Post on 10-Jan-2016
description
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 1
Thomas TriebseesUniversität der Bundeswehr München
Department of Computer Science
Thomas.Triebsees@unibw.de
Tucson, 27th March 2007
A Theory for Model-based TransformationApplied to Computer-SupportedPreservation in Digital Archives
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 2
Agenda
I. Research Context
II. Motivating Example
III. Declarative, Preservation-Centric Approach for Model Transformation
IV. Preservation Language
V. Results / Conclusions / Remarks
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 3
Research Context
Long-term preservation of digital material
Two facets:
o Preservation procedure
o Technical environment
Preservation approaches:
o Emulation
o Migration
o Hybrid
Technical environment
o usually highly modularized and complex
} preserve information
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 4
Motivating Example
PStorage 1 PStorage 2 PStorage m
...Server 1
...
Archiver Browser
Server 2 Server n
UICIINGEST REQUEST /RESPONSE
EXPOSE BII
ArchiveCustomer User
Permanent Storage Web Storage
storePerm
extractMetadata store
respond
TransformationTransformation
Contribution: Declarative, domain-specific, preservation-centric language that facilitates to specify preservation requirements
Integratable into system specifications
Appropriate language elements
Abstraction from differentimplementations
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 5
III. Declarative, Preservation-Centric Approach for Model Transformation
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 6
Approach – System Model and Transformation Semantics
-Typed Entities
-Associations
-Functions
Model elements
Archive State AArchive State ASystem State A Archive State A'Archive State A'System State A'
basic operation(create object)
System State A''
basic operation(set association)
Transformation algorithm = sequence of basic operationsTransformation algorithm = sequence of basic operations
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 7
Approach - Preservation
Preservation Preservation languagelanguage
ConstraintsConstraints
Transform.Transform.algorithmsalgorithms
defines
used togenerate /implement
respect
Concept (= semantic property)Concept (= semantic property)
Context 1Context 1 Context nContext n…
source context target context
implements implements
…specify pre-servation of
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 8
Example – Abstraction through Contexts and Concepts
Website
HTMLFileFolder
File
PDFFile
Tag
eP
0..1
11
1
*
1
*
*
0..1
<name>
<name>.html
<name>
html
resources
<name>.html
Concept Website
……
…
……
…
……
…
src
trg
Context AWebContext BWeb
φ (FOPL)
ψ (FOPL)
AWeb
BWeb
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 9
Example – Specifying the Preservation Task
source
calculation2005
overview
calc.pdf
doclist.html
start.html
Website
eP
0..1
<name>
<name>.html
<name>
html
resources
<name>.html
……
…
……
…
……
…
AWeb
??
HTMLFileFolder
11
… …
δ
Concept Website
BWeb
Calculation
EXPOSE
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 10
Example - Typed Entities
name : Stringattrs : Seq<String x String>content : Seq<(String | Tag>
name : Stringattrs : Seq<String x String>content : Seq<(String | Tag>
Tag
<html><head/>
<body> … <a href=“…“></a> …</body></html>
<html><head/>
<body> … <a href=“…“></a> …</body></html>
start.html name : Stringlocation : Stringcontent : Tag
name : Stringlocation : Stringcontent : Tag
HTMLFile start.htmlstart.html
h:HTMLh:HTML
<html><html>
t11:Tagt11:Tag
<head><head>
t111:Tagt111:Tag<body><body>
t112:Tagt112:Tag
<title><title>
t1111:Tagt1111:Tag
<a><a>
t1121:Tagt1121:Tag
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 11
IV. Preservation Language
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 12
Preservation Language – Transformation Constraints
h → HTMLFile
name = "start.html"location = "/… /source"content = <...>
name = "start.html"location = "/… /source"content = <...>
h:HTMLFile
name = ?location = ?content = ?
name = ?location = ?content = ?
δ(h):HTMLFileδ
trans(δ | h → h‘)trans(δ | h → h‘)
Archive State A'Archive State A'System State A Archive State A'Archive State A'System State A'
h h‘
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 13
HTMLFile-{ }
name
location,content
name = "start.html"location = "/… /source"content = <...>
name = "start.html"location = "/… /source"content = <...>
h:HTMLFile
Preservation Language – Object Preservation Constraints
name = "start.html"location = ?content = <...>
name = "start.html"location = ?content = <...>
δ(h):HTMLFile
δ
name
trans(δ | h → h‘)trans(δ | h → h‘)
h → HTMLFile
post(δ) ≡ { name(h‘) = name(h) }
presO(h → HTMLFile, h[HTMLFile-{ }])location,content
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 14
Preservation Language – Concept Preservation
Constraints
w: „Calculation“
f: „source“
Context AWeb
δ
presK({w → Website}, Website(w,f,h), AWeb, BWeb)
Concept Website
h: „start.html“
eP
δ(w): ____________
δ(f): ________ δ(h): __________
eP
Context BWeb
f‘: „html“ f‘‘: „resources“
presO(h → HTMLFile, h[HTMLFile-{ }])location, content
„start.html“„source“
„Calculation“
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 15
V. Results / Conclusion / Remarks
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 16
Conclusion
Constraint definitions show generality (abstraction
from specification language)
Prototypical implementation available
Website migration example:o ≈ 300 files
o ≈ 20 concepts
o ≈ 700 constraints
o ≈ 2300 affected objects
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 17
Conclusion
Strengths: State-based approach
Powerful object model
Abstraction via concepts; specification language need not necessarily
be FOPL
Intuitive constraint definition for application domain
Coherently formal underpinning
Limits: Generation of parallel migration algorithms
Undecidability of FOPL
Generating algorithms for comprehensive model reconstructions
Efficiency
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 18
Subject to your questions…
Thomas TriebseesUniversität der Bundeswehr München
Department of Computer Science
Thomas.Triebsees@unibw.de
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 19
Constraint-based Migration
easily integratable into system specifications
focus on critical issues
abstraction from implementation, thus platform independent
formal treatment possible
Advantages of declarative approach
Thomas Triebsees, Uwe M. Borghoff, Dptmt. of Computer Science 20
Approach
Conceptual overview
(1) Pre-processing
object model
source modelinstance
(4) Algorithm execution
target objectmodel
(2) Concept recognition
concept def.
extendedobject model
(3) Algorithm generation
constraints
migration alg.
(5) Post-processing
target modelinstance