Standards
I came across an article on standards called “Standards Confusion and Harmonization” in the July 2007 issue of the IEEE Computer Magazine. It offers some insight into the debate on having standards, “proponents of standards claim that having a bar to be hurdled is better than no bar at all,” while those against say “that setting a bar creates a risk of it being far lower than it should be, resulting in a free pass to do less; without the bar, users would do more.” There’s also a very interesting analogy about buying medications. An average person doesn’t have the knowledge to assess whether or not a drug is safe to use or not, but if it is a prescription, they can be sure that it has been endorsed by the FDA. Personally, I would rather pay a bit more for prescription vitamins rather than buy them over the counter because the over the counter ones are not regulated. Back on subject, there is no equivalent of the FDA in the computer software world and no way for a product to be certified. Having an overseeing entity for the software industry is quite intriguing. Of course, it’ll probably never happen. Still, if I was going to buy a text editor, or a cell phone, or any other product that has some software running it, I would feel a whole lot better if it had a sticker saying “Certified.” This would be especially comforting if the sticker was present on a side of an airplane, or an ATM, or a medical device. On the down side, getting any software into production would take a lot longer because of all the testing and certification issues. But, it would create more jobs too. Someone has to verify that the product meets standards. The article proposes “standards blending,” a way to pick and choose bits of different standards that fit a particular product. They say that “if the goal is to demonstrate that a system cannot exhibit behavior A under conditions X, Y, and Z, it might be necessary to tap into several different standards and abide by their requirements to obtain the evidence of meeting the intent.” In my opinion, it is dangerous to take bits and pieces from different standards to make one that fits. Each standard, such as IEEE 1473, is developed over a long period of time, with a lot of work and testing. It is the way it is for a reason, and if a product needs to be IEEE 1473 compliant, it should comply with the whole thing. Going back to the FDA analogy, a drug could be compliant with regulations except that it uses a chemical in the shell coloring that hasn’t been fully tested. If such a drug was allowed to go to market, then it would set precedent for other deviations and eventually ruin the FDA credibility.
Leave a Reply
You must be logged in to post a comment.