META DATA FRAMEWORK FOR ENTERPRISE

INFORMATION SYSTEMS SPECIFICATION

Aiming to reduce or remove the development phase for EIS systems

Jon Davis, Andrew Tierney, Elizabeth Chang

School of Information Systems, Curtin University of Technology, GPO Box U1987,Perth WA 6845, Australia

Keywords: Enterprise Information System, meta-data, framework, meta-data

driven application, application generation

Abstract: This paper proposes a process for implementing a meta-data approach to defi

ning a platform independent

operational computer system application. It identifies Enterprise Information System (EIS) type systems as

ideal candidates for implementation using this meta-data, based on the simplification opportunities available

due to the typically visual and transactional component bias of EIS systems. It describes architecture for the

development of a suitable meta-data based application generator system. This development could lead to the

generation of new accelerated EIS development methodologies in business modeling, analysis, design,

system deployment and global information exchange.

1 INTRODUCTION

Common core system development methodologies

such as the Waterfall, Spiral, Fountain and V models

(Sommerville, 2000), (IEEE, 1998), (IEEE, 1997)

have not been fundamentally altered as a result of

modern technologies and in general we are still

maintaining the basic paradigm for system

development; analysis, design, develop code, test

and deploy. New system development

methodologies such as Prototyping, Agile Processes,

Big Ball of Mud (Pressman, 2000), (Martin, Raffo,

2000), (Donzelli, Iazeolla, 2001), typically propose

differing levels of task decompositions, parallelism,

customer interaction etc and certainly do provide

specific advantages when dutifully employed but

they are not guaranteed to necessarily change the

magnitude of the total effort.

An Enterprise Information System (EIS) must

ecessarily star

t with a review of its requirements

and the preparation of a design. Our project proposes

that performance of the analysis combined with an

efficient collection of this information can also

perform the bulk of the design phase, largely as a

simultaneous activity. Hence the two steps may be

merged in our proposed model.

Further, with the collectiv

e design

requirements

stored and available in a suitable meta-data format

(from the first stage), we believe that most (if not

all) EIS systems can be executed automatically with

the availability of suitable runtime components. This

expectation is based on the well structured nature of

EIS applications; highly visual and interactive

applications that prompt for the entry of appropriate

data by the users, employ strong rule based actions

and utilize database transactions to complete the

action.

The introduction of such a

n approach

has the

potential to drastically reduce the time to develop

and deploy an EIS system. Effectively, once the

analysis and design have been completed the system

would become available for immediate use! The

virtual elimination of the coding combined with the

minimisation of the testing and deployment stages

has significant benefits for both the developer and

the end users.

This project aims to develop an alternative

elopment methodology using a meta-data

standard that can be extended upon for defining and

producing Enterprise Information Systems.

2 PROGRESSIVE MODELING

METHODOLOGIES

This project will be influenced by feature subsets of

these improving and emerging technologies:

• 4

Generation Language: The most common

4GL standard is Structured Query Language

(SQL) which has become the standard in

451

Davis J., Tierney A. and Chang E. (2004).

META DATA FRAMEWORK FOR ENTERPRISE INFORMATION SYSTEMS SPECIFICATION - Aiming to reduce or remove the development phase for

EIS systems.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 451-456

DOI: 10.5220/0002626404510456

 SciTePress

database management, and will be supported as

the data repository standard for this project.

Mainstream progress in application

development was restricted to proprietary tools

although UML (OMG, 2003), (Kruchten, 2000),

(Quatrani, 2001) has emerged as a popular and

quasi-standard design toolset.

• Unified Modeling Language: UML has been

progressively developed into a precise toolset

for the specification of system requirements,

and is being provided with continually

improving third party code generation options

(Lethbridge, Laganiere, 2002), (Allegrini,

2002), (Guizzardi, Herre, Wagner, 2002). A

weakness of UML is that it has such

extraordinary breadth and requires

correspondingly high technical skills, requiring

the difficult marriage of UML experts with

business analysts and business process experts

to obtain success in the commercial world. We

believe that a less complex specification can be

achieved for the subset of issues in EIS systems

(and that can subsequently be used with far less

technical expertise) and will draw on UML

features.

• Object Modeling Group - Model Driven

Architecture: OMG/MDA (OMG, 2003),

(Tolk, 2002) provides a solid guiding

methodology based on the use of UML to

provide the platform independent model, and

following through with the efficient

development and deployment of the target

system. This project will utilize the general

philosophy of MDA combined with subsets of

modified UML.

• DARPA Agent Markup Language and

Ontology Inference Layer: DAML+OIL

(DARPA, 2003), (W3C, 2001), (Qasem, 2001),

(van Harmelen, Horrocks, 2000) provides a

foundation for the classification of elements and

as a means for automating inference in data sets

and has been inspired by the burgeoning

information base available in and via the

Internet. DAML+OIL lend useful schema

elements and provide a logical documentation

and information interchange framework.

• OWL Web Ontology Language: OWL (W3C,

2004), (Kendall, Dutra, 2002) is a web ontology

language derived from DAML+OIL which is

seeking candidate recommendation with W3C.

The EIS ontology that is ultimately finalized by

this project would sensibly seek to utilize OWL.

• Interface Definition Language: IDL (IBM,

2003) is the OMG’s language for the

development of programming language

independent object interfaces. Its relevance for

this project is as a specification source for

modeling generic objects.

These models tend to be concerned with

providing methodologies for system analysis and

design. This research aims to create an automated

solution for EIS systems development from a meta-

data framework. It utilizes a meta model driven

architecture and ontology based computing

technologies to provide the target platform

independent meta-data model of an entire EIS

specification, along with prototype platform

dependent runtime components.

3 CONCEPTUAL FRAMEWORK

FOR META

-DATA APPROACH

The success of a meta-data approach will be largely

reliant on the functionality that can be provided as

defined by higher level components with the

associated efficiencies gained by the corresponding

reduction in continual individual duplication of

components. This approach is not exclusive of the

need for access to lower level components but for

EIS systems it is the ready availability of the higher

level functions that will provide the expected major

efficiencies.

An inherent feature of EIS systems is the highly

visual and interactive nature of the applications. The

success of EIS applications is reliant on the entry of

appropriate data by the application users, and they

are heavily biased towards rules based responses and

database transactions as the appropriate action.

A common paradigm in modern code generation

that has become a virtual standard following on from

the advent of event driven programming is a

separation of the actions (implemented as events)

from an underlying framework that provides a

structure that makes logical sense for the individual

applications.

The model depicted in Figure 1 provides for a

separation of the structure and events, extending

both as managed hierarchies and allowing for flow

control between the structures. The model borrows

heavily from the use of visual components to

provide application structure - this is a commonplace

analogy used in most modern Integrated

Development Environments (IDE) for the code

based generation of applications and from other

research works in meta modeling (Motik, Maedche,

2002), (Celms, Kalnins, Lace, 2003).

ICEIS 2004 - INFORMATION SYSTEMS ANALYSIS AND SPECIFICATION

452

0..1

0..*

Sibling SE

0..1

Parent SE

0..*

Child SE

0..1

Parent Event

0..*

Child Event

0..1

0..*

Sibling Event

0..*

0..1

SE Fires Event

0..1

Event Fires SE

0..*

Structural Elements

Events

Figure 1. Structural Element - Event Molecule

Basic class diagram indicates a similar hierachical structure as described in the atomic model,

for the Structural Elements and Events, with the addition of the relationships between them

that will provide the logical basis of their mutual behaviour and interoperability.

4 META-DATA BASED

APPLICATION GENERATOR

We have proposed a prototype of a meta-data driven

system for the development of EIS systems that aids

the specification of the target application, followed

by the conversion of that specification into the

appropriate meta-data that will then be interpreted

by the runtime components. Figure 2 presents an

overview of meta-data based application

architecture.

4.1 Meta-Data Definer

The meta-data definitions effectively become the

application thus it is crucial that efficient methods of

defining the meta-data are available.

As the meta-data represents the output of the

system specification process then minimal

opportunities exist to expedite the original human

business analysis effort. However substantial

shortcuts will be offered by reductions in the system

design effort due to the amount of application

infrastructure that would necessarily be provided by

the runtime engine in support of the higher level

components that the meta-data syntax is modeled

upon. This automatic execution of platform

independent models to platform dependent execution

is a source of major savings in the development

cycle.

When starting a new application system design a

comprehensive design editor is a requirement for the

efficient specification and entry of the system

design. A custom editor is required that would

represent a suitable design paradigm which is

necessarily biased towards the production of the

target meta-data syntax.

The field of system analysis and design has

developed considerable expertise in providing CASE

toolsets to reduce the overall system development

effort, particularly the introduction of UML based

tools. Opportunities will become available to

develop utilities that import design models from

existing comprehensive third party design toolsets

and convert these designs into the corresponding

meta-data syntax.

The database is an integral component of EIS

systems which have a strong reliance on data

dictionaries and rules based database transactions.

Database schemas that do not obfuscate the original

design by abstracting the names or types of schema

elements, or by relocating database constraints or

stored procedures to a remote system tier have

significant potential as a starting point to the re-

engineering of an existing system.

The development of utilities to reverse engineer

existing database schemas and convert the schema to

the corresponding data dictionary meta-data can

accelerate the meta-data system design process

appreciably. Well formed database schemas (without

obfuscation) that seek to fully utilize the validation

options of the modern RDBMS (Relational Database

Management Systems) have the potential to reverse

engineer directly to a fully working meta-data based

application, requiring optimally minimal or no

modification to the meta-data.

META DATA FRAMEWORK FOR ENTERPRISE INFORMATION SYSTEMS SPECIFICATION - Aiming to reduce or

remove the development phase for EIS systems

453

ICEIS 2004 - INFORMATION SYSTEMS ANALYSIS AND SPECIFICATION

454

4.2 Runtime Meta-Data Updater

The magnitude of the effort expended on future

system maintenance for systems represents by far

the majority effort for the lifecycle of systems from

the developer’s perspective.

Contributing causes to the ongoing high

maintenance costs continue to be the:

• lack of consistent and available system

documentation,

• inconsistently applied standards during different

(re)development phases,

• lack of structured programming techniques,

• extension of the system to provide features that

would have been better served by a full or

partial redevelopment,

• natural attrition of software development team

members knowledgeable of the system

architecture,

• natural progression of the underlying

technology to newer, richer and better supported

platforms.

The indicated high level maintenance figures of

80% and higher refers to the costs for the developers

of the system. For EIS systems this represents

person years of effort for the developer. The use of

meta-data based applications will drastically reduce

these costs to the developer and user base.

For the developer, meta-data applications will

largely be self documenting (except for the runtime

engine) which reduces the risk and reliance on

individual developers and development management

practices.

For the customers, meta-data applications are

provided as streams of structured meta-data. Unless

provided with an updated runtime engine (an

occasional requirement), the same runtime engine as

deployed to all users at a site would remain

unchanged (although this is a relatively simple

change management scenario to resolve). Changes to

a meta-data application are simply a new stream of

meta-data that represent

only the specific application

changes. To progress to any later version of a meta-

data application requires only the application of the

correct sequence of progressive update version

meta-data streams.

4.3 Application Meta-Data

The ultimate aims of a meta-data model for EIS

systems definition include abstraction from the

physical constraints of any runtime components that

are used for the final implementation and execution

by the users.

Utilising a meta-data based interpretation of the

application specification allows applications to be

executed using any simultaneous combination of

platforms that are supported by the components of

the runtime engine, providing a progression towards

complete platform independence.

The visual structure meta-data is used to

construct the appearance of the application as

presented by the user interface runtime components

to the users. Visual structure meta-data is analogous

to the drag’n’drop forms functionality provided by

modern GUI based IDE software and will draw on

advances such as XForms (W3C, 2004), (Grimes,

2002).

The program flow meta-data is used to define the

user interface and local platform actions and

procedures that are executed in response to user

actions and other data changes. Program flow meta-

data is analogous to the event processing and general

programming functionality provided by modern GUI

based IDE and traditional programming languages.

The data dictionary meta-data is used to define

the requirements of the database schema and the

data changes required in response to user actions and

other data changes. Data dictionary meta-data is

analogous to the data dictionary role provided by

modern RDBMS systems.

4.4 Runtime Processor

The application meta-data must be interpreted and

executed for the users. This role is provided by

additional meta-data to map the generic application

meta-data to the interface requirements of the

platform specific components for execution, plus the

platform specific runtime engines that will perform

the execution.

5 RECAPITULATION AND

FUTURE WORKS

In this paper we proposed a development and

deployment of Enterprise Information Systems that

uses meta-data application definitions within a

framework of the OMG/MDA and the prototype of

the meta-data application generator that we believe

will reduce the complexity and costs at each phase

of the software life cycle. We also propose to

develop a meta-data standard for the definition of

EIS type systems that will be capable of extension in

the future to model a wider spectrum of application

types. We presented an overview of the essential

components of meta-data application generator

architecture.

META DATA FRAMEWORK FOR ENTERPRISE INFORMATION SYSTEMS SPECIFICATION - Aiming to reduce or

remove the development phase for EIS systems

455

Currently the prototype system incorporates

components of meta-data application architecture

including:

• a meta-data syntax for the structure of EIS

applications and the visual presentation of the

user interface,

• a meta-data syntax for the execution of

distributed events,

• a meta-data syntax for the data schemas of EIS

applications and conversion between third party

database schema types,

• runtime processing components suitable for

executing the EIS meta-data on a selected

platform subset.

We also note this perspective of the EIS software

life cycle could lead to many new concerns in the

next generation of software engineering:

• New cross platform components that can

broaden the accessibility of meta-data

applications,

• Extended meta-data definitions and

functionality for richer application types,

• Merging of meta-data applications to create

broader and deeper EIS applications,

• Development of more accessible design

toolsets.

REFERENCES

Sommerville, I. (2000). Software Engineering (6th

Edition)

, Addison-Wesley Pub Co.

IEEE (1998). IEEE Standards Collection Software

Engineering 1993, 1998 Ed, IEEE Inc.

IEEE (1997). IEEE Guide for Developing Software Life

Cycle Processes, IEEE Inc.

Pressman, R. (2001). Software Engineering: A

Practitioner's Approach, 5

Ed, McGraw-Hill Inc.

Martin, Robert H., Raffo, David (2000). A model of the

software development process using both continuous

and discrete models in Proceedings of

Software

Process: Improvement and Practice

, Volume 5,

Number 2-3, June-September 2000.

Donzelli, Paolo, Iazeolla, Giuseppe (2001). A hybrid

software process simulation model in Proceedings of

Software Process: Improvement and Practice, Volume

6, Number 2, June 2001.

OMG (2003). Introduction to OMG’s Unified Modeling

Language. Retrieved October, 10 2003 from

http://www.omg.org/gettingstarted/what_is_uml.htm.

Kruchten, Philippe (2000). The Rational Unified Process:

An Introduction, Addison-Wesley Pub Co.

Quatrani, Terry (2001). Introduction to the Unified

Modeling Language, Rational User Conference, 2001.

Lethbridge,

Timothy, Laganiere, Robert (2002). Object-

Oriented Software Engineering: Practical Software

Development using UML and Java, McGraw-Hill.

Allegrini, Tiziana (2002). Code Generation Starting From

Statecharts Specified in UML. Universita Degli Studi

di Pisa.

Guizzardi, Giancarlo, Herre, Heinrich, Wagner, Gerd

(2002). Towards Ontological Foundations for UML

Conceptual Models in Proceedings of 1

International

Conference on Ontologies, Databases and

Applications of Semantics, 2002.

OMG (2003). OMG Model Driven Architecture – The

Architecture of Choice for a Changing World.

Retrieved October, 13 2003 from

http://www.omg.org/mda/executive_overview.htm.

Tolk, Andreas (2002). Avoiding Another Green Elephant –

A Proposal for the Next Generation HLA Based on the

Model Driven Architecture, 2002 Fall Simulation

Interoperability Workshop.

DARPA (2003). Why Use DAML ? Retrieved September

12, 2003 from

http://www.daml.org/2002/04/why.html.

W3C (2001). DAML+OIL (March 2001) Reference

Description. Retrieved September, 17 2003 from

http://www.w3.org/TR/daml+oil-reference.

Qasem, Abir (2001). A Prototype DAML+OIL Ontology

IDE, International Semantic Web Working

Symposium, Stanford, 2001.

van Harmelen, Frank, Horrocks, Ian (2000). Questions

and Answers on OIL: the Ontology Inference Layer

for the Semantic Web. IEEE Intelligent Systems,

Volume 15, Number 6, December 2000.

W3C (2004). OWL Web Ontology Language Reference.

Retrieved February, 10 2004 from

http://www.w3.org/TR/2004/REC-owl-ref-20040210/.

Kendall, Elisa F., Dutra, Mark E. (2002). An Introduction

and UML Profile for the Web Ontology Language

(OWL), Sandpiper Software Inc, 2002.

IBM (2003). Interface Definition Language (IDL).

Retrieved September, 25 2003 from http://www-

106.ibm.com/developerworks/websphere/WASInfoCe

nter/infocenter/wasee_content/corbaio/ref/rcidl.htm.

Motik, Boris, Maedche, Alexander, Volz, Raphael (2002).

A Conceptual Modeling Approach for Semantics-

Driven Enterprise Applications in Proceedings of 1

International Conference on Ontologies, Databases

and Applications of Semantics, 2002.

Celms, Edgars, Kalnins, Audris, Lace, Lelde (2003).

Diagram Definition Facilities Based on Meta-Model

Mappings, The Third OOPSLA Workshop on Domain

Specific Modeling, 2003.

W3C (2004). Xforms – The Next Generation of Web

Forms. Retrieved February, 16 2004 from

http://www.w3.org/MarkUp/Forms/.

Grimes, Richard (2002). Developing Applications with

Visual Studio .NET, Addison-Wesley Pub Co.

ICEIS 2004 - INFORMATION SYSTEMS ANALYSIS AND SPECIFICATION

456