Hacker News Reader

Comment by yongjik | Hacker News Reader

My favorite comment from TFA:

> None of your math makes sense. All your measurements have only the consumption at one exact instant, not over time, the total energy used is the integral of the energy flow over time. You'd need to measure the consumption over time, or total discharge of the battery, to actually have the energy used to load a page (in J, not W, you can't measure the energy required to load a page in W).

> You just multiplied a banana times Pi. And showed the results in meters of strawberries.

View on HN

Everyone's dodging the point of the post (versus the exact numbers and terminology). As I said on reddit as well:

> The issue is that the redesign uses more power across millions of devices, and that has an impact.

> I might have gotten the exact number wrong, but the exact number doesn't matter that much. It's about it being a significant difference.

Simply put, you don't have the expertise required to measure this. It's not just that you didn't get the exact number, it's that you haven't got a usable measurement at all. For all we know from your tests, the new site uses less power than the old one.

> I might have gotten the exact number wrong, but the exact number doesn't matter that much. It's about it being a significant difference.

Of course the exact number matters. Let's say you're off by a factor of 1000. Then the power difference isn't really worth worrying about.

How do you know it's a significant difference if you haven't measured it correctly?

The only reason we're even talking about this article is because of your outrageous number. When folks who recognize how these kind of measurements actually are done say, "this seems extremely wrong" you should definitely expect some pushback.

To be honest, this reads to me like a reddit user angry about a site change trying to make a sensational article to protest it, knowing cynically that even if it's very wrong it won't matter, the headline is too sticky for reddit to ignore.

Haha, yeah. I was considering jumping in with my usual "learn the difference between power (W) and energy (Wh)!" lecture, but couldn't have improved on that response.

This is the same mistake typically made by the "electric vehicles will overload the grid and we'll need 500 new nuclear plants!!" folks.

Also, the measurement involves reloading the same page infinite number of times which is really not even remotely represents how users probably use Reddit.

May be the new design is preloading content to make the subsequent clicks faster? May be the new design enables usage of HTTP resources with better cache lifetimes?

Reply on news.ycombinator.com