Delivering flexible applications using datamanagement patterns

1
Delivering flexible applications using
data-management patterns

• Willem de Vries, Remia C.V.
• Simon Jasperse, Kiboko

2
Abstract

• Remia uses CoolPlex for software development,
  for its ability to design abstract software
  patterns.
• A set of patterns is presented that enables
  end-users to configure parts of an application.
• The architecture of the Remia pattern-library
  will be presented from the application /
  development perspective.

3

• Willem de Vries, Remia C.V.
• IT Management and alignment
• 15 years IT experience
• Responsible for Application Architecture
• User-role for pattern requirements

4
Agenda

• Remia, the company
• Remia and Plex
• Rolodex
• Remia Patterns
• Application
• Design
• Plex
• Observations
• ? time

5
Remia, The Company

• 75 years of food manufacturing
• 300 employees
• Flexibility as a strategic approach
• IT-systems should support
• rapid change
• diversified processes
• IT-policy with strong development component

6
Remia and Plex

• Long term AS/400 - iSeries experience
• Using 2E since 1991
• Plex since 1999 for
• quality through standards enforcement
• speed through reuse of functional patterns
• Pattern development
• Framework-approach vs.
• Explicit experience

7
Rolodex the application

• Centralised customer-info repository
• Flexibility structural perspective
• different channels, different info
• Flexibility process perspective
• short time to change
• independent from IT department
• Requirement user perspective
• easy and comprehensive

8
Rolodex The Project

• Base Layer
• Separation of Logical and Physical
• Introduction of Replace - entity
• Separation of UI and DB models
• First phase of application project
• Redesign of Base Layer
• Build of additional patterns in sync with
  Application Development

9
Rolodex Categories of Patterns

• Technical patterns
• Synchronisation / consolidation of data
• History
• Supporting patterns
• Polymorphism
• Functional patterns data-flexibility
• User Managed Data
• Query
• XML

10
User Managed Data

• User perspective
• Definition of questions
• Conditional application of question-groups
• Developer perspective
• Creation of answer-formats
• Application within owner-context
• Practice
• Used within 4 different contexts

11
Comparison to _UserData

• Goals are comparable
• Difficult to adjust _UserData to Remias
  BasePattern
• UI of _UserData too constrained
• Limited run-time flexibility in _UserData
• Didnt like the licensing model..

12
Query

• User perspective
• Conceptual vs. Table/Column view
• Tree-wise presentation of compound queries
• Developer perspective
• Creation of answer-context
• Creation of support for elementary concepts
  Datatype Operators
• Technical ODBC AS/400 support

13
Comparison to Filter

• Filter can be implemented fast
• Drawbacks are
• attached to grid
• based on single view
• restricted operator-support for AS/400
  implementation

14
XML

• User perspective
• Flexible presentations layout
• Selective presentation
• Developer perspective
• New schemata
• Schema-extensions
• Practice
• XML-based presentation combines with configurable
  data support

15

• Simon Jasperse, Kiboko
• 10 years Agronomist UN Africa
• 6 years Obsydian gt Advantage Plex
• Application Development
• Pattern Design / Development
• Free-lance consultant

16
Patterns User Managed Data

• Application Demonstration
• Enter relation data
• Specify data for relation type

17
Patterns User Managed Data

• Design

Relation
Relation Type
DataOwner
Characteristic
Delivery Conditions
Product Sales
Use in kg/week (num) Type of Product (list)
Competitor (list)
Data (format)
ListData
18
Patterns User Managed Data

• Database Implementation

19
Patterns User Managed Data

• Interface Implementation, Data Entry

Data Value
Characteristics
Data Entry, Format
20
Patterns User Query

• Create and run query
• Type A users in Java land, attending Edge 2002

21
Patterns User Query

• Query example
• Type A users in Java land, attending EMEA
• Type A AND Java AND attending EMEA

22
Patterns User Query

• User Query Pattern
• ERD
• Combine queries (AND /OR)
• UI to build combinations
• Query by Platform
• AS400 - file pointer
• ODBC - SQL

23
Patterns User Query

• Polymorphism Pattern
• Super - Subtype
• Subtype
• owned by supertype (query)
• shared functionality (run query)
• distinct functionality (query view / fields)

24
Patterns Polymorphism
Call RunQuery
Dispatcher
Super Type
Dispatchers
Runquery
Common Functionality
View1
View7
25
Patterns User Query

• ERD

Relation
Query_ Owner
Query
26
Patterns User Query

• Implementation

Call RunQuery
27
Patterns XML/XSL

• Application Demo

28
Patterns XML/XSL

• XML generation
• DOM for xml generation
• Scripting for DOM control
• Conform with Plex conventions

29
Patterns XML/XSL

• XML Writer
• Abstract functions
• Write xml data from block/single fetch
• XML generation options
• Keys in node Attributes
• Field Names for nodes
• Impl Names for nodes

30
Patterns XML/XSL

• XML Writer Tree structure through object scoping
• Function name gt Node Name
• Meta calls and parm mapping
• Field names for field tags

MijnAgenda.xml
XML Generator function
31
Patterns XML/XSL

• XML Writer Implementation
• Inherit from abstract functions
• Replace View / Fetch functions

32
Patterns XML/XSL

• Display XML - XSL, ActiveX

33
Patterns XML/XSL

• Display XML - Runtime script generation to
  combine XML and XSL

lthtmlgt ltbodygt ltscript languageVBScriptgt set xml
CreateObject("Microsoft.XMLDOM") xml.async
false xml.load("klnkaart.xml") set xsl
CreateObject("Microsoft.XMLDOM") xsl.async
false xsl.load("Basis.xsl") document.write(xml.tra
nsformNode(xsl)) lt/scriptgt lt/bodygt lt/htmlgt
34
Patterns XML/XSL
Relatie.xml

• XML for data exchange between platforms
• Data transfer
• 1 table 1 file
• Keys gt Attributes

35
Patterns XML/XSL

• XML for data exchange
• Process Data Communication Files

DownloadInfo.xml
36
Observations

• Flexibility and patterns
• Balancing between developer and user
• Weighing reusability and initial investment
• Application Building
• Poor impact analysis
• Huge number of objects
• Practical results
• Fast implementation when needed
• Small team delivers

37
Discussion Questions
deVries_at_remia.nl simon_at_kiboko.nl www.kiboko.nl