A. Lexicon.dtd

Note: this file is slightly out of date because of a number of recent changes. Around the end of January 2002 a final version is expected to be available at http://bcxml.sourceforge.net/

<!-- =================================================== -->
<!-- DTD for XML files holding lexicon data -->
<!-- version 0.3 of 2001-05-02 -->
<!-- Changed by Reinout van Rees (R.vanRees@ct.tudelft.nl) to -->
<!-- accommodate some changes in the lexicon xml export, like -->
<!-- nesting of subjects and properties -->
<!-- version 0.2 -->
<!-- Changed by Reinout van Rees (R.vanRees@ct.tudelft.nl) -->
<!-- version 0.1 -->
<!-- Started by Kees Woestenenk (kwoestenenk@stabu.nl) -->
<!-- =================================================== -->

<!-- Let's start with an explanation of the various -->
<!-- declarations below: -->
<!-- <!ELEMENT [ElementName] [AllowedContent]> -->
<!-- <!ENTITY % [EntityName] '[ReplacementText]'> -->
<!-- <!ATTLIST [ElementName] 
	[AttributeName]	[AttributeType] [Modifier/default]
	[AttributeName]	[AttributeType] [Modifier/default]
	...> -->

<!-- On naming: elements are CamelCase, attributes are -->
<!-- lowerCamelCase. -->
<!-- An attribute "somethingRefs" contains ID's of -->
<!-- "Something"s with which the element has some sort of -->
<!-- relation. <Subject propertyRefs="P1 P34"> means that -->
<!-- <Property id="P1"> and <Property id="P34"> are -->
<!-- properties of this subject. -->

<!-- "Lexicon" is the top-level element with as it's -->
<!-- children a number of elements containing the actual -->
<!-- information. All elements directly below "Lexicon" are the -->
<!-- top level elements of their tree structure. The elements -->
<!-- below it are interconnected -->
<!-- using ID's and IDREF's like the properties that are -->
<!-- attached to subjects. -->

<!ELEMENT Lexicon	(Subject, Activity, Propcollection,
			Property, Measure, Unit*, Reference ) >

<!-- A combination of subelements that's quite common is -->
<!-- that of one or more "Longname"s, optional -->
<!-- "Description"s and optional "Filename"s. In some cases -->
<!-- this is extended by optional "Shortname"s. -->
<!-- This calls for two reusable templates. -->
<!-- Apart from the initial "Longname", the elements may be -->
<!-- freely mixed. -->
<!-- LDF = "Longname, Description, Filename" -->
<!-- LSDF = "Longname, Shortname, Description, Filename"--> 

<!ENTITY % LDF 	'(Longname, (Longname| Description|
		 Filename)* )' > 

<!ENTITY % LSDF	'(Longname, (Longname| Shortname|
		Description| Filename)* )' > 

<!-- Something that is also used in various places is a -->
<!-- "language" attribute. It is good to specify this in a -->
<!-- central place to be able to add languages more -->
<!-- easily. The default language is "en" (english). In version -->
<!-- 0.3 the allowed element content was changed from a list of -->
<!-- possible languages to just plain NMTOKEN, thereby ditching -->
<!-- the need to adapt this file everytime a language gets -->
<!-- added. -->

<!ENTITY % language	'
	language	NMTOKEN	"en"' >  

<!-- There is a standard set of attributes attached to most -->
<!-- elements. -->
<!-- id = an unique id. In xml it cannot start with a -->
<!-- number, so a prefix has to be added. As the ID's aren't -->
<!-- globally unique in the lexicon, the prefix indicates the -->
<!-- kind of element like 'S' for Subject. -->
<!-- version = "1.0" or something like that. -->
<!-- date = date of the last change in 2001-02-14 format. -->
<!-- status = signifies the status of the element (see -->
<!-- possible values for a clue). -->
<!-- update = kind of a -->
<!-- status-after-the-last-big-version. --> 

<!ENTITY % standard	'
	id		ID	#REQUIRED
	version		CDATA	"1"
	date		CDATA	""
	status		(-|Draft|Release|Deleted)
				"-" 
	update		(-|New|Updated|Expired)
				"-" '>

<!-- Now follow the elements that are allowable directly -->
<!-- underneath the "lexicon" top-level element. -->

<!-- A subject is an object or a service, like door or -->
<!-- paintjob. -->
<!-- The nested subjects should be placed last, behind the -->
<!-- "longnames" etc. -->
<!-- WARNING: The attribute "componentRefs" points towards -->
<!-- multiple Components which themselves are "Subject"s, so -->
<!-- implementors should keep in mind that it points towards -->
<!-- "Subject"s (despite the name "componentRefs") but that -->
<!-- the "Subject"s pointed towards are components. -->
<!ELEMENT Subject	(%LDF;,Localproperty*,Subject*)>
<!ATTLIST Subject	
	%standard; 
	componentRefs	IDREFS	#IMPLIED
	propcollectionRefs
			IDREFS	#IMPLIED
	referenceRefs	IDREFS	#IMPLIED >

<!ELEMENT Activity	(%LDF;,Activity*)>
<!ATTLIST Activity
	%standard; >

<!-- A Localproperty is a mechanism to attach multiple -->
<!-- properties of the same kind to a subject, for instance to -->
<!-- define multiple heights h1, h2, h3. Those extra -->
<!-- identicators (h1, h(w), etc.) can be put into this -->
<!-- element's content. -->
<!ELEMENT Localproperty	(#PCDATA)>
<!ATTLIST Localproperty	
	propertyRef	IDREF	#IMPLIED
	extnr		NMTOKEN "0"
	extlanguage	NMTOKEN "en" >

<!-- A propcollection is a collection of properties, grouped -->
<!-- together for some purpose. -->
<!-- The nested propcollections should be placed last, behind -->
<!-- the "longnames" etc. -->
<!ELEMENT Propcollection (%LDF;,Localproperty*,Propcollection*)>
<!ATTLIST Propcollection	
	%standard;
	referenceRefs	IDREFS	#IMPLIED >

<!-- A property is something like "height" "length" -->
<!-- "fire-resistance". -->
<!-- The nested properties should be placed last, behind the -->
<!-- "longnames" etc. -->
<!-- WARNING: The attribute "measureRefs" is required, so -->
<!-- you ought to connect at least one Measure to a -->
<!-- property. -->
<!ELEMENT Property	(%LSDF;,Property*)>
<!ATTLIST Property
	%standard; 
	measureRefs	IDREFS	#IMPLIED
	referenceRefs	IDREFS	#IMPLIED >

<!-- A measure is a quantification of a property. This means -->
<!-- that a property derives it's meaning from the attached -->
<!-- measures. -->
<!-- The nested measures should be placed last, behind the -->
<!-- "longnames" etc. -->
<!-- WARNING: The attribute "unitRef" references a *SINGLE* -->
<!-- unit, not multiple! -->
<!ELEMENT Measure	(%LDF;,Value*,Measure*)>
<!ATTLIST Measure
	%standard;
	unitRefs	IDREF	#IMPLIED
	valueRefs	IDREFS	#IMPLIED
	referenceRefs	IDREFS	#IMPLIED
	type		(Single|Enumeration|Bounded|Range) 
				"Single">

<!-- A unit is something like "meter" "inch" "kelvin" -->
<!-- The nested units should be placed last, behind the -->
<!-- "longnames" etc. -->
<!ELEMENT Unit		(%LSDF;,Unit*)>
<!ATTLIST Unit
	%standard; 
	referenceRefs	IDREFS	#IMPLIED>

<!-- References are the mechanisms to construct a different -->
<!-- view on the lexicon's contents. You can use it to -->
<!-- attach for instance classification system's names to -->
<!-- subjects in the lexicon. -->
<!-- References can be nested. -->
<!-- The actual data is in the "description" -->
<!ELEMENT Reference (Reference*)>
<!ATTLIST Reference 
	id		ID	#REQUIRED
	%language;
	edition		CDATA	""
	publdate	CDATA	""
	shortname	CDATA	""
	title		CDATA	""
	subtitle	CDATA	""
	description	CDATA	"" >

<!-- Now follow a few elements that are not allowed directly -->
<!-- under the top-level "Lexicon" element. They also do not -->
<!-- have an "id" attribute and therefore are not pointed -->
<!-- towards, they are simply contained within other -->
<!-- elements. -->

<!-- A value's content is in the NominalValue element -->
<!-- seqNr = This is used to distinguish between possible values -->
<!-- in an enumeration. -->
<!-- valueType = What sort of a value is included in the -->
<!-- content. -->
<!-- upperTolerance = The optional maximum value (only -->
<!-- applicable for reals, integers or numbers). -->
<!-- lowerTolerance = The optional maximum value (only -->
<!-- applicable for reals, integers or numbers). -->
<!ELEMENT Value		(NominalValue)>
<!ATTLIST Value		
	measureRef	IDREF	#IMPLIED
	propertyRef	IDREF	#IMPLIED
	%language;
	id		ID	#REQUIRED
	seqNr		CDATA	"0"
	valuetype 	(-|Real| Integer| Enumeration|
			String| Boolean| Number) 
				"-"
	upperTolerance	CDATA	#IMPLIED
	lowerTolerance	CDATA	#IMPLIED>

<!-- A NominalValue's content is the actual value, like "2", -->
<!-- "red", "red, blue, green, yellow". -->
<!ELEMENT NominalValue	(#PCDATA)>


<!-- A longname is the normal name for something, like --> 
<!-- "door", "fire-resistance", "kelvin" -->
<!-- language = Indicates the language of the Longname. -->
<!-- preference = A value of "1" indicates that this name is -->
<!-- preferred for this 'something' (provided that there are -->
<!-- more "Longname"s), whereas a value of "0" indicates -->
<!-- that it's best not to use this name. -->
<!ELEMENT Longname (#PCDATA)>
<!ATTLIST Longname 
	%language;
	preference	(0|1|2)	"1">

<!-- The shorthand name for something, used for properties -->
<!-- and units. So: "l" for property "length" and "K" for -->
<!-- unit "Kelvin". -->
<!-- language = Indicates the language of the Longname. -->
<!-- preference = A value of "1" indicates that this name is -->
<!-- preferred for this 'something' (provided that there are -->
<!-- more "Longname"s), whereas a value of "0" indicates -->
<!-- that it's best not to use this name. -->
<!ELEMENT Shortname (#PCDATA)>
<!ATTLIST Shortname 
	%language;
	preference	(0|1|2)	"1">

<!-- A description is a textual explanation of something. -->
<!-- language = Indicates the language of the description. -->
<!ELEMENT Description (#PCDATA)>
<!ATTLIST Description 
	%language; >

<!-- A filename is either a local filename or an internet -->
<!-- hyperlink containing more information like a VRML file -->
<!-- or a picture. -->
<!ELEMENT Filename (#PCDATA)>