LESSWRONG
LW

Comment Permalink

DanielVarga14y00

I don't describe it as "not completely trivial" because I think it's some sort of laudable standard, but because you are implying that it is a completely trivial task.

It seem like our difference is about whether to give a worst-case security analysis or a best-case security analysis. :) It is a completely trivial task for a high percentage of potential targets. Some people aren't geolocated or are incorrectly geolocated. Some live in large cities. But a long tail of lesswrong users is almost completely identified by their geolocation.

Using information which in all likelihood is known to no-one else on this site but you, you've found a reliable method of uniquely identifying yourself.

No. The only information I used was that I am from Hungary. Nothing else. This fact about me is public here. (Obviously, nobody cares about it, but that does not make it a secret. Again, I am talking about worst-case analysis.)

I am really not monomaniac about privacy. I agree with you that it is not a big deal that somebody can be followed like that. But at least let's realize that lesswrong is unusual in this regard, and unusual in a bad way.

29

Isn't this sitemeter logging a bit too excessive?

2nd Feb 2011

1 min read

29

I have just realized that sitemeter has the following data published about my visit, in a searchable and browsable format:

Domain Name

broadband.hu ^? (Hungary)

IP Address

80.98.73.# (UPC Magyarorszag Kft.)

ISP

UPC Magyarorszag Kft.

Location

Continent	:	Europe
Country	:	Hungary (Facts)
State/Region	:	Fejer
City	:	Etyek
Lat/Long	:	47.45, 18.75 (Map)

Language

English (U.S.)
en-us

Operating System

Macintosh MacOSX

Browser

Firefox
Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.2.13) Gecko/20101203 Firefox/3.6.13 GTB7.1

Javascript

version 1.5

Monitor

Resolution	:	1296 x 810
Color Depth	:	24 bits

Time of Visit

Feb 2 2011 12:34:19 pm

Last Page View

Feb 2 2011 12:34:40 pm

Visit Length

21 seconds

Page Views

2

Referring URL

http://lesswrong.com..._meetup_is_easy%252F

Search Engine

lesswrong.com

Search Words

hungary

Visit Entry Page

http://lesswrong.com/r/discussion/

Visit Exit Page

http://lesswrong.com...a_lw_meetup_is_easy/

Out Click

Time Zone

Visitor's Time

Feb 2 2011 9:34:19 pm

Visit Number

3,497,452

I am not a privacy geek, but isn't this a bit too extensive? By the way, I am not from Etyek, Hungary, I am from Budapest, Hungary. Etyek is a very small village, so if sitemeter consistently identifies me as someone from Etyek, then it will be even easier to track my lesswrong browsing habits. It is very easy even without that.

Privacy / Confidentiality / SecrecySite Meta

29

New Comment

19 comments, sorted by

Click to highlight new comments since: Today at 6:23 AM

[-]wedrifid14y60

I have just realized that sitemeter has the following data published about my visit, in a searchable and browsable format:

While it doesn't bother me personally I can certainly see why people could object to that much information being published publicly.

[-]NihilCredo14y60

RequestPolicy is a Firefox addon that blocks (and whitelists with a click) all cross-site requests, such as sitemeter.com's or google-analytics.com's, or for that matter any funny guy who embeds terrorismandpaedophilia.com in the middle of poniesandkittens.com

If you are bothered by this sort of stuff, install it. Note, however, that it makes it quite cumbersome to visit multimedia-heavy websites for the first time, especially in conjunction with the near-mandatory NoScript.

[-][anonymous]14y30

I use Ghostery. It's less radical than NoScript or RequestPolicy, but it knocks out most tracking scripts and ad networks, including SiteMeter, and it's more pleasant to use.

[-]nick01200014y00

Does this offer any functionality NoScript doesn't? I've already got the latter installed, but I'd want to know if it would be a waste of time to install this as well.

[-]NihilCredo14y30

https://www.requestpolicy.com/faq#faq-noscript

Is RequestPolicy an alternative or competitor to NoScript?

No! :)

NoScript is a tool that gives you a default deny policy for JavaScript, Java, Flash and other plugins. NoScript allows you to whitelist scripts and objects from domains you trust.

RequestPolicy is a tool that gives you a default deny policy for cross-site requests. RequestPolicy allows you to whitelist cross-site requests you trust.

How does RequestPolicy help you where NoScript does not? RequestPolicy will protect you from various attacks that NoScript will not (such as CSRF attacks, though there some special cases that NoScript protects against) and will give you greater privacy while browsing.

Also, RequestPolicy will give you finer-grained control over JavaScript and plugins when you use it with NoScript. For example, if you whitelist a domain with NoScript to allow it to run JavaScript, then that domain will also be allowed to run JavaScript when you are on any other site that you have whitelisted with NoScript. RequestPolicy makes sure that when it is JavaScript from a third-party site, it will still be restricted unless you have allowed those cross-site requests.

Conversely, NoScript gives you protection that RequestPolicy does not. RequestPolicy will not keep you safe from malicious JavaScript or vulnerable plugins on the current site you are visiting, So, NoScript is absolutely essential for browser security.

Having two separate tools that each do their specific jobs well is the best approach. NoScript is an amazing extension and is absolutely essential (like RequestPolicy) to using Firefox securely. It is best to use both RequestPolicy and NoScript.

[-]luminosity14y30

Just about any site you visit nowadays will be recording a substantial set of data on your visit, even if it just records the pages you visited, your IP and your browser's id string. If it's a problem I suggest masking your IP / changing how your browser identifies itself, rather than changing one specific site.

(Personally I find it interesting, if not demonstrably useful to know where less wrong readers are coming from. Of course you could view this using only aggregate stats.)

[-]DanielVarga14y90

I think I wasn't clear enough. I maintain several webservices. I perfectly understand that lesswrong collecting such data about me is normal. But I am not happy that sitemeter publishes the whole stuff in a non-aggregated form, for everyone to see in real time. Do you suggest that this is usual, too?

[-]Quirinus_Quirrell14y20

I have just realized that sitemeter has the following data published about my visit, in a searchable and browsable format:

Searchable my behind! I looked into what it would take to use this to, for example, unmask Clippy, and it was less usable than the marginal next-best strategy.

[-]DanielVarga14y20

(JoshuaZ) I'm also curious why you would be interested in promoting the unmasking of users.

I think you are more interested in avoiding the unmasking.

Anyway, you are right that sitemeter is not very convenient for this task, but the data is there. And unmasking is not the only possible application. Right now, it is publishing the fact that I looked at Quirrell's and JoshuaZ's user profile. And I might have just figured out Eliezer's current IP address. (Okay, maybe it was somebody else who tried to visit Eliezer's password-protected drafts page. Wait, is Eliezer writing a post?)

[-]JoshuaZ14y10

Searchable my behind! I looked into what it would take to use this to, for example, unmask Clippy, and it was less usable than the marginal next-best strategy.

I'm curious what the marginal next best strategy is.

I'm also curious why you would be interested in promoting the unmasking of users.

[-]Quirinus_Quirrell14y60

I'm curious what the marginal next best strategy is. I'm also curious why you would be interested in promoting the unmasking of users.

Not all users, just the few I happen to be curious about. And no, I won't say anything more about what the marginal next-best strategy is other than that I'm immune to it too, and -1 Quirrell point for asking.

[-]JoshuaZ14y10

This is a general problem with sitemeter. There's no way (as I understand it) to have your sitemeter stats be openly viewable without all the details viewable also. It would be nice if it had middle setting but I'm not aware of any.

I found myself, but only by switching browser and hitting a obscure page.

Sitemeter thinks I'm in Poland. Which I suppose is on the right continent.

[-]sixes_and_sevens14y-20

It took me over five minutes to identify myself on there, and I already know my own IP address.

[-]DanielVarga14y00

What is this supposed to prove? It took me one minute to identify myself, without checking my IP address.

[-]sixes_and_sevens14y00

It's not supposed to prove anything, it's just observing that identifying individuals using this information isn't a completely trivial task.

[-]DanielVarga14y10

It is not always completely trivial to identify individual visitors? That is a very low standard.

[-]sixes_and_sevens14y00

I don't describe it as "not completely trivial" because I think it's some sort of laudable standard, but because you are implying that it is a completely trivial task.

Using information which in all likelihood is known to no-one else on this site but you, you've found a reliable method of uniquely identifying yourself. If you hadn't published it, the amount of effort required to identify you would probably be enough of an obstacle to deter anyone with a sudden hankering to investigate your recent LW browsing history from doing so.

My original comment was observing that I'm me, I know (and probably care) more about me than anyone else in the world, including specific technical information salient for tracking myself in a web activity logging context, and it was still far from immediately apparent as to who I was. If someone else knew (or cared) enough about me to try and do the same, they'd probably have better methods available to them.

[-]DanielVarga14y00

I don't describe it as "not completely trivial" because I think it's some sort of laudable standard, but because you are implying that it is a completely trivial task.

It seem like our difference is about whether to give a worst-case security analysis or a best-case security analysis. :) It is a completely trivial task for a high percentage of potential targets. Some people aren't geolocated or are incorrectly geolocated. Some live in large cities. But a long tail of lesswrong users is almost completely identified by their geolocation.

Using information which in all likelihood is known to no-one else on this site but you, you've found a reliable method of uniquely identifying yourself.

No. The only information I used was that I am from Hungary. Nothing else. This fact about me is public here. (Obviously, nobody cares about it, but that does not make it a secret. Again, I am talking about worst-case analysis.)

I am really not monomaniac about privacy. I agree with you that it is not a big deal that somebody can be followed like that. But at least let's realize that lesswrong is unusual in this regard, and unusual in a bad way.

More from DanielVarga

Curated and popular this week