Future of Application Development Keith Jaeger

1
Future of ApplicationDevelopmentKeith Jaeger
2
Unprecedented Change

• Huge amounts will be spent to change applications
  in the next 5 years
• The rate of change will increase the risk to
  development projects - which already fail up to
  90 of the time (IDC)
• The opportunity to make MIS a competitive
  advantage consequently grows for those who know
  how to succeed

3
Driving Forces Behind Change
Technology Driven
Cost Driven
Feature Driven
4
Mandated Business Changes
Technology Driven
Feature Driven
Cost Driven
Year 2000 European Currency
5
Two Strategies
Application
Patch
Application
Rewrite
Application
6
Budget Impact of Patching
1.00 a line of code
Technology Driven
Feature Driven
Cost Driven
Year 2000 European Currency
7
Driving Forces Behind Change
Technology Driven
Distributed Computing Components
Cost Driven
Feature Driven
8
Increased Pace of Technology Change
VS
9
Technology Risks
Smalltalk
Glockenspiel
CGI Bin
Taligent
Gupta
OpenDoc
ADcycle
NEXT
MOTIF
Token Ring
OS/2
10
(No Transcript)
11
N-Tier Client Server
NT
MVS
HP/UX
HP/UX
Windows client
NT
AS/400
AS/400
12
Dynamic Partitioning
Mobile computing, Failover, Distributed servers
NT
PgmB
Client
PgmA
HP
PgmB
PgmB
AS/400
PgmB
AS/400
PgmB
13
SMPs and Clustering
Application Task
Processor 1
Processor 4
Processor 2
Processor 3
14
Cost Pressures
Technology Driven
Feature Driven
Cost Driven
New Platforms Labor Shortages Packages
15
Labor Challenges

• Labor Shortage
• Skills Mismatch
• Shortening Tenure

Cost of skilled developers
16
Software Economics
Software becomes a commodity
Price Fixed Cost Marginal Cost
Units

• Bill Joys Law (SUN)Dont write software for
  less than 100,000 customers - 10 million 1000
  price
• Bill Gates Law (Microsoft)Dont write software
  for less than 1,000,000 customers - 10 million
  100

17
The 90s - Move to Packages
Efficiencies of Multiple Implementations
18
Groves Law - Horizontal Consolidation

• Breaks old vertical model
• Each layer picks best from lower layer
• A dominant player tends to emerge in each layer
• Different layers for different price
  points/hardware

19
Generic Layer Within Applications

• Pressure to integrate with complete solution

• New applications move to generic list

20
Dominant Players Stretch Layers

• Extend to other platforms

• Redefine layers to leverage dominance

21
Moores Law

• Hardware technologyimproves 4X every 3 years
• While someimprovementmeans morecapacity at
  thesame price,there is a downward pressure on
  prices

22
Gordon Bells Platform Economics
Software prices follow hardware prices
23
Cost Pressure
Software price follows hardware
However, costs remain constant
24
Application Growth Segments

• New Hardware Markets
• NT
• Internet
• Hand held
• Home applications
• Lower price, little differentiation
• New Vertical Niches
• High price, highly customized

25
Customization to Fill Gap
SolutionPrice
Customization
PackagePrice
26
Reuse Through O-O
Procedural Relational
Object Oriented
27
Failed to Take Hold in Applications
Procedural Relational
Object Oriented
28
New Answer Components
Procedural Relational
Object Oriented
Components
29
Components
30
Component Types

• VBX controls
• Sliders, graphs, fuel gauges, grids etc.
• ActiveX controls
• similar to VBX
• 32 bit COM controls
• Internet enabled
• Java beans
• Internet enabled
• Downloadable to browser

31
Benefits
AS/400
NT
Allows you to mix your objects with others on
other platforms at any time
32
Move to Business Components
Packages
Business Components
Controls
33
Large Scale Business Components
34
Jaegers Law
The more complex a component, the smaller the
number of opportunities for its
use Corollary The more business content in a
component, the smaller the number of
opportunities for its use
35
Gartner View of Options
Packages
Components
Templates
Cost
Custom
Flexibility
36
Next Major Leap

• Packages/components that can be customized at a
  lower cost
• More flexibility like templates
• Reuse like business components

Vendors (or in-house development) that achieve
these two objectives will establish dominance in
the next cycle
37
Template Types
38
Building a Kitchen

• Plug-ins add customization into a pre-built
  kitchen
• Frameworks customize by following rules to
  extend a kitchen blueprint
• Patterns customization draws a new blueprint

39
End User Control of Desktop
Fixed Character
Windows Desktop
Browser Navigation
Customized Display
40
End User Control of Behavior
Fixed Programs
.Ini File Control
Control ofBusiness Rules
41
Driving Forces
Technology Driven
Distributed Computing Components
Cost Driven
Feature Driven
New Platforms Packages
Year 2000 European Currency
42
Conclusions

• Massive redevelopment in next 3-5 years due to
  Y2K, EUC, Web, NT, components
• NT and Moores law will push package prices
  lower, but not necessarily solution price
• Vendors who lead the next wave will be easiest to
  customize
• New tools and development technologies will take
  on increased importance

43
(No Transcript)