Forum

Pamela Samuelson’s “Legally Speaking” column (“Reverse Engineering Under Siege,” Oct. 2002) does a masterful job of explaining how a court’s interpretation of a law can affect the legitimate work of researchers and technologists.

ACM members and other Communications readers might be interested to learn that USACM, the U.S. Public Policy Committee of the ACM, signed a legal brief that was submitted to the California Supreme Court in the DVDCCA v. Bunner case profiled by Samuelson. The brief (co-authored by Samuelson) seeks to affirm that reverse engineering of mass-marketed products is a lawful way to acquire a trade secret. In addition, the brief also repudiates the notion that an anti-reverse engineering clause in a mass-market license can override the right to reverse engineer.

In our interest statement, USACM states that reverse engineering is critical for systems interoperability and facilitates the research and testing of information systems and the development of programs that impede the spread of viruses and other kinds of malicious software.

The interest statement concludes that restrictions on reverse engineering would have serious stifling consequences for software engineers, the computing community, and the overall security of the information infrastructure and e-commerce.

For more information about the brief and other USACM activities, please see www.acm.org/usacm.

Barbara Simons
Eugene H. Spafford
Co-Chairs, USACM

Back to Top

Criteria for Integration Technology

In the Article “Integrating Web Sites and Databases” (Sept. 2002) Mike Morrison, Joline Morrison, and Anthony Keys compare several technologies for integrating databases into a Web-based environments.

From my point of view, some of the main decision criteria for an integration technology are missing in this area.

Most notably, J2EE (Java 2 Platform, Enterprise Edition) is not rated correctly.

If I have to integrate databases in my own home-brewed Web site and am forced to keep my costs as low as possible, some of the script and client-driven features will be of interest.

Without having to be on the leading edge I can use simple scripts to build a browser front-end to a small database.

But to deliver a real e-business solution, many other questions have to be answered.

Following is an incomplete list of issues commonly seen in the evolution of Web-based applications with one or more databases in the back-end:

• Separation of data, presentation, and user input management. How can a Web designer build respectable front ends without knowledge of the underlying Web and database protocols?
• Security. How can the Web designer control different user roles, user groups, and access rights to databases?
• Session management. How can the Web designer use the stateless HTTP protocol for transactions, including at least two requests from the Web user?
• Distributed systems. What can be done if a business transaction involves transactions from different databases with different protocols and access methodologies?
• Scalability. Though we hope the number of users grows, will it be possible to serve 20,000 users as well as we serve 500?
• New channels. Can voice-driven input and output to and from the Web site be provided without deep modifications of the server-side applications and infrastructure?

For such issues and questions, J2EE helps conform application servers (such as IBM’s WebSphere, which offers standardized and platform-independent solutions).

JSPs and Servlets play a dedicated role in the complete J2EE architecture, which is influenced by approved design patterns and is more comprehensive than a compiled program or hybrid technology.

Christof Schmalenbach
Duesseldorf, Germany

Back to Top

The Current State of Hypercomputation

In their column “Hypercomputation: Hype or Computation.” (Aug. 2002) Christof Teuscher and Moshe Sipper describe the current state of discussions about hypercomputation. There are several doubts that persist on effective realization of this theoretical construction as they state correctly.

However, they forgot the most important one. There is no evidence that the infinitesimal physically exists. The first warning— energy is not continous—shocked the physicist community about hundred years ago. Since then a mixed model has been adopted and widely accepted: everything is continous but energy. This has led to serious and unsolvable problems. Quite dirty tricks like (re-) normalization were necessary to overcome the most obvious. The price is paranoia (Nature is watching us by watching her) and schizophrenia (parallel universes). Yet we still have a lot of unsolved problems.

We should admit and accept that the physical world is discrete, not only in terms of energy but of space, time, and all the other quantities as well. Indeed, this is a “New kind of science,” as Stephen Wolfram states.

If the infinitesimal is just a human notion, without physical equivalence, as I suppose, then all constructions, including hypercomputation, are just a kind of daydreaming. There are nice mathematical constructions of this kind. But that’s all, I’m afraid.

László Karafiát
Munich, Germany

Back to Top

Still No Licensing in Texas

I enjoyed reading Ken Kennedy and Moshe Vardi’s article (“A Rice University Perspective on Software Engineering Licensing,” Nov. 2002), but I am afraid some readers may be misled.

I have never met a Texas-licensed software engineer, and though the Web site for the Texas Board of Professional Engineers (www.tbpe.state.tx.us/downloads/ sofupdt.htm) says there is not yet an available exam for software engineering, there is a procedure for obtaining an exam waiver. I am not surprised the authors have not had many certification or licensing questions from Rice students.

I think readers may get the wrong message from the statement that four years of licensing in Texas “… had practically no effect, positive or negative.” For all practical purposes, there is no software engineering licensing in Texas.

Dave Isaacs
Dallas, TX

 Authors Respond:

We appreciate Isaacs’ comment. We definitely did not intend to imply we have removed our objections to the licensing requirement for software engineers in Texas. We stated our objections to the licensing requirement, and we continue to stand by this position.

We also agree with Knight and Leveson, who offered well-reasoned objections to the licensing requirement. Issacs is right in observing that our optimism regarding the situation in Texas is perhaps premature.

Our hope is that the marketplace will in effect neutralize the requirement in Texas or at least minimize its harm.

Back to Top