|"Those who do not learn from the mistakes of the past are bound to repeat them"|
Ryan1 argues that despite new approaches and development
aides, the issues of time consuming and difficult systems development will
not disappear. He cites the wide variety of user interfaces, the continuing
need to understand roles, the lack of an underlying theory in system development
and the continual need for significant testing as the bases for his claim.
|Do you agree with Ryan's contentions? For those of you who have been in system development for a decade or more, what evidence, if any, have you seen of improvement in the system development process?|
Everybody is a Customer
|For the customer Instant Gratification isn't fast enough
Use the best, not the rest
Human Factors bigger than ever
Document, Document, Document
|Ubiquitous Internet Transport Systems
Multiple Customers Everywhere
|Stick to your business
Re-learn on a regular basis
Tight Control still needed
|Widespread Desktop Level Client Server
|Single Server Concept
Specialization within the system Development Profession
The Personal more savvy customer
LOC in hundreds of thousands
|Modularize and Capitalize
Partner w/ Customer
Hire good people
|Division level & specialized
Adoption of Packaged Software
3rd Generation Languages
LOC in tens of thousands
The Outside Customer
Limit Mix and Match Code
|Large Scale enterprise wide mainframe based
Internal Custom built applications
Low Level Coding
LOC in thousands
Lack of Documentation
The Inside Customer
Very Tight Control
This is most likely attributable to a Silicon Valley phrase - "eating your own dog food." In other words, system developers tend to employ those tools which they are producing - programs. Programmers and system analysts have a natural inclination (as Myers Briggs INTJs) to improve their development processes. And those programs or other processes are the embodiment of automated processes to shorten, and make more effective the development process. Important questions now arise. 1) Have these improvements been linear or are they increasing in nature, and 2) will this trend continue into the following generations of computing.
The Next Trend - Invisible Computing
Almost certainly the next "trend" in the computer age will be invisible computing. Invisible computing will be most readily signified by lower and lower prices for the computing machine until it becomes an "appliance." This era will concurrently be identified by the disappearance of the physical machine - as we now know it. Additionally the applications - as we now know them - will also disappear.
The Change in Hardware
We now typically associate the physical computing machine with screens, circuit cards, keyboards and mice. But within the next twenty years these artifacts will disappear. Screens will be replaced by images directed from electronic eyeglasses onto our retinas, and by large wall sized flat, yet 3D, panel displays. Computer chips will become pin-head sized micro circuits and our interface with the machine will be driven by commands from our voices, facial expressions, and hand movements. They will become more and more specialized, but able to communicate through common protocols.
The Change in Software
We also typically associate the computer with applications such as word
processing, data base processing and spreadsheets.
The reader should be mindful that the development of these applications stems from the industrial age. Before computers, written communications were transferred through pieces of paper called memoranda or via formal salutation type letters. In the information era, this paradigm was transferred to e-mail and word processing. In the industrial era, basic data were kept on file cards. In the information age, these file cards were placed into the computer in the form of databases. Also in the industrial era recorded numbers were stored on green columnar pads. In the information age these rows and columns are known as spreadsheets.
But the concept of moving our communication, storage and computation
into e-mail, databases, and spreadsheets has only been transitional. Already
the communication concept is being surpassed by intelligent agents,
storage through object and knowledge base schemas, and spreadsheets through
packaged add-ins, system dynamics and real time processing mechanisms such
as dynamic data exchange.
|From Michael Dertouzos book What Will Be, "A bank's cost to process a simple transaction like a withdrawal or a check the old way from teller to back office is $1.40; for an ATM transaction it is $0.45, and for an electronic bank-to-bank exchange. $0.08."|
1. Ryan, H.W. Information Systems Management, Winter 1993. Vol 10 Issue 1. p 62-4.
1. Full Research Papers and Compiled Organizational Analysis due.