All posts tagged ‘Security’

File Under: Glossary

PGP


Pretty Good Privacy is a flavor of algorithmic encryption that uses two cipher keys, one public and one private. Anyone can use a public key to send a scrambled message to the receiving party. The private key is then used only by the receiving party to unscramble incoming messages. The two-key system was developed by RSA Data Security, Inc. and PGP is the most popular type of two-key encryption available for public, non-commercial use.

File Under: Glossary

Cryptography

Cryptography is a constantly changing and evolving field of mathematics that on the internet refers to the practice of encrypting data for safe transmission.

Cryptology is the basis for many types of secure transmission over the internet. Regular data is coded into a cipher (which looks like scrambled text) then transmitted and deciphered by the receiving party.

File Under: Identity, Security

EFF Reveals How Your Digital Fingerprint Makes You Easy to Track

Think that turning off cookies and turning on private browsing makes you invisible on the web? Think again.

The Electronic Frontier Foundation (EFF) has launched a new web app dubbed Panopticlick that reveals just how scarily easy it is to identify you out of millions of web users.

The problem is your digital fingerprint. Whenever you visit a site, your browser and any plug-ins you have installed can leak data. Some of it isn’t very personal, like your user agent string. Some of it is more personally revealing, like which fonts you have installed. But the what if you put it all together? Would the results make you identifiable?

As the EFF says, “this information can create a kind of fingerprint — a signature that could be used to identify you and your computer.”

The EFF’s test suite highlights what most of us probably already suspect — we’re readily identifiable on the web. We ran the test on a Mac using Firefox, Safari and Google Chrome, all of which leaked enough data to make us identifiable according the EFF’s privacy explanations.

The purpose of Panopticlick is to show you how much you have in common with other browsers. The more your configuration mirrors everyone else’s, the harder it would be to identify you. The irony is, the nerdier you are — using a unique OS, a less common browser, customizing your browser with plug-ins and other power-user habits — the more identifiable you are.

For example, say you’re running Firefox on Ubuntu with the Gnash plug-in instead of Flash — way to stick it to the man — but you’re also showing up with a unique configuration of browser, OS, installed fonts, plug-ins and more which can be combined to identify you via a unique online fingerprint.

So what can you do to make yourself less identifiable? Well, by disabling cookies, the Flash plug-in, the Java plug-in and most of our extensions we were able to blend in better. Actually, the fact that we didn’t have Java or Flash turned on made us more identifiable in those categories, but it also denied the test access to our installed fonts and other bits of data, so overall, less identifiable.

Obviously that approach has a downside — without Flash there’s not much in the way of online video, a lack of cookies will cause issues with logins, and without Java, you won’t be able to crash your browser or cause it to get hung up for hours.

In short, the disabling method isn’t much fun. Strange though it may seem, the best way to lose the unique online fingerprint is to blend in with the herd. As the EFF points out, mobile browsers are hardest to identify since there are few customization options and, for the most part, one version of Mobile Safari looks just like another.

By the same token, if you want to blend in, stick with stock system fonts, run Windows XP, use Firefox with no add-ons and turn off cookies. You’ll be much harder to identify.

We should point out that, no matter how well you blend in the fingerprint test, you are of course still identifiable by your ISP. Advertisers and websites generally can’t access the information your ISP has on you, but of course governments — with the cooperation of your ISP — always can. So don’t think just because you’ve eliminated your fingerprints no one knows who you are.

Front door photo: Brian Lane Winfield Moore/Flickr (CC)

See Also:

File Under: Identity, Security, UI/UX

Warning: This Site May Be Sharing Your Data

Aza Raskin, head of user experience at Mozilla, is leading a charge to make privacy settings more explicit to users by creating visual cues in the browser. Raskin’s idea uses a set of small icons to denote the limits of a website’s privacy policy.

Raskin likens the idea to how Firefox (and other browsers) currently handle phishing attack warnings, using visual icons and simple language.

For the active social web user, keeping track of which bits of your data are public and which are private on different sites is a chore. Some websites share your photos, status updates, your list of friends, who you’re following and other data on the open web by default. Some share nothing. The rest are somewhere in the middle.

Part of the problem is the privacy policies themselves. They are complex, mind-numbingly long legal documents. We routinely ignore them, breezing past them by clicking “I agree.” Dangerous behavior, indeed.

Raskin and his supporters have borrowed some ideas from the way Creative Commons licensing works, and the way licensing options are denoted on content sites. Originally, the idea was to create a Creative Commons model for privacy policies — that is, a common, readable, reusable set of policies much like the Creative Commons licenses for content — but that plan was abandoned because policies differ too much from site to site. There’s no easy boilerplate for privacy like there is for content publishing.

But the icon concept remains: A website creates a privacy policy and chooses from a limited set of standard icons that reflect the written policy. Is your profile public by default? Your photos, or status messages? Each setting has its own icon, and the group of settings are indicated by a short stack of icons. The icon set is then detected by the browser and displayed to the user. If there are no icons chosen, the browser offers a warning along the lines of its phishing warning, something like: Be careful, this site might be giving away or selling your data.

Raskin is very clear that, so far, this is a work in progress. There are, as of yet, no icons designed, and the details of how they would be implemented remain vague. Nor has Mozilla made any official announcement that it would support such a system.

However, recent events have proven there’s clearly a need for a standardized, front-and-center privacy notification system. In December, Facebook began a shift towards looser default privacy settings that encourage users to share more of their data. Just last week, Facebook CEO Mark Zuckerberg, in an interview with TechCrunch’s Mike Arrington, noted that people’s notions of privacy on the social web evolve often, and that social web sites will have to continually update their own privacy policies to reflect those changes. As a result, Facebook’s new defaults will offer less privacy. Zuckerberg’s words set off a fierce debate on the topic, with Marshall Kirkpatrick of ReadWriteWeb presenting the clearest counterargument that changing social mores should not lead to looser default privacy settings on the social web.

We’ve often said the browser is the most logical place to display identity and privacy information to the user. As people surf from site to site, they should be able to see, at a glance, what level of privacy they’re currently working with. Raskin’s model sounds like a pretty good plan, though implementing it might be a bit more difficult.

One obvious problem: What’s to stop a site from using icons that are totally different than what the written policy actually says? Raskin and crew want the icons to supersede the written policy so, in that scenario, the written policy is trumped by the icons and the user retains their rights. Whether or not an icon can legally trump a written document is something Raskin doesn’t directly address, and, as one commenter points out, the situation gets much more complex when you start considering international legal systems.

If you’ve got ideas or would like to participate in the discussion, head over to Raskin’s blog or sign up for the upcoming privacy workshop hosted at Mozilla on Jan. 27 (see Aza’s post for full details).

See Also:

File Under: Databases, Frameworks

Django 1.2 Alpha Offers Multiple Database Support, Improved Security Features

Django, the popular web development framework written in Python, has released the first alpha for its much-anticipated new version, Django 1.2.

Among the new features coming in Django 1.2 are support for multiple databases — a key feature for larger websites running Django — improved security features and a messaging framework that works much like Ruby on Rail’s “flash” feature.

The multiple database support will likely be the most important part of the next version of Django since it will allow for much easier application scaling. Django 1.2 makes it easy to target individual databases within your apps using some new queryset methods which make it easy to read and write to specific databases.

The security features include much-improved protection against Cross-Site Request Forgery (CSRF) attacks. For more details on how the CSRF protection works, have a look at the new CSRF documentation page.

If you’d like to test out Django 1.2, or see how your apps run on the new release, head over to the downloads page or update your Subversion checkout. Keep in mind though that this is still an alpha release and should not be used on production sites. The final release of Django 1.2 is scheduled to arrive in March 2010.

See Also: