Challenges for Future Software Development

1
The Innovative Force Behind Tomorrows Technology

• Challenges for Future Software Development
• Dr. Markus Lauff on behalf of Dr. Joachim Schaper
• Director CEC Karlsruhe, SAP Corporate Research

2
Agenda

• Who is SAP?
• What is the problem space to cover?
• Is there help?
• today
• near time future

3
What Does SAP Do?

• Delivering on the promise of e-business
  integration to help our customers foster
  growth, innovation and value ...

4
Summary of SAP Today

• SAP AG in 2001 revenues 7.34 billion
• 50,000 installations
• 18,000 companies run SAP
• Providing 21 Industry Solutions
• 29,400 SAP employees (as per June 2002)
• 12 million users in 120 countries team with us
  to
• Integrate their business processes
• Extend their competitive capabilities
• Get a better return on information at a lower
  total cost of ownership
• Focused on users in all enterprises regardless of
  size
• Increased customer satisfaction and stronger
  customer loyalty
• Heavy investment into SAPs worldwide business
  community

5
Expanding Leadership in Business Software
SAPs Main Competitors by License Revenues
The Worlds 7 Leading Software Companies By
Equity Market Capitalization
389.7
PSFT 12
ORCL 17
89,9
SAP 43
SEBL 20
42,9
ITWO 8
21,1
20,5
18,7
14,1
PSFT
SAP
SEBL
Veritas
ORCL
MSFT
CA
MSFTMicrosoft, ORCLOracle CAComputer
Associate, PSFTPeopleSoft, SEBLSiebel, ITWOI2
Source Datastream
6
SAPPHIRE 01 Defining SAPs Vision

• Integration Through Portals and Exchanges

Private and Public Exchange
Networked SoftwareComponents
Portal
7
mySAP.com The E-Business Platform

• A Total E-Business Solution
• Providing business process that span all
  functional areas within and between
  organizations
• Providing a flexible, standard framework
  supporting constant change and adaptation
• A technical framework that is open to allow the
  leveraging of existing assets and evolution into
  new areas

8
Problem space for SW Eng. in the large

• Example SAP ERP System (Sales and Distribution)
• 18,500 users
• 5,578,000 dialog steps / hour,
• average dialog response time 1.94 s
• Software - R/3 Release 4.6 C - Win2K / MS SQL
  Server 2000 SP1,
• Hardware 1 Database server 32 x Pentium III
  Xeon 700 MHz, 2 MB L2 cache, 12 GB main memory
• 92 Application 76 dialog servers 8 x
  Pentium III Xeon 700 MHz, 2 MB L2 cache, 4 GB
  main memory
• Total disk space 1,256 TB

9
SAP Software

• R3 Kernel
• 1000 developers gt 100 mio LOC
• Application Development
• 5000 developers
• Customizing...

10
How to build very large Software systems?

• Software Eng. Research hasnt found THE concept
  yet!
• Widely used Methods
• Methods and Tools for modeling and
  Specification(Huge Area informal Description,
  Diagrams, ... , formal specification)
• Regression Testing
• Code Review/walk through, ..., Documentation
  of code by non-coders
• certification (ITSEC, CC,...)
• ...
• Classical structured programming (Dijkstra). No
  GOTOs
• Controlled information exchange (well defined
  interfaces) between parts of the program. /
  Information Hiding / Modules
• Object orientation
• ...
• Lose binding of Software components with
  standardized Interfaces
• ...

11
Details Info. Exchange in Programming Languages

• No Control a program manipulates data and even
  code (Assembler) of another program, (1950)
• Communication using global variables (C,
  FORTRAN), (1960)? in most cases causing
  problems!
• Exchange of control information, z.B. global
  flags, (1960) ? high error rates!
• Data exchange using variables in procedure calls,
  (1960) ? standard method
• Messaging, (1970) ? method of choice for
  distributed systems
• modules with explicit defined (z.B.
  Export/Import) Interfaces, (1980) ? today's
  standard mandatory structuring procedure- to
  avoid uncontrolled exchange of messages or
  procedure calls!

12
How to build large Software Systems?
Requirements
Design
Implementation Testing
Integration
Operation
13
Key to Moduls reusability

• Problem
• How to find an appropriate Modul/Program
• Example. 17 GByte source tree at SAP kernel
• Blackbox vs. Whitebox How to modify an existing
  Modul?
• How to ensure that a piece of software is doing
  the intended semantic even if is written for a
  different reason

14
44 seconds of the life-time of ARIANE 501

• T - 3 sec. Main engine is on
• T 0 sec. Booster started
• T 7,5 sec. ARIANE 501 raises into the sky of
  French Guyana
• T 37 sec. altitude 3500 m, speed (857 km/h).

T 41 sec. the flight computer executes the
mechanism for self destruction
REASON a piece of reused code which has not been
modified of ARIANE 4
15
Conclusion...

• Modules encapsulate program - logic and data
• benefit easy to change
• Only reliable method if changes will be the
  default (Application Programming)
• Object-orientation extends the modularization
• classes could bee seen as modules ( inheritance,
  polymorphism,)
• Software-Development
• No one excellent method, some approaches/tools/met
  hods are in place
• Future no end to the software crisis so far
  huge amount of research but only small steps to
  improvement

16
Change Management

• Where?
• Business Logic / Backend
• Development (Rev.)
• Customization
• User Interface
• R2
• R3
• Web Interface (new Browsers)
• Mobile Devices

17
Web Dynpro Separation of Content and Layout
18
WebDynpro Simple/Advanced Clients
19
WebDynpro Separation of Presentation / Business
Logic
20
Conclusion Web Dynpro / Web AS

• State of the Art
• Separation Content / Layout
• Separation Presentation / Business Logic
• Missing
• Automatic (Business Logic) Adaptation

21
EU Project Consensus

• Research Project
• 5th Framework, 7th Call for Proposal
• Work Program Information Society Technology
• Key Action Functionality models and building
  blocks for end user services (IST-2001-IV.3.2
  )
• Time Frame Efforts
• Two years time frame, Project Start March 2002
• Total effort 340 Person-Month
• Project partners
• IBM, Nokia, Fujitsuinvia, Center of Usability
  Research (CURE), UbiCall

22
Objectives I

• Cost-efficient development of usable device
  independent Applications

23
Dilemma Cost vs. Usability
Usability
Cost
24
Adapting applications I Automatic

• A multitude of screens
• 15 numbers have to be entered
• The usage of this adapted application is too
  complicated!

25
Dilemma Cost vs. Usability
Usability
Cost
26
Adapting applications II Manual
27
Dilemma Cost vs. Usability
Usability

• Integrated Adaptation
• semantic information
• context information
• annotated application

Todays technologies
Cost
28
The Innovative Force Behind Tomorrows Technology

• Challenges for Future Software Development
• Dr. Markus Lauff on behalf of Dr. Joachim Schaper
• Director CEC Karlsruhe, SAP Corporate Research