'Speed wins in the Net age, even over big structures and CMM levels'

He

is the guru of software engineering methodology. He is the man who brought order

and sanity to large system development–projects with hundreds of engineers.

His work helped describe systems from the top level down to small pieces that

could be programmed, so that they would all link together later. This served as

the foundation for CASE tools. Edward Yourdon is widely known as the lead

developer of the structured analysis/design methods of the 1970s, co-developer

of the Yourdon-Whitehead object-oriented analysis and design and of the popular

Coad/Yourdon OO methodology.

In the late 1990s, Yourdon began to say that those methods were not for

everybody. Finally, in the most recent of his stream of books–Death March: The

Complete Software Developer’s Guide to Surviving ‘Mission Impossible’

Projects–the creator of software methods said that those formal structured

methods are irrelevant in today’s Internet-speed world.

Today’s developers may argue about whether Yourdon’s assumptions and

methods became anachronisms in the Net age, or whether they were always wrong;

but his impact on and contribution to software is undeniable. There are few in

the software world with his span and depth of experience. In 1964, he started

off with creating a Fortran math library for the PDP-5 microcomputer, with

Digital Equipment. For the rest of the twentieth century, he worked on over 25

different mainframe computers and a number of pioneering technologies such as

time-sharing operating systems and virtual memory system. In 1974, he founded

Yourdon Inc, providing educational, publishing and consulting services in

software engineering technology. He has authored over 250 technical articles and

26 computer books. He serves as the chairman of the Cutter Consortium, a

research and analysis firm. He also serves on the board of directors of iGate

Capital Corp, a holding company of 10 global e-services companies, and its

Bangalore-based subsidiary Mascot Systems, a provider of solutions in the

e-business, business intelligence, custom solutions, application maintenance

outsourcing and application re-engineering areas. DATAQUEST met Ed Yourdon in

Bangalore, and listened to his views on a range of subjects, from the US economy

to the new economy. Excerpts:

On the present state of the US economy:

Nobody knows what’s going to happen to the US economy. Even professional

economists don’t really know, and I’m not one. Part of the reason is simply

that it’s so enormous, with a $10 trillion GDP. Some indicators do suggest

that that economy is slowing down. Automobile sales are down, the Christmas

shopping season, crucial for the retail industry, grew much less than it did

last year. The consensus is that IT spending in 2001 will probably increase by 7—8%.

But last year it went up by up to 12%, so there’s going to be less growth.

On whether this really is a recession:

The official US definition of a recession is two consecutive financial

quarters of negative growth–not a slowdown, but a decline. That has not

happened yet. A lot depends on consumer spending, attitudes, opinions and

consumer confidence. And of course, there are strong political forces that would

like to keep up this optimistic outlook. To keep consumers feeling very happy

can earn us a lot of money. Two-thirds of our GDP is based on consumer spending.

Then came the complication of the recent election, which involved a change of

political parties. Keep in mind that what you see often

overemphasizes things. "Decline in IT spending" really means decline

in growth in IT spending.

On the impact of the slowdown on Indian companies:

Well, it’s already forced down the stock prices of American high-tech

companies, and what happens in the US will happen here in India too. A visible

impact will be lower venture capital funding in IPOs, and the overall lower

pricing of hi-tech companies. But a positive impact is more opportunities for

outsourcing, for if there is some cost-cutting or belt-tightening in the US IT

market, they’re going to be looking at how and where to save money.

On what Indian companies need to watch out for:

Competition. More than ever before. Many US software and service companies

still see India as "this low-cost, off-shore place that does all the boring

work that we don’t want to do". We don’t do Cobol programming…let the

Indians think about it. That’s fine as long as we have lots of exciting work

and contractual jobs, Web development, e-business. But when all of this

disappears, we suddenly face the prospect of unemployment. And Cobol programming

doesn’t look so bad after all.

And then there’s the distance, which means that you may not have such deep

customer relationships. A typical Indian company may have a small local office

for marketing, whereas the American company is right there in the small town.

Finally, the chance that North American or European companies might put off

by a year or more those plans to outsource their entire IT operations. There are

risks and up-front costs associated with that sort of outsourcing decision–and

the return on investment can happen two years later. IT departments are going to

be under severe costs pressure this year. "Do we really need to upgrade all

our computers this year? No, we could do that next year." They’ll tend to

defer any unnecessary investment decisions, anything that looks risky, where the

return on investment isn’t quick.

On how companies should approach software engineering and quality:

In the Internet and e-business age, a key issue is speed. We’ve got clients

with a need for critical systems, the world over. There are heavy methods, and

light ones. The former use lots of SEI-CMM levels, which was a growing

phenomenon in the 90s. Then there’s a new software development process called

XP–extreme programming, a term coined by Kent Beck.

There is nothing magic about it: it is relatively informal and so helps you

deliver high-quality systems.

On the future of dot-coms:

I think there’s going to be a bit more shakeout, collapse of some of the

dot-coms because they’ve run out of cash. You may see a further decline.

Things will stagnate for a while, which will be frustrating.

The American media headlines–"20,000 layoffs in 2000"–are

over-hyped. 20,000 is only a tiny percentage of the 2—3 million US people

employed in the IT industry, while there is a shortage of 200—300,000 IT

professionals.

On how dot-coms can succeed:

The success mantra is an "old-economy idea": a credible business

plan that provides a continuous revenue stream. You need to go back to basics

again. You need to understand the technology and its importance. What happened

in the US is very interesting: even after the dot-coms started falling apart

last year, venture capital groups were planning to continue investing.

A DQ report