, the begining and end of the
simultaneous label group are the same for all labels in the
group. This is these labels are simultaneous. The first label of the group
should be unmarked, and all following in the group should be marked.      </doc></value-el>
  </valueset>
  <valueset name="ActorColors">
    <doc>
    </doc>
    <value-el>Black
      <doc>
      </doc></value-el>
    <value-el>Brown
      <doc>
      </doc></value-el>
    <value-el>Orange
      <doc>
      Same as "peach".
      </doc></value-el>
    <value-el>Blue
      <doc>
      </doc></value-el>
    <value-el>Pink
      <doc>
      </doc></value-el>
    <value-el>Magenta
      <doc>Same as "purple"
      </doc></value-el>
  </valueset>

</valuetype-def>

</head>

<body>

<!--   <track-spec name="Audio" type="waveform" height="1" > -->
<!--     <doc> -->
<!--     </doc> -->
<!--   </track-spec> -->

<!--   <track-spec name="Silence" type="primary" edit-takeover="false" > -->
<!--     <doc> -->
<!--     </doc> -->
<!--     <attribute name="silence" valuetype="String" display="false"> -->
<!--       <doc></doc></attribute> -->
<!--   </track-spec> -->


<!--**************   SPEAKER LAYER  -->


<group name="Speaker">
    <doc> This track has specified only the necessary attributes for
    each action.

Utterances such as "we want a xxxx in yyyy ..." show that at the point
of hearing "we want" we are unaware which action speaker wants the actor
to perform (i.e. it could be a move or a paint).
  The disambiguation of intended action occurs later in the utterance.
This means that the actual annotation of the intented action must
occur when there is no doubt of what the speaker meant.
    </doc>

  <track-spec name="Transcript" type="primary" edit-takeover="false" >
    <doc>
    </doc>
    <attribute name="utterance" valuetype="String" display="true" emptyvalue="false" defaultvalue="" >
      <doc>
      </doc></attribute>
  </track-spec>



<!--*********** OBJECT LAYER *****************-->

  <track-spec name="Object Layer" type="primary" edit-takeover="true" color-attr="action" >
      <doc>  
      To incrementally annotate an object in this domain, we make use
      of an anchor at the beggining of the refering expression. This
      anchor will be the first word uttered at the begging of the
      object (i.e. "a", "the", etc.). 
      </doc>  


    <attribute name="Anchor Type" valuetype="AnchorTypes" display="true"  >
    <doc>
      First word in a reffering expression will act as the anchor
      which will be incrementally described (i.e. the articles "a" or
      "the" will work as an anchor). 
      Chose the appropiate anchor type e.g. indefinite, definite,
      pronoun, name, etc. 
       </doc></attribute>

    <attribute name="Simultaneous" valuetype="Boolean" display="true"  >

    <doc>If this option is checked, the begining and end of the
simultaneous label group are the same for all labels in the
group. This is these labels are simultaneous. The first label of the group 
should be unmarked, and all following in the group should be marked. 
  </doc></attribute>


  </track-spec>



<!--**************   LOCATION LAYER  -->

  <track-spec name="Location Layer" type="primary" edit-takeover="true" color-attr="action" >
    <doc> 	

Location layer has the following attributes: Vertical Direction, Horizontal Direction, 
Topological Properties and Relation Modifiers. This layer describes the spatial 
relation between two objects; an object and its ground. It is used to describe 
location semantic roles for Move and Rotate actions. 
    </doc>
    <attribute name="Vertical Direction" valuetype="VerticalDirection"  display="true" >
      <doc>
      </doc></attribute>

    <attribute name="Horizontal Direction" valuetype="HorizontalDirection"  display="true" >
      <doc>
      </doc></attribute>

    <attribute name="Topological Properties" valuetype="TopologicalProperties"  display="true" >
      <doc>
      </doc></attribute>

    <attribute name="RelationModifier" valuetype="RelationModifiers" display="true"  >
      <doc>
      </doc></attribute>

  </track-spec>

<!--**************   ATOMIC LAYER  -->

  <track-spec name="Atomic Layer" type="primary" edit-takeover="true" color-attr="action" >
    <doc> 	
This layer has a list of all possible role values not included in
location and object layer. Thus we have: Distance, Heading
and Endpointorientation. The following already have entries at the
object layer: Color, Decoration type, Decoration location, Size.

    </doc>


    <attribute name="Object Type" valuetype="ObjectTypes" display="true" >
      <doc>  
      </doc></attribute>

    <attribute name="Unique ID" valuetype="Objects" display="true"  >
      <doc> Mark unique id when object is disambiguated. Then link it
      to the anchor with an id-role triplet. 
      </doc></attribute>


    <attribute name="Distance" valuetype="String"  display="true" >
      <doc>Please enter the numeric distance or angle distance.
      </doc></attribute>

    <attribute name="Color" valuetype="Colors" display="true"  >
      <doc>
      </doc></attribute>

    <attribute name="Size" valuetype="Sizes" display="true"  >
      <doc>
      </doc></attribute>

    <attribute name="EndPointOrientation" valuetype="String" display="true" >

      <doc>This attribute handles cases such as "rotate triangle so
      that star on the corner is on top" and "rotate it like a martini
      glass" </doc></attribute>


    <attribute name="Simultaneous" valuetype="Boolean" display="true"  >
    <doc>If this option is checked, the begining and end of the
simultaneous label group are the same for all labels in the
group. This is these labels are simultaneous. The first label of the group 
should be unmarked, and all following in the group should be marked. 
  </doc></attribute>



  </track-spec>


<!-- **************   ID-ROLE LAYER -->
   <track-spec name="Id-Role" type="primary" color-attr="speech act" edit-takeover="true" >
     <doc> This layer identifies the roles that go with and link to
   objects or actions. If there is a role relation between to objects one of them is the anchor and the other is the role value. 
To know which one is which, one can say "the role of [something] is [that]". Then [something] is the anchor, 
and [that] is the role value.  

            Id-role has the following three attributes: 
                Anchor (link to object or action whose role we are attaching);
		Role Relation    (obj, src, destination ...  all semantic roles);
		Role Value (appropiate value);


     </doc>

     <attribute name="Anchor" valuetype="ReciprocalLink"  display="true"  link-color="red" >
       <doc>The anchor can be on the Object, Domain
           Action, or Actor Layer. For instance, when describing the final location for a move action. The action entry is the anchor. The value is the location entry. If we are describing the location of an object, the object itself is the anchor (the article 'the' for instance). 
       </doc></attribute>


     <attribute name="Role Relation" valuetype="Roles"  display="true"  link-color="red" >
       <doc>
       This identifies the role being instatiated by the role value. 
       </doc></attribute>


     <attribute name="Role Value" valuetype="MultiLink"  display="true"  link-color="red" >
       <doc>
     Roles value fills the role relation of the anchor. E.g. "Blue" is
       the value of the "color" role of some object or action. 
       </doc></attribute>


    <attribute name="Simultaneous" valuetype="Boolean" display="true"  >

    <doc>If this option is checked, the begining and end of the
simultaneous label group are the same for all labels in the
group. This is these labels are simultaneous. The first label of the group 
should be unmarked, and all following in the group should be marked. 
  </doc></attribute>


   </track-spec>

<!-- **************  DOMAIN ACTION -->

  <track-spec name="Domain Actions" type="primary" edit-takeover="true" color-attr="action" >
    <doc>
    </doc>
    <attribute name="Action" valuetype="DomainActions" defaultvalue="Select" emptyvalue="false" display="false" >
      <doc>Annotate which action the speaker intended at the first word that
disambiguates such action. In examples such as "there needs
to be" is an ambigous, thus we annotate at the start of an argument
which disambiguates it fully.

      </doc></attribute>
  </track-spec>

<!-- **************  previously LINK ACTION -->
   <track-spec name="Speech Act" type="primary" color-attr="speech
   act" edit-takeover="true" >

     <doc> These are speech acts which may or may not refer to other
   actions. If they do, link to the former action. Request speech acts
   must have a domain action specified on the layer "Domain Actions". 
     </doc>
     <attribute name="Speech Act" valuetype="SpeechActs"
   emptyvalue="false" display="true" defaultvalue="Accept" >

       <doc> 
    Speech acts are labelled at the point where it is identified. 
However they span over all the domain specific actions 
until speech act terminates. This means that SA are 
not annotated incrementally. We dont mark the start of 
the speech act but we do mark the end. 

       </doc></attribute>


     <attribute name="Content" valuetype="MultiLink"
 display="true"  link-color="red" >
 
       <doc>

Content of SA request is the domain action entry.

Content of SA reject points to actions only.

Content of SA self-correct points to id-role only.

Content of SA apology does not require a link.

Content of SA refine points to id-role which is elaborating the action. 
	Add "in response to" link to the actor action. 

     These actions refer to a another action to which they are linked.
      These are Accept, Reject, Undo, Refine, Stop,
     etc. An argument can have a different value from the original
     speech act.
       </doc></attribute>


    <attribute name="Overlapping" valuetype="OverlapChoices" display="true"  >

    <doc> If Overlapping is chosen, this label is on top of a previously introduced label. However,
different from simultaneous, it does not share the start and end point
of the previous label. The time stamps for this label are indeed the begining 
and end of this label only, and not the group of overlapping labels. 
    If Simultaneous option is checked, the begining and end of the
simultaneous label group are the same for all labels in the
group. This is these labels are simultaneous. The first label of the group 
should be unmarked, and all following in the group should be marked. 
  </doc></attribute>




   </track-spec>

 </group>


<!--********************   Actor Layer  **************************   -->

  <track-spec name="Actor" type="primary" color-attr="action" edit-takeover="false" >
    <doc>
    </doc>

    <attribute name="Speaker Action" valuetype="MultiLink" emptyvalue="true" display="true"  >
      <doc> Use this only to save annotation. If the actor action was
      not requested, or different from speaker request in some
      attribute, then create a new action using attributes at this
      layer.  </doc></attribute>

    <attribute name="Actor Action" valuetype="DomainActions"
    emptyvalue="true" display="true" >
      <doc> Annotate this attribute when information can *not* be linked
      from speaker layer. 
      </doc></attribute> 

    <attribute name="Autonomous" valuetype="Boolean" emptyvalue="true" display="false"  >
      <doc>
	An autonomous actor action is one which was unsolicited by the speaker. 
  </doc></attribute>


 <attribute name="Object"  valuetype="ActorObjects" display="true" emptyvalue="true"  >
  <doc>Actor only acts on uniquely identified objects. 
  </doc></attribute>

    <attribute name="Destination Location" valuetype="ActorLocations" display="true" >
      <doc>  Location will link to an entry in the location layer
      which specifies region, spatial relation, and ground object. 
      </doc></attribute>


    <attribute name="Distance" valuetype="String"  emptyvalue="true" display="true" >
      <doc>Distance can either be linear or angular. 
      </doc></attribute>

    <attribute name="Heading" valuetype="ActorHeadings"  emptyvalue="true" display="true" >
      <doc>Heading of mouse-motion or rotation.  
      </doc></attribute>
    

    <attribute name="Color" valuetype="ActorColors"  emptyvalue="true" display="true" >
      <doc>
      </doc></attribute>

</track-spec>


<!--********************   Transaction Layer  ************************** -->
<!---->

  <track-spec name="Transactions" type="primary" edit-takeover="true" color-attr="action" >
 <doc>
 </doc>

 <attribute name="Success" valuetype="Boolean" display="true" emptyvalue="false" defaultvalue="true" >
  <doc>
  </doc></attribute>


 <attribute name="Transaction" valuetype="Transactions" display="true"
 emptyvalue="true"  >

  <doc> This track (layer) encompases all actions performed to
    modify one simple object in three ways an object was moved,
    rotated or painted. Transaction may be viewed as the highest level
    summary of what happened at the end of a set of actions which may
    include accept, reject, refines, etc. 

    
  </doc></attribute>


     <attribute name="Action" valuetype="ReciprocalLink"
     display="true" link-color="orange" > 

       <doc> The speaker and actor will both specify most of attributes
       if not all of the transaction. Therefore no double annotation
       is needed when transaction was succesful. 
       </doc></attribute>

 <attribute name="Object"  valuetype="MultiLink" display="true"
       emptyvalue="true"  >

  <doc>Points to object layer. 
  </doc></attribute>

    <attribute name="Source Location" valuetype="MultiLink" display="true" >
      <doc>  Location will link to an entry in the location layer
      which specifies region, spatial relation, and ground object. 
      </doc></attribute>

    <attribute name="Destination Location" valuetype="MultiLink" display="true" >
      <doc>  Location will link to an entry in the location layer
      which specifies region, spatial relation, and ground object. 
      </doc></attribute>


    <attribute name="Distance" valuetype="String"  emptyvalue="true" display="true" >
      <doc>Distance can either be linear or angular. 
      </doc></attribute>

    <attribute name="Color" valuetype="Colors"   emptyvalue="true" display="true" >
      <doc>
      </doc></attribute>
  </track-spec>


</body>

</annotation-spec>


