letters.you

Software Magazine, August, 2001

08.01/09.01

Threat Is Larger Than Language

I read your article, "Programmer Shortage Threatens Mainframe Future," (April/May 2001, Elizabeth Harding p. 10) with great interest. I was especially intrigued by the premise that a shortage of skilled programmers may be the issue that brings the mainframe to an end of life. I agree with your observation.

I wonder if you have given much thought to which skills are actually at issue? References to languages such as Java and C and technologies such as XML would lead one to believe that the skills you refer to are primarily language and environment based.

The issue is much larger. The most important skills have to do with design and deployment of large mission-critical systems--the ability to consider backup and recovery strategies, journalization of input messages for subsequent replay following disk drive failures, and writing code for performance rather than elegance. I have not seen many fresh-outs who give these issues much thought.

The issue is not a lack of Cobol or CICS skills, but rather a lack of 25 or 30 years of in-the-trenches experience with enterprise systems. And it may well be that mainframes are not the only platforms affected by this shortage.

Michael Giroux

Phoenix

Absence of Scope Document a Concern

Your recent article, "Micro-project Methodology Defined" (April/May 2001, Jim Johnson, p. 20), pushed very small incremental projects. I don't disagree with your statements, but I have some concerns about the things that were missing.

One is the absence of a complete scope document before launching into design and development. Projects like this could go on forever and nobody would know when they were complete or whether they have been successful. This smacks of major scope creep.

Another concern is a technical one. Before launching into development, one needs to be sure that the design approach and infrastructure are adequate for the final product. For instance, an inadequate database design could require serious amounts of rework and conversion in the latter phases of implementation.

This article is really important. Rarely do articles step back and give such an overview of success. Keep this type of article coming. Many developers and managers have never been on enough large projects to have learned these "truths" by experience.

Gerald Davison

Naperville, Ill.

THE AUTHOR RESPONDS: My problem with a complete scope document is the meaning of complete--I think software should be a more adaptive and growing creature which is done in small steps. Sure you need a scope document, but it should be very basic. Given that, what would you suggest be in this document? JJ

ISP As Inoculator Limits Market

I enjoyed reading your column, "Virus Inoculation" (April/May 2001, Paul Raines, p. 72) and agree with what you say, but you may have overlooked one thing. Virus-protection software companies would soon go out of business or have to charge extremely high prices for their updates. You would take away millions of customers and limit the antivirus market to only the several thousand ISPs.

Other than that, it's a great idea, and one that we think about all the time. Security of data and technology is becoming more and more important to all of us.

Mike McLallen

Shreveport, La.

THE AUTHOR RESPONDS: I'm not sure antivirus vendors would go out of business if such an idea were implemented. Corporate desktops would still need some AV protection to guard against viruses contracted through diskettes. Also, ISPs would have to pay AV vendors license fees, depending on the number of users subscribing to the site. I'm sure the engine of capitalism would grind out some viable solution. PR

CORRECTIONS/CLARIFICATIONS

In the Software 500 Financials chart (June/July, p. 56), the column % Services Revenue contained a calculation error, making some of the percentages incorrect. The correct percentages are available at www.softwaremag.com. We regret the error. Also:

* Artemis International Corp. under-reported its revenue. The correct revenue is: $22.6 million software license, $30.6 million services, $15.8 million support, with total corporate revenue of $69 million. As such, Artemis would have ranked #194.

* Computer Sciences Corp. did not include outsourcing and consulting revenues. Total software/services revenue for 2000 was $10.2 billion, which would have ranked the company at #8.

* EMC Corp. did not include services revenue of $906 million in its total software/services revenue. EMC's total software/services revenue is $2.3 billion, which would have ranked the company at #16.

* Evolutionary Technologies Inc. reported its 1999 revenue. The company's software/services revenue for 2000 was $31.5 million, which would have ranked the company at #293.

* SS&C Technologies Inc. did not include maintenance revenues in its total software/services revenue. SS&C's total software/services revenue is $60 million, which would've ranked the company at #212.

* Veritas Software Corp. under-reported its revenue and was not included in the ranking. Total software/services revenue was $1.2 billion, which would have ranked the company at #27.