I gather (from other comments in this thread) that the O/S did some of the command/parameter parsing for you, so I'd be inclined to argue that it was part of the O/S.
Now that I think about it, I think you could register your programs with the help system, and help for your program would automatically get included if someone just typed "help".
Yup. The important inner nervy bits (technical term) weren't so much stretchy like elastic, more stretchy like a spring.
TFA - "What the whale has developed is a beautiful system to protect its nerves, to package it in a way, probably in a corkscrew or undulated way, surrounded by elastin fibres"
So we could avoid nerve damage by preemptively replacing our at-risk nerves with stretchy whale nerves, in much the same way we could avoid bone breakage by covering our skeleton with adamantium... or some non-fictional alternative.
I'm basing it on how the similar Cayla doll works. Doll -> Tablet/Phone -> House WiFi -> Internet. Also in TFA it says the parents enable it by "signing into an app". That's most likely something they need to do each time the doll's turned on. Most of the local processing power is likely to be in the App.
It also makes sense from a maintenance point of view. It's a pain to update the doll's firmware, but easy to update mobile apps and server-side software. I doubt if there's any processing going on in the doll at all.
As for security, it's as secure as any other Android app I allow on my phone with "full internet access".
Yes, the professional liars could be lying about "Its listening function is activated only when a button on Barbie's belt buckle is pressed." but battery life will suck if they do it any other way.
Mattel didn't call it the "world's first interactive doll", they called it the "very first fashion doll that has continuous learning". The key words here being "fashion doll", which allows them to differentiate themselves from competitors they don't consider fashionable, and "continuous learning", i.e. the server is keeping a log of previous conversations, and using them to have better conversations.
The Cayla doll for instance was brilliant at answering questions like "What is Polytetrafluoroethylene?", and rubbish at answering questions like "Do you like kittens?" Sadly, it was a rather dull toy for a 7yr old girl, despite looking initially exciting.
What it didn't do, and this Barbie won't be doing, is snooping on us. I'd expect it to use Bluetooth to communicate with an App, which will then be using WiFi to connect to the Mattel server. If you don't have the App turned on, and by the sounds of it, if you don't press the button on the doll, it won't be listening. So if a child is playing with it, don't discuss national secrets, and if a child is not playing with it, there's no snooping.
The online conversation history is interesting, but not worth getting worked up about. It's recording how you've played with it, like all sorts of other online games keep logs.
Sure, there's a potential for abuse - "Hey - you say you like ponies! Did you know Mattel make a Barbie pony that's only $9.99 from your local Wal-Mart?" but I think if they start doing that, the backlash of people not buying their products out of sheer disgust will stop them pretty quickly. I'd say the worst is probably more like "Hey - you say you like ponies! I've got a pony called Muffin" without mentioning that Muffin costs $9.99 from Wal-Mart (or might do if it existed).
Can they say:
Does A cause B? Probably not.
Does B cause A? Probably not.
So there's probably a C causing A and B.
There's a lot of probablys in that.
To drift even more off-topic, I used to work on an "Expert System" where BOOLEAN variables had one of FOUR possible values:
It's not quite as daft as it sounds - It populated variables by asking users a question, to which they could answer yes/no/don't know. The fourth value was the case where it hadn't asked the user yet.
One small step for man, one giant stumble for mankind.