The Collapse and Rebirth of Information Systems … · project manager repeatedly reminds team that the one thing that's absolutely beyond consideration is any further slippage; the

The Collapse and Rebirth of Information Systems

InfrastructureTechnology Management Issues

for 2001-2010 Conrad WeisertJanuary, 2001 forACM Chicago Chapter

copyright 2000, Information Disciplines, Inc. 1

Copyright 1997, Information Disciplines, Inc., Chicago

1998 Hypotheses

Organizations are abandoning any systematic approach to project planning & control.

Most other aspects of I.T. infrastructure are similarly being abandoned.

The once-discredited "theory-X" school of management is staging a strong comeback. Can these be valid?

What are the causes?Is there a remedy?



Dawning awareness:3 situations in the late 1990s

Brief consulting roles on three huge death march projects1. an integrated insurance application2. a telecom company software package

installation project3. an employee benefits system for an

international telecom firm.

Typical or isolated situations?copyright 2000, Information Disciplines, Inc. 3


Dawning awarenessSituation 1 (1995 - 1998): A huge casualty insurance company plans to replace nearly all of its applications with an integrated system.

A major software services vendor gets a contract to develop the new system using:

object-oriented technology> 100 programmers (in-house, vendor, & freelance)

First deadline established (early 1997)

Y2K ignored as irrelevant

First deadline missed.



Dawning awareness:situation 1 (continued)

More people added to the project (>200)Second deadline established (mid 1997)Second deadline missedContractor fired3 key contractor people (O.O. experts)

leave the contractorform a new companyget the contract!

Third deadline established (January, 1998)Client insurance company declares any further slippage absolutely unacceptable




Project status October, 1997400+ use-case scenariosno other specs. or requirements documentationno formal project planno user documentationdaily (!) team meetings of ~24 key people

project manager calls on each in turn to report progress since yesterdayproject manager repeatedly reminds team that the one thing that's absolutely beyond consideration is any further slippage; the system will be operational by end of January!

Guess what happened?



Dawning awarenessSituation 2: (1998-1999)Fast-growing telecommunications company needs to automate screening job applicants' résumésThey find a packaged software product that scans résumés submitted by E-mail, matching them to open job requisitions on the basis of combinations of keywords and storing them in a database for future retrieval.They commit to replacing their current résumé processing by the new system on a geographically phased implementation schedule starting in 10 months.




A project manager is appointedA young "fast track" executive.He has never managed an I.S. project before.He forms a project team with members located in 11 different cities (none of which is where he is) across three countriesHe calls a "kickoff" meeting, at which he announces that the deadlines (February, 1999) are non negotiable.




Project status (November, 1999)No formal project planThe weekly scheduled conference call

The whole project team and the software vendorReview and update the "action list"

Starting (reluctantly) to document "requirements"

Discovering numerous "discrepancies" where the product fails to satisfy the requirements.




Dawning awarenessSituation 3 (1999-2000)Fast growing international company, after several mergers and acquisitions, needs to overhaul its employee benefits.Each employee is to be given a menu of options for such things as:

stock purchasesavingspensionhealth care

We need an online system for each employee to use (once) to make his or her selections




Employees' selections are constrained by many complicated rules and U.S. & Canadian government regulations.Several of the benefits plans are outsourced to banks, insurance companies, etc.Governments mandate that the process must be complete by end of February, 2000Project manager appointed (January, 1999)

Calls a meeting of all affected key people to define system requirements, September (!), 1999




Conclusions?Were those isolated instances of unusually weak management, or did they illustrate an increasingly common pattern?

Causes and cures?



Growing awareness

It's not just project management; it's all I.S. infrastructure!

But didn't organizations solve most of these problems years ago? New generation of decision makers Extreme short-term rewards system: get it done now and move on.Theory-X resurrectionpoor organizational memory



What is (was) I.S. Infrastructure?

Support servicesLibrarian services

Reusable software component repositoryDocumentation repositoryPublications library

Education services

and lots more!

MethodologySystem development life cycleStandards and tools for

Project planning & controlSystems analysisProgramming Database design



Management by cliché

"I agree with you in principle, but this deadline is so critical that we just don't have the luxury of doing it right."

"We probably should have gotten you guys in a year ago, but now it really istoo late."

"We've just got to roll up our sleeves and do whatever it takes to get the job done!"



Return of Theory-XmanagementAssumes:

People are lazy; they dislike and shun work.

Employees can't be trusted to carry out their responsibilities without detailed supervision and frequent prodding.

Arbitrary "deadlines" are needed to get people to work harder.

Douglas McGregor: The Human Side of Enterprise, 1960, McGraw Hill.



Poor organizational memory

An organization may have documented and disseminated its best practices

But after a merger, decentralization, or other reorganization:New management views the documentation as:

Bureaucratic overheadAssociated with the hated old regime.

"We're going to be leaner and meaner".Deliberately throw it out or just let it wither away.



Professional staff development (a.k.a. "training")

Large I.S. organizations typically allocated between 3% and 6% of staff-member's time to learning new concepts, skills, techniques.

The most enlightened training programs emphasized mastering concepts over memorizing rote skills.

2 approaches: Proactive vs. reactive



Proactive Reactive education education

Education director compiles catalog of approved courses and course vendorsLine organizations initiate requests for a course based on their immediate needs.Most such courses emphasis rote skills over concepts ("hit the ground running . . .")

Role definitions specify skills requiredCourse descriptions specify behavioral objectives based on those skills.Education director plans curricula to prepare staff to assume roles.



Programming standards1980: Nearly every organization that had programmers coding in Fortran or COBOL disseminated programming standards.

1997: C++ and Java present 100 times more choices of technique than Fortran or COBOL

Many of those choices lead to extreme maintenance costs.

Practically no organization disseminates C++ or Java standards to its own staff or to outside contractors.



Quality Assurance1980:The most enlightened I.S. organizations conduct formal and informal reviews of work products:

Note deviations from applicable standardsDesign and code walkthoughs

1997:Most application software installed with no quality review, esp. when developed under contractThe term "QA" persists in recruiting ads, but only in the limited sense of late-stage testing (finding "defects")



Discouraging reactions to thecollapse of I.S. infrasturcture

"It's not really a widespread problem. You're overreacting to a few isolated situations."

"It is indeed a serious problem, but there's nothing we can do about it."

"It's not a problem for our graduates. We require a project management course."

"It's a real phenomenon, but it's a good thing. We have to be able to react in Internet time. The old methods don't apply any more" Should we give up?



ConsequencesMassive, conspicuous fiascosobject-oriented fiascosdot-com fiascosclient-server fiascosweb-site fiascos

Not just initial project failures, but ongoing maintenance nightmares.Cause & effect unrecognized.

Lessons are being learned again, but will they be remembered? Do decision makers care?



System development in the 1990's

Then why do all our systems still turn out to be excremental?

We've adopted an incremental approach

to developing our applications!



Situation 4An "executive briefing" at a breakfast meeting sponsored by a major university.Keynote speaker CEO of a highly successful Internet based enterprise related his experiences and advice:

Spent nearly $30 million on software development originally estimated at < $5 million."Dont trust I.T. people.""Don't trust consultants"Watch out for packaged software (SAP) that doesn't do what you require.

What can we conclude?What was the audience reaction?



Positive counter forces

SEI Capability Maturity ModelEmphasis on repeatable processes and predictable results

ISO 9000-3

Are they still relevant? Which way is the trend?



Capability maturity model

5 "stages" of an organization's software-development maturity:

1. No methodology

2. Repeatable

3. Defined

4. Managed

5. Optimizing

Project management

Methodology focus

Measurement



An encouraging sign:Increase in public airing

Current examples:DePaul CTI alumni group seminar on SEI CMM Rich Rod, United Airlines, October 20, 2000Ed Yourdon: "Long Term Thinking", ComputerWorld, October 16, 2000Lisa J. Roberts: "Join the Revolution", Software Development, November, 2000This ACM meeting

Once again a respectable topic?



Remaining negative signsRecruiting obstacles:Recent ads (ComputerWorld, Chicago Tribune, etc.) as well as on-line recruiters (Dice, Monster, etc.) list no openings at all for infrastructure staff roles, such as:

Methodology administratorProject-control consultantProfessional education directorQuality assurance managerData administrator

Brokers and recruiter representatives: have no awareness whatever of infrastructure. are actually repelled by productivity claims.



Justification obstacles1980: Justifying the "overhead" of staff and other infrastructure is virtually automatic:

Huge multiplier: every project in the organizationBenefits: ?

Even if benefits only 5% the multiplier justifies the investment

1998: Decentralization removes incentive:"Yes, we really need all that, but my department can't afford it."Requires understanding and strong support at the highest level of the organization.



The bottom line

Is is hopeless?

What needs to happen?


The Collapse and Rebirth of Information Systems … · project manager repeatedly reminds team that the one thing that's absolutely beyond consideration is any further slippage; the

Documents