A JAB in the Dark

by Steven Myers

Not too long ago, the American Electronics Association (AEA) issued an alert regarding a plan by the Japanese Accreditation Board for Quality System Registration (JAB) to implement a Software Quality Registration Program in Japan. Under this proposal, companies seeking to sell any type of software in Japan would have the "voluntary" option of submitting their product to a third-party JAB-accredited reg- istrar for inspection and registration.

Concerned that this type of program would hinder foreign access to Japan's software markets, representatives from the American Embassy and the AEA have voiced strong opposition to the JAB proposal. In response, JAB has dismissed the AEA press releases as "extremely one-sided," and in July prepared a document purportedly addressing the issues raised by the US and presenting JAB's view of the situation.

The JAB position

Eager to hear the JAB side of the story, I called their office to request a copy of all papers written by JAB that refute the points raised by the AEA. Upon reading carefully through the documents, however, I found them to be sadly lacking in several key areas. Essentially, JAB emphasizes the following points:

* Great pains have been taken to ensure that the proposed program is in sync with those of other countries, and not based on a "uniquely Japanese" perspective.

* US claims that the ISO 9000 standards are not meant to be applied to software are misleading since one of the areas listed for coverage under ISO 9000 is "computers and related industries."

* The program is strictly voluntary, and as such does not represent a non-tariff trade barrier.

* The program will not affect the price of software products in the way that the US claims, because the worth of such products is decided by the market. And even if it does raise vendor costs, if the software product has merit, it will sell well (and if it doesn't, it won't).

Obfuscating the issues

These assessments may be valid, but even if we assume that they are, the JAB rebuttal appears to deliberately ignore the main issues, completely omitting any reference to several important points brought up by the US side. Further, it is sorely lacking in specifics to substantiate the claims that JAB does make.

Most people are not so concerned about a software quality program per se based on the ISO 9003 standard. The central issue in this case is that the current proposal does not appear in any way to be international. JAB would have tremendous power in deciding which foreign companies are able to get registered (because JAB will, according to the guidelines of the proposed program, control accreditation of the registrars). This is the salient point: companies must be registered by a JAB-accredited registrar, which at present are all Japanese. If certification by foreign-accredited registrars will not be accepted (which is what JAB seems to be saying), this completely goes against the general policy for international standards employed by other countries.

Which brings up a second issue: How exactly will the registrars determine whether a product or design process is of "acceptable quality"? For that matter, how are evaluation criteria for software quality to be set? So far, JAB has said nothing specific or otherwise about the actual auditing process. This has raised strong suspicion on the part of many US companies that the Japanese registrars will insist on having access to proprietary source code and design procedures. This important concern has been voiced repeatedly by the opposition, but it is not even mentioned in the JAB documents.

Poor quality US software?

JAB also sent me a paper by Professor M. Azuma of Waseda University, chairman of the Investigative Committee for Software Quality Registration System. Prof. Azuma's paper was also written in response to the American criticism, and it basically follows the same lines as the JAB paper. Azuma concludes his essay by saying:

"I have marveled at the abundance of ideas found in American PC software (including freeware), but at the same time I have had trouble with its poor quality. I would like to see American software vendors take responsibility for improving the quality of their products. Rather than using their energy for useless fighting, they should be making efforts toward this goal of improved quality."

Upon, reading this, I frankly wondered, "Poor quality? Compared to what?" If Prof. Azuma has been having such a hard time with his free American software, why doesn't he just use something else? Something Japanese. Of course everyone gets irritated when Windows 3.1 produces one of its famous GP faults, but that doesn't mean they would rather pay more to use a JAB-accredited version of Windows (which no doubt would still cause problems now and then).

Let the market decide

Software products evolve naturally in their own way. Initial versions are rarely fault-free, but if the program offers useful features and conveniences, then users will willingly put up with minor bugs. At any rate, no amount of "auditing" is going to improve this situation. New software offerings will never be perfect, but many of us (including the Japanese) would rather use them anyway than wait for that elusive "bug-free" version. The JAB rebuttal shows an apparent disregard for this fact and fails to address key questions about implementation of the proposed program. Furthermore, no answer is provided to the question of exactly how JAB hopes to improve software through the program. In short, the rebuttal is entirely insufficient, and ends up raising more questions than it answers.

I'm still in the early stages of investigating this issue, however, and in the interests of fairness I plan to meet with JAB representatives in the next couple of weeks to hear their side of the story. Watch for a full report of this volatile issue in next month's Computing Japan.ç