Content

• Generic model
• Representations
• Stream
• Tree
• Events
• Transformation
• Measurements
• Basic Transformation processes
1. Stream -> Event
2. Stream -> Tree
3. Event -> Tree
4. Event -> Stream
5. Tree -> Event
6. Tree -> Stream
7. Stream -> Stream
8. Tree -> Tree
9. Event -> Event
• Testing

[representation] -> [transformation process] -> [representation]

The representations and the transformation processes are characterized by a number of properties and each performance measurement is valid with respect to a specific set of properties.

A baseline measurement is a measurement taken with respect to a specific set properties which other measurement are related to.  Usually the baseline set of properties are selected in order to create, what is believed to be one of the fastest possible measurement.
Baselines are used to minimize the effect of machine configurations and create a relative measurement which easier could be used to compare measurement from different machines.
 

Representations

NOTE:  Properties are name-value pairs. Property names are shown  in  BOLD and its possible values are enclosed on quotes.
Boolean properties are however shown with their name in quotes with implicit true/false as possible values.

Stream

Properties:

• Name: name of stream representation.
• Identifier: identifer of stream representation. Filename, URI, etc.
• Encoding: "UTF8", "UTF 16", "ISO-8859-1", ...

Type of character encoding used
• StreamType:

"File"
The stream is stored in a operating system file. Information in the stream is accessed through I/O which involves a performance degradation relating to machine configurations but also how the algorithm that accesses the stream is constructed.
"ByteArray"
The stream is stored in memory, in an array of bytes. No I/O necessary to access stream information.
• "W3C XML 1.0"

The stream syntax is valid according to W3C XML 1.0 specification.
• "W3C XML Namespace"

The stream syntax is valid according to W3C XML Namespace specification.
• Size: Stream size expressed as bytes

Tree

Properties:

• Name: name of tree representation.

"W3C DOM"
The tree representation is according to W3C DOM specification.
• Threading:

"S" - Single reader and writer
"M" - Multiple readers and writers
• "defer-node-expansion"

If this feature is set to true, the tree nodes in the returned document are expanded as the tree is traversed. This feature allows the parser to return a document faster than if the tree is fully expanded during parsing and improves memory usage when the whole tree is not traversed.
• "Include ignorable whitespace"

Includes text nodes that can be considered "ignorable whitespace" in the tree.
• "create-entity-ref-nodes"

Create EntityReference nodes in the tree. The EntityReference nodes and their child nodes will be read-only.
• Size: In.memory size expressed as bytes

Events

Properties:

• Name: "SAX2", "SAX"

Name of representation.
• "namespace-prefixes"

Report the original prefixed names and attributes used for XML Namespace declarations.
• "java-string-interning"

All element names, prefixes, attribute names, XML Namespace URIs, and local names are internalized using java.lang.String.intern.

Properties:

• Iterations: The number of iterations the transformation was performed.
• "validation"

The transformation process includes an activity which validate the input representation against a schema.
• ValidationSchema: "XML-DTD-x.x", "XML-Schema-x.x", "TREX-x.x", "RELAX-x.x",...
• "external-general-entities"

Include external general entities
• "external-parameter-entities"

Include external parameter entities or the external DTD subset.
• "W3C XML Namespace handling"

Perform XML namespace processing: prefixes will be stripped off element and attribute names and replaced with the corresponding namespace URIs. By default, the two will simply be concatenated, but the namespace-sep core property allows the application to specify a delimiter string for separating the URI part and the local part.

Duration

CodeSize

size

• stream : size of stream
• tree : size of in-memory representation
• event: not applicable

Memory
The memory size measurement is measured as the difference between the size of runtime memory before and after the transformation.

General rules:

• Java
• forced garbage-collection SHOULD be performed before each transformation
• Streams
• the activity of flushing intermediate buffers MUST be included in any time measurements.

NOTE: Absence of a baseline property name means that the value is either empty or false

1. Stream -> Event

Baseline properties:

• Encoding: UTF8
• StreamType: ByteArray
• W3C XML 1.0
• W3C XML namespace

 
• SAX2

 
• NoValidation

2. Stream -> Tree

Baseline properties:

• Encoding: UTF8
• StreamType: ByteArray
• W3C XML 1.0
• W3C XML namespace

 
• W3C DOM

 
• NoValidation

3. Event -> Tree

Baseline properties:

• SAX2

 
• W3C DOM

 
• NoValidation

4. Event -> Stream

Baseline properties:

• SAX2

 
• Encoding: UTF8
• StreamType: ByteArray
• W3C XML 1.0
• W3C XML namespace

 
• NoValidation

5. Tree -> Event

Baseline properties:

• W3C DOM

 
• SAX2

 
• NoValidation

6. Tree -> Stream

Baseline properties:

• W3C DOM

 
• Encoding: UTF8
• StreamType: ByteArray
• W3C XML 1.0
• W3C XML namespace

 
• NoValidation

7. Stream -> Stream

Baseline properties:

• Encoding: UTF8
• StreamType: ByteArray
• W3C XML 1.0
• W3C XML namespace

 
• NoValidation

8. Tree -> Tree

Baseline properties:

• W3C DOM

 
• NoValidation

9. Event -> Event

Baseline properties:

• SAX2

 
• NoValidation

TODO: create testcode/ testcases based on unit testing and the code package JUnit.
.