When is something good enough?
I think the answer is complicated but if we boil it down to two simple measurements it becomes easier to answer: perfection level and expectation level.
The perfection level measurement is, of course, the most difficult because there is nothing perfect ever made. So one must then set a reference level of how many imperfections are allowed to be able to measure it – and then accept the imperfect results. Not an inspiring model for building a product.
The expectation level, however, is probably a much more attainable measurement and far more meaningful to the end user. Let’s focus on that one.
If we produce a product that meets or exceeds every expectation you might have then by default it was not only good enough it was better than good enough. That’s a solid win even though the product is only “good enough” and is imperfect and perhaps has multiple issues.
This subject came about because my engineering director and I have been studying how to become better software programming managers (we both come from hardware backgrounds). What we’ve learned is that in an environment where we don’t control the hardware, like a program running on your PC, software becomes a black hole sucking in resources at a constant and infinitely long rate. That’s a reasonably eye opening concept to get a grip on.
It’s an even bigger eye opener to learn that “good enough” is the new standard for development.
And all along I thought it was perfection.
Forward to a friend and help us engage more readersYou must be logged in to post a comment.
Andy Stoneman
I guess that, beyond the basic mechanics of your product, what you are trying to achieve is quite different to most programming, Paul. In that the majority of software has clearly defined objectives; add these two numbers, display this, actuate that etc. Whereas how something ‘sounds’ is open to interpretation?
Paul McGowan
Exactly, it’s subjective and when software is working on something objective it really gets crazed. Certainly there are clearly defined objectives we work with all day long and these are relatively easy to hit, verify and implement. It’s when subjective goals get involved added to the complexity of today’s programs that it gets truly interesting.
Tom Devey
I have never heard of measurements being described in terms of ‘expectation level’. I would argue that the expectation level in the measurement of a product’s parameters are its specifications as measured in accordance with a recognized standard which takes into account the tolerances associated with the measurement process.
Paul McGowan
And I would agree with you on the surface – however as I mentioned in my reply to Andy, there’s a subjective nature to high-end audio performance that’s never been quantified and is elusive to measurement – and therein lies a big problem.
Tom Devey
What physical parameters in the audio reproduction chain are elusive to measurement? The only parameter I can think of, and its not a physical one, is the psycho-acoustic interface between the speaker output and the listener. Even that has been addressed over the last 40 years, though perhaps there is more research to be done on the subject.
I would suggest that if there is a problem, it is not one of measurement, but correlation between cause and perceived effect.
David Bevin
“That’s a solid win even though the product is only “good enough” and is imperfect and perhaps has multiple issues”. How interesting that this was related to a software issues: Microsoft has been making a killing on that premise for years!
Paul McGowan
Exactly! No Microsoft or Apple software products are perfect or even close. They are good enough for most of us.
petewilson
Paul
The effects of software on sound are always interesting. But if you’re in the early days of learning how to do embedded software, I suggest you sign up with The Embedded Muse, a bi-weekly email newsletter by Jack Ganssle.
http://www.ganssle.com/tem-subunsub.html
Ganssle isn’t always right, but he is more right than wrong. And there’s always interesting stuff.
hahax
I was glad to see you mention expectation. I think of the times when a system sounded better than expected and boy didn’t I think, wrongly of course, that we were really getting close to the ‘real’ thing.
And I recall the stories of live versus recorded demos when the phonograph was new and the sound was worse than an old telephone and the listeners were amazed, some thinking that it sounded just like the real thing – expectations.
Andy Stoneman
Agreed. Expectations are contextual and evolving. Early digital photographs looked amazing – for a digital photograph, but compare recent shots to even those only 5 years old and try blowing a 35mm shot to 30″. Early plasma screens were amazing – so thin and 42″ across, but the picture was awful. A 1960′s Ferrari had breathtaking performance, but today it’s outpaced by a hot hatch and would fail any safety test.
I recently heard a superb valve/horn analogue system – it was great, but no better than my fully digital system. And we both agreed on that!