How Do You Test a Battery?

I’ll say one thing for the Fandroid army: when they don’t like something, they let you know.

This week, I tried out Flash on an Android phone. I noted that I didn’t actually run battery tests, but that Adobe says you can expect 3 or 3.5 hours of battery life if all you do is watch Flash videos.

Android fans didn’t care much for that. “I’m really disappointed how sloppy this article is,” wrote one. “You say that battery life could be an issue, but you’re too lazy to test it.”

Another: “You completely made a side issue of the most important aspect to mobile users: battery life.”

And: “Thanks for not doing the battery test. That could have been the only interesting part of this piece.”

Then: “You didn’t conduct battery tests?!!! That’s only ONE OF THE BIGGEST ARGUMENTS AGAINST FLASH ON A MOBILE OS. Jeez, get a NYT intern to test the battery.”
And of course: “How could you then blithely just skip a battery consumption test?”
Hmm. It almost seems like you guys think I should have run some battery tests.
To you, my failure to test Flash’s battery hit is “lazy.”

To me, it’s a little more complicated. My problem was, “How do you fairly test the battery?”
One way to do the test: spend a whole day visiting Flash Web sites and playing Flash videos. On a second day, visit exclusively non-Flash sites. Compare the battery life.

But come on: how many people actually use their phones that way? Who would use the cellphone for nothing but watching Flash videos, all day long? Isn’t it more likely that the typical person will, during a typical day, make some phone calls, do some e-mail, play a game or two, surf a bunch of non-Flash sites, and maybe visit a few Flash sites? In that real-world situation, how measurable would Flash’s hit on your battery life be?

Furthermore, almost nobody uses the battery continuously, draining it until it’s dead. Most people use their phones sporadically during the day — a pattern that will provide totally different battery-life results. How to factor that in?

Some of you suggested using sites like YouTube, where videos are available in both Flash and HTML 5 (the kind that the iPhone can view). Test the battery by playing Flash videos on an Android phone, and compare with watching the same videos in HTML 5 format on an iPhone.

Well, surely you see the problem with that protocol. You’re not testing the effect of Flash. You’re testing two completely different phones. They have different processors, storage, batteries and other components. It’s not a meaningful test at all.

So here’s my challenge to you: design a testing protocol for Flash’s impact on an Android phone’s battery life, and I’ll do it.

Descend on the comments. Let’s get this thing rolling! Can’t be lazy, now!