On classification of XML document transformations Jana Dvo rkov - - PowerPoint PPT Presentation

On classification of XML document transformations

Jana Dvoˇ ráková FMFI UK, Bratislava jana.dvorakova@dcs.fmph.uniba.sk

DATESO 2005

DATESO 2005 – p. 1/14

Outline

Introduction Formalization of XML concepts Classification hierarchy Type transformations Formal models Conclusion

DATESO 2005 – p. 2/14

Introduction

XML - popular standard for production of structured documents

document standards, data exchange between applications

Various transformations needed Many transformation systems exist, based on different techniques

SynDoc, Scrimshaw, TREX, XSLT, Grif, Thot, VXT, CoST,...

Aims

• 1. to define classification of XML document transformations
• 2. to examine relationships among defined groups of

transformations

DATESO 2005 – p. 3/14

Formalization of XML concepts (1)

XML document

tree, where internal nodes are elements and leaves contain textual content

Document Type

context-free grammar valid XML documents - derivation trees

Restrictions

XML documents without attributes element names from finite and known set

DATESO 2005 – p. 4/14

Formalization of XML concepts (2)

Transformation

set of pairs of trees (d1, d2) d1 - tree representing source XML document d2 - tree representing target XML document

Transformation model:

tree Source Transformation process specification Transformation tree Target grammar Target grammar Source

DATESO 2005 – p. 5/14

Classification hierarchy (1)

Different possible criteria

driving element, scope, purpose, level of user interaction, transformational power, source and target correctness...

Basic groups:

Source grammar transformations Target grammar transformations Two grammar (type) transformations

In each group different transformation techniques used

DATESO 2005 – p. 6/14

Classification hierarchy (2)

Source grammar transformations transformations Event−driven Target grammar transformations applications, CoST, SAX OmniMark Metamorphosis Scrimshaw, XSLT processors, SynDoc, SDTT, ICA, HST DTT SynDoc HAG SIMON TTG DGTS, Alchemist SIPEQ, Two grammar (type) transformations element Driving Parsing transformations Tree−based SDTS ... TREX technique technique Mapping XML transformations

DATESO 2005 – p. 7/14

Type transformations

Transformation specification is created by grammar mapping Both source and target corectness Formal models used:

Syntax directed translation schema (SDTS) Tree transformation grammar (TTG) Descending tree transducer (DTT) Higher order attribute grammar (HAG)

• 1. Common framework development
• 2. Mutual comparison

DATESO 2005 – p. 8/14

Formal models - SDTS (1)

Translates a sentence of a source grammar (ws) into a sentence of a target grammar (wt) according to translation groupings Simulates derivation of both sentences from the start symbol simultaneously Sentential form = current frontiers of the source tree and the target tree Translation grouping:

A → u1 . . . un, v1 . . . vm

Translation:

(S, S) ⇒ .. ⇒ (xAy, x′Ay′) ⇒ (xu1 . . . uny, x′v1 . . . vmy′) ⇒ .. ⇒ (ws, wt)

DATESO 2005 – p. 9/14

Formal models - SDTS (2)

Standard defininition is not suitable in the case

• f XML

Extension proposed: trees as sequential forms First sequential form - source derivation tree, last sequential form - target derivation tree

DATESO 2005 – p. 10/14

Formal models - SDTS (3)

Translation step:

t β z 1 β z 1

1

u un A A

1

t

1

s

n

s m v v

1 m

Ω

1

... ...

A

m

v ...

1

v , u ...u n

DATESO 2005 – p. 11/14

Formal models - comparison

Results obtained by comparing syntax directed translation schema (SDTS), descending tree transducer (DTT) and their modifications: SDTS ESDTS d-DTT DTT SDTS

• ESDTS
• N

N

d-DTT

• N

⊆

DTT

• N

⊇

DATESO 2005 – p. 12/14

Conclusion

Summary of results:

formal definitions of XML concepts classification for XML document transformations introduced formal models for two grammar transformations defined in common framework several results of comparing these formal models

Future work:

add other comparisons of formal models include attribute transformations

DATESO 2005 – p. 13/14

Thank you for your attention!

DATESO 2005 – p. 14/14