Hacker News Reader

Comment by 23david | Hacker News Reader

I'm surprised that the google engineers and managers involved in the project were so stupid/arrogant/lazy that they deployed a program that seemed to just vacuum up payload data for later analysis instead of processing and sanitizing the data as it was collected. Maybe google needs to tweak their HR algorithms to focus on hiring people with a bit more common sense. It's nice to have a powerful legal department that can help get your engineers and company out of trouble when they really screw up, but it's still a PR disaster... Google has a big problem now that on a corporate level they seem unable to do the right thing when they have clearly screwed up. Just come clean about what really happened, sincerely apologize, and then take serious steps to ensure that it doesn't happen again.

View on HN

Why does "street view" need to "collect data", much less sanitize it?

How could this possibly be a error of judgment? Take pictures, fine. Snoop on people's WIFI? What possible purpose could that serve in context of Google Maps?

> Just come clean about what really happened, sincerely apologize, and then take serious steps to ensure that it doesn't happen again.

I do not understand this attitude. They were spying on people. It is that simple.

Ever heard of Skyhook [1]? Not that I know what it was used for, but generating statistics on the types/coverage of wireless networks could prove useful in the future.

[1] http://en.wikipedia.org/wiki/Skyhook_Wireless

Because they want to triangulate all the wifi access points. This helps you to fast and with low power determine your position using wifi on your android phone. This is valuable data and I think both google and apple used a third party for this before but have since build up their own database of wifi networks.

The payload is used to help and triangulate the accesspoints more precisely (according to the article).

Its more than just Android/iOS- their location data also works for computers with wifi as well (eg when I press the "share my location" button on Google Maps from my MBP, they pin my location down to within 50 feet)

They certainly did not need to keep full packet contents though, just IP/Macs of nearby address to map to an approximate lat/long

I don't understand how the actual data being streamed through the WIFI could be useful for anything other than spying. How does knowing what's in the packets change anything about how they use it for location?

It's easier to just collect every packet that passes through your antenna and batch-process it later than to sanitize it in real-time - there's handy off-the-shelf software to do the former.

That's probably exactly what happened. Every time the privacy chicken littles start up, remind them that Google has precisely no use for the random data passing through each hotspot. None. Nada. It's completely useless to their organization. Junk data to be deleted later.

Then why store the data? And why would Google refuse to release the data to regulators because "it might break privacy and wiretapping laws if it shared the material?"

Because it's automatically stored? Google is still comprised of humans who make mistakes, sometimes horrible ones.

Because they were in a pretty bad position once they found out what was happening? I can see an "Oh sh-, we broke the law" moment happening at the highest levels of the organization.

The answer to this specific question is two replies directly above your comment...

> It's completely useless to their organization. Junk data to be deleted later.

At least one person at Google thought it could be potentially useful:

> The report, quoting the engineer’s original proposal, gives a somewhat different impression. The data, the engineer wrote, would “be analyzed offline for use in other initiatives.” Google says this was never done.

I think Google is guilty of arrogance wrt privacy, rather than a fiendish plan to spy on people.

Let's do it the "right" way. You drive around and collect ssid for each network. Then you discover half of them are "linksys". Oops, should have collected mac address instead. So now you have to drive the trucks around all over again. Or, if you'd saved complete packets, you just analyze them instead. You never know what you're missing until you're missing it.

You're right that collecting these fragments is useless for the purpose of street view or Google maps. But it's also completely useless as a method of spying on people's personal affairs.

You can't build personal profiles based on that data, but that is what Google does for a living. They have an incredibly dense set of personal information from gmail, adsense, doubleclick, search, etc. What's the additional information they could possibly gain from arbitrary street view payload fragments?

I'm usually very concerned about the impact of Google's information collection on my privacy, but in this case I just don't see how it fits in with any kind of goal they might have.

My diagnosis would be one of institutionalized compulsive hoarding. Maybe Googlers just don't have it in their genes to let go of any information they got, regardless of how useless it may be. Or maybe we need a new term: BBDR - Blind Big Data Rage :-)

My feeling is that your diagnosis is spot-on. But I don't know enough about organizational psychology to know whether such a thing is really possible.

I suppose that logically if one of the google founders had a similar approach to dealing with privacy and data collection, naturally they would hire and attract people with similar attitudes.

I've seen from google employees a general sense of emotional detachment and even contempt of their own customers that I haven't seen in other successful tech companies... it's as if they're not really dealing with fellow human beings.

It's no wonder that they haven't been successful with creating meaningful social services. On an institutional level, they just don't get or even respect how normal people think. Too bad for them...

It's not as stupid as you think. If you don't know what you are going to do with the data, particularly with low level network packet data like this, it is tricky to sanitize it in a way that it will be useful. That actually requires a ton of thought, and it is way simpler collect the data, keep it isolated, and then come up with ways of accessing/collecting it that protect against privacy violations.

Here's the thing: almost all of the privacy violations that people actually reference in the articles about this issue are data (e-mails, login passwords, URL's, etc.) that are already being transmitted to ISP's (where ironically, there has been a lot of discussion about them being required to archive this data). ISP's have a far less fragmented and transient view of the data than a Google Street Car, and they know precisely where their customers live. The only possible privacy violation here is with data exchanged between systems within the LAN of the home, which is a very different kind of information, is generally not that useful when viewed as a few isolated packets, and which requires a degree of technical sophistication such that you'd really think the same people doing it would also know to encrypt their wireless networks, even if only with something as lame as WEP.

"I'm surprised"

When it comes to Google collecting information which would surprise the average person and would cause concern for the citizenry were there government do so, one shouldn't be. This is simply another example of what Google will do in the interest of "developing new services."

Reply on news.ycombinator.com