The Web Won't Be Safe Or Secure Until We Break It
CowboyRobot writes "Jeremiah Grossman of Whitehat Security has an article at the ACM in which he outlines the current state of browser security, specifically drive-by downloads. 'These attacks are primarily written with HTML, CSS, and JavaScript, so they are not identifiable as malware by antivirus software in the classic sense. They take advantage of the flawed way in which the Internet was designed to work.' Grossman's proposed solution is to make the desktop browser more like its mobile cousins. 'By adopting a similar application model on the desktop using custom-configured Web browsers (let's call them DesktopApps), we could address the Internet's inherent security flaws. These DesktopApps could be branded appropriately and designed to launch automatically to Bank of America's or Facebook's Web site, for example, and go no further. Like their mobile application cousins, these DesktopApps would not present an URL bar or anything else making them look like the Web browsers they are on the surface, and of course they would be isolated from one another.'"
Broke it. Does that mean it's safe now? http://www.google.com/404
(let's call them DesktopApps)
Let's not.
Everything is better with chainsaws.
So we would have no clue as to where we were taken?
Yeah, that must be good security
So then I'd end up with about 100 "Apps" on my desktop, which all might or might not behave a bit differently, and every time I want to switch to another site, I have to switch the app? How would I follow links outside of the app? Would there still be a way to find websites/desktopapps? If so, what makes sure that those aren't malware?
Yeah. Because nobody would ever hack/write a virus for the BofA DesktopApp that would collect login credentials, etc.
Most of what we want on the web is text and static images. Tables are nice. Maybe you need a handful of tags. Let the browser handle layout. That would be much easier to secure than the dynamic fustercluck we have now. There are probably more APIs than there were tags in 1999. There are probably hundreds of functions in your browser that expose security flaws. We could dump all of them and they wouldn't be missed.
Slashdot needs a handful of tags and good old CGI. That's all.
For all intensive purposes, "whom" is no longer a word. That begs the question, "who cares"?
Frankly, I'll take the current internet with all its warts and diseases over some centralized, walled-garden approach that will STILL suffer from the same things, just in a different mechanic. The bottom line is how you decide what to trust in any system.
I'd submit that the problem isn't that the internet is the Wild Wild West, it's that it is the Wild Wild West without any sheriffs or cowboys. No, I'm not talking about regulation of the internet; I'm talking about people who break laws (fraud, theft, etc.) being found and prosecuted regardless of what tool (postal system, telephones or internet) they used to do it.
Did he just re-invent client-server desktop apps?
outlining why, everyone else is covering it pretty well, but this is an incredibly awful idea. And its originator is an idiot as is he who decided this was worthy of posting to /.
The same holds for these apps. Same difference.
I want the wild wild web, where the deer and the antelope roam, and the skies are (not cloudy) all day, not some locked-down police-state prison-cell silo-world of commercial money-sucking, mind-***king apps.
Where are we going and why are we in a handbasket?
Dan Walsh, one of the principal developers of SELinux has blogged about a way to do this on your linux desktop box. You can start a "virgin" browser in it's own Xserver with optional presets you copy in the loopmounted container. Every time you run it, it starts the same fresh image built on the fly when you run the command. This makes it easy to have separate browsers for each task you want isolated from the rest of your web experience or your desktop computer. Even if it gets infected, it will not remain on your computer and the infection is gone as soon as you close the browser. He's not the only one that has written about it, there are many more people giving useful examples on the web.
I was promised a flying car. Where is my flying car?
The idea is just completely tangential to what the problem is. The problem isn't that "If we just had a secure little app that could ONLY go to my Bank, everything would be OK". The problem is that the internet is a series of interconnected sites, many of which you discover without even realizing what the site is, compounded by the fact that browsers aren't secure. We all know once the machine is infected from visiting a compromised site, all bets are off.
Drive bys happen because the browser isn't secure, not because people are supposed to have some inherent understanding of what sites are "good" and what sites are "bad". I've worked security in multiple different capacities, and even I can't tell you if a site is going to be "safe" or not. That's because a lot of drivebys are from the 3rd party adware server getting infected. Despite what some totally uninformed IT professionals will tell you, you can't protect yourself by just "knowing where not to click" or "knowing not to click on the fake anti-virus thing". Sadly, I know IT professionals that absolutely SWEAR that this is how people get malware, despite me repeatedly providing them examples of how that's just not that case.
AccountKiller
My thoughts exactly. So - my google search app wouldn't point me to web pages - instead it would point me to apps I could download and install for each different web page. So now I'm installing thousands of web apps? THAT sounds like a security nightmare! Who is going to watch over the security of the apps? Google? They are already having problems with the Android apps.