1. First, I agree fully with James Bach. Bias aside, best practices are like best friends: You never know when you’ll have a falling out. That’s sad, I know, but since I was little, I never wanted to choose a “best” friend because I thought it devalued the qualities of the other friends I had. Perhaps I’ll choose one when I’m old?
2. Through conversations with coworkers, most on the programming side, there seemed to be this implicit understanding that “best practice” really meant “the best practice we know of right now.” This still doesn’t pass the context argument, but there seems to be an understanding that “best,” here, is not absolute over time… though it sometimes seems to be touted as so. There also seems to be this notion of, “Most bridges are built in way X, and therefore, way X is the best practice.” This might actually hold some weight in the development side of software engineering just as I assume the phrase holds in some contexts within structural engineering — I mean, let’s not forget Tacoma Narrows. But in testing, I’d assume stakeholders’ interests (among other variables) are too fickle and disparate to choose one “best” way of managing them. (I’d love to hear more from structural engineering folks to see how this phrase is used, if at all.)
3. In my company, “best practice” feels like a marketing term. This is one of James’ arguments. If the quality of the software produced is a measure of the practice (this isn’t always the case), I wouldn’t say our practices are “the best.” That said, if there’s truth in advertising, maybe there’s a context — perhaps a geospatial one, an area where TOPP excels — where we really are leading the practice. Does that mean we’re the “best”? I don’t know.

I talked to the CEO of an optics-based software company recently, and he described his product not as the “best” product on the market, but as the best for a certain context — in his case, the high-end precision optics market. Not the low-to-mid end, where software is of lower quality but is much cheaper. I felt his description and use of the term “best” worked because he gave a context, though he was sampling over a single variable, quality, which to each person is fairly subjective — that is, if you subscribe to Jerry’s view, described here.

As humans, though, we seem to place the label “best” by ignoring variables we think are uninteresting, all with the intent of rallying behind a common leader. At one point, the software industry was led by IBM. After that it was Microsoft. Now it’s probably Google. Which is the best, you ask? That’s like asking which is the best car company. Maybe what you’re really asking is, “Which company hasn’t turned into an 800 lb. gorilla?”

I’ve probably belabored the point already, but the word “best,” in my opinion, is simply one person’s perception of the environment in which things exist, sampled at a certain point in time. They choose variables that they’re interested in, explicitly or implicitly, then find a leader amongst those available. What they don’t take into account when assigning their label is the concept of time, or that over time perceptions change. Or maybe they take time into account, but the “best” label simply doesn’t hold as things change. And we all know how fast the software world changes…

But again: The variables they sample over may not have value to others, even if the label holds, for them, over time.

My last analogy, simply because I had to hit home while making a cheap shot at Britney Spears, is that maybe the concept of “best” in “best practice” is just as fickle as “’til death do us part” in 50% of American marriages. Perhaps we assign the term too quickly. Or, maybe, we’re just looking for social status in a world that rewards that sort of thing.