Hacker News Reader

Comment by HL33tibCe7 | Hacker News Reader

Strong disagree. “Did you just fall?” is more confusing: in that case, the watch is not directly telling you what it thinks happened; instead, it is asking a leading question.

It’s way less clear. You’re relying on implication to convey the fact that the watch thinks you’ve fallen, rather than just saying it.

If you’ve hit your head and are concussed, that extra level of indirection and confusion that your question introduces could have a significant effect on the outcome of the situation.

View on HN

I don't understand your point of view. There is no indirection. It's just a simple question, yes or no, did you just fall?

It’s entirely unclear to me what the consequences of the yes/no choice would be.

Furthermore, falling is not binary, so there’s the question about where the boundary is. In an emergency situation, I don’t want to have to deal with any of this. It’s terrible UX.

So maybe a message like

"Call emergency services?"

"Yes" "No"

Although it removes the context of the watch' reasoning.

"Fall detected. Call {911 || regionally appropriate number/name}?"

Exactly.

Again "fall detected" is quite un-human and mechanical. Is this a technical term? Who/what fell? Are you talking about the season? I think the existing Apple language is good and has clearly been through a lot of discussion!

As befits a nonhuman, mechanical device like a watch.

Which makes it sound like the watch fell, not me

Given the consensus understanding of physics and gravity, it did. It just so happened that it was attached to a larger thing that also fell, and the larger thing likely either knows it and can comprehend this message trivially, or does not and summoning help is the proper action to take.

Reminder that the screen is a false positive protection mechanism and will only ever be seen in a specific context.

Sorry, but that's just a pedantic take when we're discussing what is the best UI /UX for wearers to know how to instinctively react in a possibly stressful moment.

You have to solve for the lowest common denominator, not the hacker's technically correct preference.

This ain't about the watch falling, it's about someone possibly needing to call emergency services and contacts when they're hurt. "I'm OK" is a great prompt in that context, even if it's not an explicitly accurate response to whether the watch or wearer fell (which in turn is why the focus is not on the fall but on their need for help).

There also must be some joke about special relativity hidden in that wearer-and-their-watch-falling incident.

> the larger thing likely either knows [about the fall] and can comprehend this message trivially, or does not and summoning help is the proper action to take.

You did say "likely" but to provide an example: skiers who know they fell cannot comprehend the message trivially (not having noticed it's existence) and summoning help is often not the proper action to take. Call centers serving ski towns are swamped with false positives due to this issue.

Are you imagining someone who, having just fallen while wearing their Apple Watch, wondering whether they or their watch fell when their watch says, “fall detected” While they lay on the floor?

I'm imagining a scenario in which someone who, having just fallen while wearing their Apple Watch, concludes the watch is talking about itself when it says "fall detected" and they will have to have it repaired

"Fall detected. Call 911?"

“Why is my watching asking me to call 911 to get it repaired? My hip is broken!”

I can’t tell if you’re trolling me or not but I’ll assume you’re not.

If you’re laying on the floor I think you’ll know who fell.

If you didn’t fall, then you know it’s wrong.

But the watch didn't detect a fall. The watch saw signals and data points and thinks it could be a fall. Saying "Fall detected" is too confident and not a reflection of the feature.

You’re right. Better explain all that to the user of the watch who is laying on the floor.

If they are not laying in the floor then they already know it’s wrong.

"Fall detected. Call 911?"

“Huh? No, I didn’t fall. Don’t call 911”

“I did fall, but I’m fine. Don’t call 911”

Maybe you can explain to the user of the watch who didn't fall but the Watch suddenly says "Fall detected".

Apple's version is better. It's more accurate of what the feature does (guessing based on factors), and it's more human.

I wrote a Fall Detector app for Android, ten years ago.

When a fall was detected the sample played: "A fall has been detected" When the caretaker was alerted: "The caretaker has been alerted" When the caretaker acknowledged: "The caretaker is on their way" ..

I dunno why its so hard. Talk to the user, re-assure them that the app is doing its job.

But what is it going to do if I answer yes? If I did fall but don't need medical attention, is answering "yes" going call emergency services anyway? I don't want that. But if I lie and answer "no", is it going to somehow modify its detection criteria and not work the next time I fall and really do need help? I don't know, I have no idea what it will do.

Besides, it's all a moot point because it's the wrong question to ask. It doesn't really matter if you actually fell or not. It detected what it "thinks" was likely a fall and the real question to ask is: do you need help?

If you turn the corner and see someone laying on the ground, do you ask them if they they just fell, or do you ask them if they need help?

Reply on news.ycombinator.com