An Engineering Perspective on Structural Computing: Developing ComponentBased Open Hypermedia System

1
An Engineering Perspective on Structural
Computing Developing Component-Based Open
Hypermedia Systems
Michail Vaitis, Manolis Tzagarakis and George
Gkotsis
2
Structural Computing Motivation

• Inadequacy of information structuring
  abstractions
• Web applications
• Links are second class entities
• Starting-points for unidirectional jumps
• Hypermedia applications
• Employed for incarnating all information
  structuring situations

3
Research on Hypertext

• Hypermedia domain research
• how the human mind perceives structure in
  different problem domains
• Hypermedia system research
• designing and building the computational
  foundations to support people working with
  structure

4
Hypermedia domains

• Hypermedia domain
• a coherent set of structure abstractions that
  solves a particular data organization problem

• Navigational
• Taxonomic
• Spatial
• Argumentation support

• Configuration management
• Workflow
• Hyperfiction
• Linguistics

5
Component-Based Open Hypermedia Systems
Third-party programs requesting structure
functionality
Provide domain-specific structure services to
clients
Fundamental functionality available to all
structure servers (persistent storage, naming,
event notification control, versioning)
6
Architecture of CB-OHS
7
Structure Server

• Domain specifications specifications about the
  structure abstractions of the domain
• Structure syntax
• Templates / Patterns
• Behavior semantics
• Internal operations
• External operations
• Services Implement the domain-specific interface

8
Structural Engineering

• Motivation the lack of an engineering framework
  guiding the development process, coupled by the
  appropriate tools
• Definition a systematic and disciplined approach
  to the development, operation and maintenance of
  applications and infrastructures that solve
  structure-oriented problems

9
Characteristics of structure servers

• Structural completeness (structure abstractions
  should completely solve the structure-oriented
  problems of the domain)
• Size (small to medium software projects)
• Performance (short response time to clients
  request)
• Distribution and Heterogeneity (operate in a
  distributed environment of different hardware and
  software platforms)
• Specifications evolution (small possibility for
  the specifications to be changed during the
  development)
• Reusability and Extensibility (structure services
  at a fine granularity level)
• Life time (long duration of structure services)
• Robustness, Scalability and Availability (due to
  a huge number of applications)
• Introspection capabilities (communicate behavior
  to other applications)
• Interoperability (transformation of structures
  between different hypermedia domain)

10
CB-OHS life cycle
11
Scenario description

• Goals (name of each goal, plus a description of
  it)
• Characters (the different kind of users of the
  service)
• Data (data examples that may be associated
  together)
• Requirements for third party applications
  (requests)
• Structure configuration (description and
  constraints among the structural elements)
• Behavior description (operations and propagation
  of them, synchronization among elements)
• Infrastructure requirements (storage, naming,
  etc. )

• Facilitate discussions among hypermedia
  application designers and developers
• Avoid reinventing the wheel.

12
Structure specifications

• Syntax
• Specification of the structure model of the
  domain
• Contain the basic structural elements, their
  properties and the connection constraints among
  them, or to data items
• Detection of relevant elements that maybe reused
• Structure is a first class entity
• Behavior
• Computational aspects of a domain
• Tightly coupled with the structure model

Structure syntax and behavior may be considered
as different views of the same whole,
introducing the possibility of specifying both
within the same language
13
Component implementation

• Casting of the neutral structural objects stored
  by the infrastructure, to the specific structure
  elements concerning the application domain.
• Abstraction Factory Layer (AFL). Responsible for
  reifying un-typed structural objects to domain
  specific abstractions
• Abstraction Utilization Layer (AUL). The domain
  specific abstractions may be used by clients

14
Deployment and Documentation

• Structure server in operational mode
• Clients ability to discover and locate the
  required service
• Registration of the structure server to the
  dedicated directory services
• Configuration of introspection capabilities

15
Structural Engineering Tools

• Theoretical armory tools
• Aim at supporting problem analysis within the
  structural computing framework
• Development tools
• Attempt to assist developers when working with
  CB-OHS

16
Theoretical tools and armory

• Needed primarily by analysts and to a lesser
  degree by developers
• Required on two important research fronts
• Structural completeness To what degree structure
  specifications may cover or solve structural
  problems?
• Structural analysis and decidability
• Comparison of models and determination of their
  differences
• Decision whether a particular organizational
  problem belongs to a hypermedia domain, or not

17
Development tools

• Tools for structure model definition
• Structure specification formalism
• Open to extensions,
• Model-neutral
• Common ground for cooperation
• Tools for structure behavior specification
• Modelling dynamic and computational aspects
• Controlling structure abstractions life-span,
  interaction and reaction to messages

18
Development tools (contd)

• Tools for discovering components in CB-OHS
• Aim to locate available structure servers without
  prior knowledge of their existence or their name
• Special protocols for browsing structure servers
  according to certain properties, such as
• the name of the domain
• the interface the structure server is using to
  receive requests
• the availability of the server

19
Structural computing environments

• Themis
• Construct
• Callimachus

20
Future Work

• Tools for structural computing
• Behavioral modeling
• Stimulus absorption
• Propagation of operators
• Theoretical armory
• Structural completeness
• Structural decidability
• Peer-to-peer Hypermedia
• Structure services discovery