 |
Date Posted:
Friday October 05, 2001 12:50:20 PM
|
|
This article is an example of what I call assumptive reasoning. The basic argument, as far as I can tell, goes like this:
1) Software developers do not generally have sufficient appropriate skills to produce good software in the real world.
2) Traditional computer science programs do not teach enough about physics, engineering, accounting, etc. to produce graduates with the required real-world skills.
3) The fix consists of licenses and certifications, or at least they are a step in the right direction.
In an assumptive argument, the conclusion (in this case 3) is simple assumed, rather than supported with evidence. All arguments are peripheral. The main tactic in an assumptive argument is to deflect or rule out questioning the assumption. By the responses of mmarcos and Red Rubbie I guess that it has been largely unsuccessful.
An excellent example of assumptive tactics is near the beginning of the article:
"No, there are frequent complaints about them; nobody would claim perfection. I would still argue that we're better off with them. Let's take doctors as an illustration. Some people say that doctors are too conservative and don't use natural medicines and things like that. They want to broaden the rules so that chiropractors and those who use natural remedies can practice as doctors and so that doctors can use some unconventional therapies. On the other side there are many people who notice that doctors make many mistakes, are often negligent and even incompetent. Those who note those failings want to tighten up the system. It's a compromise. I don't think doctors are perfect, and I don't think engineering is perfect."
This contains two shibboleths of assumptive arguments: 1) the setting of options along a single, linear, monotonic axis ("conservative" vs. "broaden") and 2) the notion of compromise or balance.
Assumptive arguments are common in politics and criminal law, where they are usually strongly emotive. For example, the rights of victims are often "balanced" against the rights of the accused. The hidden assumption is that the intensity of a victim's perceived suffering is positively, monotonically related to the probability that any given individual committed the crime. Because this assumption is so ludicrous and has recently been directly refuted by DNA evidence of sex offenders imprisoned many years ago, nobody tries to support it with logic, hence the assumption.
Dave Parnas' assumption is hardly ludicrous and not nearly as emotive as this, but yet I believe it is incorrect, and at the very least it needs to be examined. Here is why I think it is wrong.
1) Certifications
The term "certification" already means something with respect to computers. It is an item of vocational training which is widely viewed as an alternative or in some cases an unrelated supplement to a degree. Most certifications are offered by specific vendors and are about the tricks needed to operate their products. There is currently a blue-collar attitude amongst some certificate holders and many managers that looks down on education. While the criticism of computer science programs is to some extent valid, it is also true that there are many certificate-holders that not only wouldn't know DeBroglie's laws if they sat on their faces, but are fiercely proud of the fact. There is a strong culture, not universal but just short of dominance, which holds that knowing anything more than the bare minimum is bad. Finding this out was one of the most disturbing things I learned upon making the transition from a career as an academic research scientist to a career in industry.
Dave Parnas, by his description, has produced a pretty good program, though I wonder about the courses in thermodynamics. This may be a US/Canada cultural difference, though. In the US most students learn about thermodynamics in general physics classes, sometimes in high school. At the CS level, a course in information theory, which builds upon knowledge of thermodynamics and ties it into coding, compression, etc. is essential. Anyway, I'm sure that he means something very different by "certification." However, given the fact that a well-established culture of certification already exists, it would be worse than useless to try to take back the concept.
2) Licenses
Are licenses useful for doctors? Possibly, as they are backed up by extensive training and peer review. On the other hand, there are doctors who have lost their licenses yet continue to write diet books. Are they useful for hairdressers? Possibly, as hairdressers do seem to be pretty good about cleaning their combs these days.
However, licenses are also useful to people who treat cancer with coffee and tobacco enemas, people who massage inflamed gall bladders, people who claim to see the future in a lump of quartz, and the like, who are all "licensed professionals."
Licenses may provide two things: 1) a positive correlation between having the license and being competent, and 2) a positive correlation between having the license and being perceived as competent. For licenses to be good, not only does 1 have to be true; it has to be demonstrably stronger than 2.
This is difficult, because a license essentially means only the ability to persuade a bureaucrat to allow one to practice, and bureaucrats do not have much of a clue. The public who would use their services have a hard time telling the difference between a piece of paper that means something and one that does not. Licensing of physicians and lawyers works moderately well only because both professions have hard, centralized, hierarchical control organizations (boards and bars) that are cultural descendents of a time when both professions were managed by the Church (hence the name "profession," an abbreviation of "profession of faith").
Developers, on the other hand, have a widely non-hierarchical organization. Although old-time hackers speak of wizards and gurus, these are strictly meritocratic terms and are in fact being driven out by the modern "IT professional" mindset. A comparable universal hierarchical organization does not exist and possibly cannot exist. However, it is a prerequisite for any talk of licensing to begin. A person of mediocre competence who is perceived as such is not a real danger. A person who is perceived as more competent than is true, on the other hand, is a danger. At the present time, licenses would increase the danger.
I may be wrong about this. However, at minimum, it is part of what needs to be discussed. The nonexistent ideal fix always seems better than reality, but that says nothing about the effects once implemented and made real.
|
