NO, that's what proper design & development process is for. Testing is only a (small) part of that process and as we've known for decades, you can't test quality into a product. Hell, just a few minutes ago, someone brought me a piece of code that has been HAMMERED on for the last nine months and it still has a bug that wasn't noticed until today!
gdp is right: certifications are useful if they are rigorous, targeted and demand useful knowledge and training
I doubt most developers would say that licensing pilots or physicians is useless but that's essentially the same thing as claiming that certifying programmers is useless.
When a pilot or a physician makes a mistake, there is no chance for someone else to correct it. That's why these professions (physicians especially) require expensive certifications, and continuing expenses to keep these certifications current.
Engineers are probably a better comparison: while engineers do have expensive certifications, they aren't the ones doing most of the engineering. Instead, they're the ones verifying that the engineering (done mostly by non-certified engineers) meets the design requirements.
So because there's a chance for someone to fix a bug is a reason that certifications are useless in software? Not buying it for a second. Whether or not the bug can be fixed is immaterial: it may have already caused thousands of $$ of problems even if fixed. Tell that to a business whose eCommerce site has been down for a day because of a dumb error by an inexperienced programmer.
As far as your comment about engineers being a better comparison and axod's comment that "most software is fun..." (WTF!!! do you think businesses expend millions of dollars yearly on IT so they can have "fun?"), I call bullshit! As someone with experience and in both hardware and software engineering, I can assure you that most engineering is verified by non-certified (I assume you mean PE's) engineers.
I live daily with software that has to work. Serious bugs are not tolerated in medical devices. The attitude that "oh, we can fix that in the next release, just tell the operator not to hit F8 before F7 or they'll misdiagnose the patient" simply doesn't fly around here.
Sorry for the rant, but attitudes like this are exactly why software development isn't taken seriously by most professions!
You're making a leap here. The chance for someone to fix a bug does not make developer certifications useless. The chance for someone to catch a bug during review, before it harms anyone (which is what I was referring to) still does not make developer certifications useless. The latter does, however, decrease their marginal utility. This is why it's an apt comparison to engineering: an engineer can't "patch" a defect in his design once it's been produced, but he can have other engineers look it over beforehand.
Now, what I mean by "certified" varies a bit across disciplines. The professional engineer cert is a general certification, which is generally much less useful than specific certifications for particular disciplines. The considerations in electrical engineering are considerably different from those in, say, structural engineering. While there is also significant overlap, you're not going to get a general-purpose "engineer" who can verify both.
gdp is right: certifications are useful if they are rigorous, targeted and demand useful knowledge and training
I doubt most developers would say that licensing pilots or physicians is useless but that's essentially the same thing as claiming that certifying programmers is useless.