News for hackers. Stuff that matters.
Re:Carrier IQ Speaks Out: We ignore extraneous user Data by salimane on Thursday December 01, 2011 @09:17PM (#3302378)
you ignore extraneous user data?!??! what!?!?! why do you get it in the first place if you're only interested in a subset? bunch of lying rootkit developers
Re:Carrier IQ Speaks Out: We ignore extraneous user Data by ghshephard on Thursday December 01, 2011 @09:43PM (#3302443)
From the article:

"If during a support call a technician asks a customer to enter a short code, CIQ will be listening for it; when it’s entered, CIQ will relay the appropriate diagnostic information to the carrier. Any keystrokes beyond that are ignored."

Their explanation then, for monitoring keystrokes, is that they are waiting for a diagnostic code to be entered by the user. Presumably this can be done from anywhere within the phone, on any platform, without regards to running applications.

Re:Carrier IQ Speaks Out: We ignore extraneous user Data by latch on Thursday December 01, 2011 @09:46PM (#3302449)
How do you capture a subset of data without capturing it all (and discarding what doesn't match)? If you are doing a find in a document, don't you scan through the entire document and discard whatever isn't what you want? It sounds like that's the explanation they are trying to give.
Re:Carrier IQ Speaks Out: We ignore extraneous user Data by salimane on Thursday December 01, 2011 @10:28PM (#3302554)
the thing is, given the context, you can do this without looking at everything the user is doing, you can basically wait for the trigger (error, call drop...) to happen before trying to submit error reports. that's the way it's done in software applications
Re:Carrier IQ Speaks Out: We ignore extraneous user Data by latch on Thursday December 01, 2011 @11:56PM (#3302760)
Just to be clear, I'm not saying they are right, I'm just trying to look at this objectively. If you read what they said, I don't think your solution works:

"During a support call a technician asks a customer to enter a short code, CIQ will be listening for it"

Without knowing a ton more, maybe there's a _much_ better way to do this..like having the user explicitly launch a program rather than having something always running and always listening. No argument there. But if you're thinking they should abstract away some "code_entered" event, ultimately something lower-level is going to be listening to each keypress and looking at sequences to raise those events.

Some of this stuff happens at a lower level than a lot of us might be use to programming nowadays. I'm not sure that there's necessarily a pretty error event they can hook into. Something about all of this reminds me of using stuf like Spy++ on Windows and you see just how much raw data is available at the lowest levels

Re:Carrier IQ Speaks Out: We ignore extraneous user Data by derefr on Thursday December 01, 2011 @09:41PM (#3302436)
A topical metaphor: sometimes, to get a toy, you have to buy the whole happy meal.
Re:Carrier IQ Speaks Out: We ignore extraneous user Data by hboon on Thursday December 01, 2011 @10:44PM (#3302590)
But here, you are forced to eat it too.