We’ll still be playing piano in 10,000 years. I would hope we would have the gumption to be the generation who painfully went through the transition to make the piano accessible to all.

Can we switch from base-ten to base-eight while we're at it? That'd be far more painful, but also hugely beneficial in the long run. For one thing you could use the metric system everywhere while maintaining the ease of halving and doubling that the imperial system offers. It would obviously make various aspects of programming simpler too. But it's hard to imagine it ever happening, since the transition would just never be worth it (within the lifetimes of the people making the decision).

Not to derail this conversation, but I'm really curious: what ease is there in halving/doubling in the imperial system?

Halfs are easy enough in either. Quarters are hard in metric as 4 is not a factor of 10. Which is why imperial systems are often base 12 (gives you 3 and 4 as a factor). circles are base 360 so that you get factors form 2 all the way to 12.

How much this matters is open to debate, but base 10 is clearly a terrible thing in this. It might be a good compromise, but I am not sure.

Yeah, I meant repeated halving or doubling. So for example if you start with 1, and keep dividing by 2, in base 10 you get 1, 0.5, 0.25, 0.125, 0.0625, 0.03125, etc. In base 8, 1, 0.4, 0.2, 0.1, 0.04, 0.02, 0.01, 0.004, etc. Much simpler.

Obviously multiplying and dividing by 5 and 10 are more natural in a base 10 system, but dividing things in halves, quarters, etc. (or multiplying by powers of two) are a much more common operation. (We only think of 10s as round numbers because of the base-10 decimal system.)

12 is useful in that it has so many different factors, but it misses all the benefits of being a power of 2, so I feel 8 is superior. Hexadecimal also has benefits, but my feeling is that 16 symbols would be too many for everyday use. Difficult to count on fingers and such. 8 really seems ideal to me. In fact, I've mused in the past that it's kind of a shame - you can imagine a world where people ignored their thumbs when finger counting and just lucked into a base-8 system thousands of years ago. (Although of course in that case we wouldn't call it base-8, it'd be base-10!)

> I've mused in the past that it's kind of a shame - you can imagine a world where people ignored their thumbs when finger counting and just lucked into a base-8 system thousands of years ago.

What a fascinating hypothetical!

But to the main topic - sorry to be dense about it, but I'm actually just not used to the imperial system. Which parts of it are easy to manipulate in this way? I know there are 12 inches in a foot but that's the only "easy" conversion I remember off the top of my head.

The metric system, since it's all based around base-10 counting, tends to operate in decimals. So you wouldn't normally talk about for example a quarter cm; rather you'd have 25mm. In the imperial system working in fractions is much more common, so you have quarter inches, eighth inches, etc. Halving is more intuitive using fractions since it amounts to multiplying the denominator, rather than dividing.

Anyway, I'm not really arguing for the imperial system, just saying that aside from legacy use, that's the one advantage I see for it vs metric. But a metric system based on base-eight... that would be beautiful.

Oh interesting, you're saying it's not about the specific conversion factors, but rather the idea that people tend to subdivide units into halves/quarters/etc in imperial, whereas they don't in metric.

I just meant in imperial, it's common to use fractional measurements. IE 1/16", or 1/4 cup or whatever. Doubling or halving those fractions, especially repeatedly, is simpler than with metric measurements. IE taking 1/4 of 25mm.

How about a very slow and gradual transition? Say so that in one or two hundred years we have reached the optimal size by reducing the standard every decade just by a little.

That's a cool idea, but even that rate would probably be too fast. The piano I bought for a few hundred bucks on craigslist is 113 years old this year and still going strong. Good pianos last a long time.

I was surprised when I recently learned that the big concert pianos that are used in the music halls usually lasts less than a decade. But of course, they are played constantly. I assume a home piano that in best case gets played an hour a day or so last longer. And most home pianos are rarely played at all.

I was with you until this bit:

> This is also why I'm not so enthused by USB-C...

Please elaborate?

I hate USB-C with a passion -- the spec is a shitshow, some cables will catch on fire if used for PD, the connectors can be snapped off the PCB, as have happened to me pretty frequently, and the plugs are regularly made too damn long and are nice cantilevers. Everything is wrong about its design and ruggedness for consumer use.

Consumer electronics needs to live upto higher standards than even military. One should be able to plug their phone into an external battery, shove the entire mess in their pocket, and go skiing. That's what consumer electronics should live upto. The cable housing should be supported snugly and firmly by the housing of the device, like one would do on a BNC connector, but on a smaller scale. The PCB shouldn't be taking all the torquing forces, as is currently the case.

USB-A has become practically a universal standard for DC charging, and USB-A drives and peripherals are ubiquitious. USB-C breaks all of that. It's definitely smaller, and that's great, but I'll take a slightly slicker laptop that doesn't make me carry around an adapter.

The other side being USB mini-B, though, has problems. The mini-B connectors are notoriously fragile. C is clearly a win here. Using the same connector for both ends is clearly a win here. And moreover: laptops can and do (and imo should) have both A and C ports. So USB-C isn't as disastrous as a piano, which can only have one key interface.

Mini-B seemed to have an exceedingly short lifespan, even by tech standards. I have some mini-B gear, but micro-B seemed to quickly supersede mini-B. (Micro also has some fragility issues of its own, of course.)

You could have said exactly the same thing with USB itself was first introduced, though. And lots of people did: there were endless, massive complaints about the first iMac.

Right—I totally see both sides here! Some transitions are absolutely necessary, and worth the trade-off.

Part of the problem in my view is that, if you're going to make these sorts of switches, you ought to be planning ahead 30+ years. If you told me USB-C was going to be the last iteration of the format (at least for several decades), I'd be okay with some weirdness in the interim. By contrast, if you told me the USB forum was going to introduce USB-D in 2030, I would absolutely not be willing to put up with this.

USB type A/B connectors lasted for >10 years, but were obviously inadequate for most recent devices (which is why there were mini and micro B types, a type A “superspeed” connector with an associated "extended" micro B type, alongside a large number of proprietary other connectors), without even mentioning the worst type A misfeature of appearing symmetrical at a glance but not being symmetrical. Keeping type A/B as the default for another 10+ years would be horrible.

There is no apparent reason why USB Type C connectors will not last significantly longer. They support a much wider range of use cases (collapsing what were previously at least 4 or 5 different connector types into 1, and using the same connector at both ends), and were explicitly designed to be future proof. The people involved in the design were thinking decades ahead.

It is unfortunate that many device makers (e.g. people making power strips) have taken a while to adopt Type C. But in another 5 or 10 years the transition will just be a memory. At most you’ll need an occasional dongle for your legacy hardware.

The biggest future competitor for USB type C plugs is various kinds of wireless transmission.

USB4 should help somewhat here by effectively being a streamlined subset of the initial USB-C + Thunderbolt 3 specs, as well as finally giving a clear definition for how hubs should work (in my eyes, the one fatal flaw of USB-C so far).