So, Google has made it very hard to install Chrome extensions outside of the Chrome Web Store – out of security concerns. In addition, they sprung this on users and extension developers without much consultation or consideration for their concerns. As always – understandable to protect users, but the handling has an almost Apple-like bluntness to it. Next up: how to jailbreak your browser?
The issue deserves more than sensationalist phrases like “apple-like” and “jailbreak your browser”. I expect much more from osnews, to be honest.
I just signed up to to post this message, this piece of news is that horrible.
Didn’t see a submission from you…?
I’m not a robot, I can’t pump out an extensive item on command. I just don’t care about browser extensions enough to write 6 paragraphs about it, especially not after a long workday – sorry.
Install Chromium? Does Chromium also require web store extensions by default?
Not much point, the only difference between Firefox and Chrome now is a few -webkit- properties and NativeClient. And since Chromium doesn’t support NaCl, most of the ‘chrome web store’ content doesn’t actually work.
Where do you see that Chromium doesn’t support NaCl? I’m running Chromium on ArchLinux, and I’m pretty sure it works. Perhaps you’re referring to it not working on Debian/Ubuntu builds of Chromium?
http://askubuntu.com/questions/91789/why-is-nacl-disabled-for-chrom…
According to the green-highlighted comment added in this commit…
https://chromiumcodereview.appspot.com/10511015/diff/1/chrome/common…
You can re-enable side-loading by starting Chrome/Chromium with the “–enable-easy-off-store-extension-install” switch, and then drag-n-drop the *.crx file onto “chrome://extensions/”.
I’d say this is a nice compromise. It still allows user freedom, while not opening up an easily-exploitable area for users to be duped by. This is pretty comparable to Android, which offers a checkbox to enable side-loading. This is no where near the Apple/iOS defective-by-design model, in which you have to exploit a bug in the OS in order to side-load.
jasutton,
Yea, it’s better than apple’s control. But google is weaving precariously on the line between open and closed. I am extremely disappointed they’re using a “security” excuse to justify new restrictions, which users won’t know how to bypass. A better approach would have been to give users better tools to view/limit what extensions can do and just set the defaults to restrictive.
Speaking from tech support experience, I’d say if a user doesn’t know enough to google for that switch, they have no business side-loading. The more checkboxes you give users, the more they will check out of annoyance just to avoid the alert dialogs, and then your security becomes null and void. I’d agree that having this switch is a nice compromise, and it’s not as though you have to hack your browser to enable this.
darknexus,
“Speaking from tech support experience, I’d say if a user doesn’t know enough to google for that switch, they have no business side-loading. The more checkboxes you give users, the more they will check out of annoyance just to avoid the alert dialogs, and then your security becomes null and void.”
The spread of malware happens because users lack the tools to make informed decisions. Often the choice is between “run” and “do not run” and the only information presented is to identity the software. Even knowledgeable users will be at a complete loss to know if something is harmful, so I fully agree that this type of security model is flawed. But I disagree very strongly with the “remedy” of a walled garden (even if more savvy users can disable it). It’d be both more open and more secure to add metadata about what the extension does and then enforce it in a sandbox. Given the right tools & information, users may be even more secure than simply trusting everything in google’s repository.
I call bs. Malware spreads because users treat their computer as a magic box. They expect their computer to protect them, to do their common sense thinking for them, and to be the always-on tool. They don’t wish to make informed decisions. I speak from experience with a large number of users, some of which have actually said this to me. I can’t count the number of times I’ve heard “well, if that’s not a safe program, I shouldn’t be allowed to install it.” I shit you not, I have heard these words.
If you’re a carpenter, you don’t expect your tools to maintain themselves. You don’t expect your vehicle to keep itself going without maintenance, nor do you expect it to survive in tact if you drive it straight into a tree. Yet, people expect their electronics to magically just work no matter what sort of crap they put on them. They want that? Fine, but that comes at a heavy cost. They don’t want to think, so fine, we don’t make them think. However, we have to allow those who still wish to think and employ their common sense to not be limited by the idiocy of those who do not wish to use their brains. Therefore, a compromise is needed.
I wish these sorts of things weren’t necessary. I wish people would use their damn brains for something other than watching crap tv and window shopping. The sad reality is, however, that the greater part of the market decides these issues and we could easily find ourselves buried by them. Allowing us to enable this ability in an easy way that is, at the same time, not obvious to those who don’t have the brains to search for it anyway seems to be the best way to keep both groups happy.
And before anyone mentions it: Yes, I know how elitist and arrogant I sound. That’s what happens when you see the same mistakes repeated over and over and over again, and every time they ask: “Why didn’t my computer protect me?” I don’t know what the situation is in other countries, but in the USA that’s what I get 95% of the time. I’m almost glad for lockdowns imposed on these types of people, if only so they stop bothering me with the same crap and stop spreading their malware. As long as those of us who do know our stuff can legally and uncomplicatedly bypass said lockdowns, I have no problem with it whatsoever, as that approach keeps both groups happy.
darknexus,
“I call bs.”
There’s really no need for sarcasm. Your opinion is that it’s ok to submit users to third party control for safety’s sake, which is fair enough. I hope you are least aware that such philosophies, especially when taken collectively, tend to erode our freedoms over time.
“If you’re a carpenter, you don’t expect your tools to maintain themselves. You don’t expect your vehicle to keep itself going without maintenance, nor do you expect it to survive in tact if you drive it straight into a tree. Yet, people expect their electronics to magically just work no matter what sort of crap they put on them.”
You are speaking metaphorically about how physical tools relate to software. I don’t like using metaphors since comparing different things as though they are the same is inherently flawed as details are worked in. But to be more complete the metaphor must account for how end user restrictions affect software. For example, your tools would need to refuse to work with unauthorised components that are never the less compatible. Artificially restricting tools would generally be considered a bad thing, even if the freedom to use the tools the wrong way may damage them.
“And before anyone mentions it: Yes, I know how elitist and arrogant I sound. That’s what happens when you see the same mistakes repeated over and over and over again, and every time they ask: ‘Why didn’t my computer protect me?'”
To which I say, the goal should be addressing the lack of software sandboxing rather than having users acquire all their software from centralised sources.
“As long as those of us who do know our stuff can legally and uncomplicatedly bypass said lockdowns, I have no problem with it whatsoever, as that approach keeps both groups happy.”
But you’ve completely overlooked that the walled garden approach (whether it can be disabled or not) doesn’t directly solve any security problems on it’s own. For that you need additional vetting, otherwise there’s nothing in place to stop covert distribution of malware through official channels. In fact it creates a false sense of security that anything downloaded through official channels is safe. Though one may be happy under a false sense of security, it’s still not something to be happy about. At best this lock down offers reactive security, which is better than nothing, but not as good as having the ability to run software in a security sandbox in the first place.
Sandboxing doesn’t help in this situation. Even if a piece of software can’t get outside the sandbox, if you voluntarily run it inside of your browser, it has access to whichever features the parent process does. If you install an extension that happens to be malware in a sandboxed browser, it might not be able to get at your files or other data but anything you put in that browser is compromised in either case. That means web history, form entries such as credit card numbers and passwords, and any other information said malware wishes to collect. As it’s running inside your browser, which has network access, so does the malware. Network access and data, that’s all they’re after anyway, and you can’t effectively block browser extensions’ access to these facilities since they depend on such things to function. Both sandboxing and walled gardens offer you a false sense of security in the same way. I prefer to call Google’s approach a gated garden, since you can easily get out if you wish. The one advantage such systems have over sandboxing is that malware, if detected, can be revoked and killed. That power can, of course, be abused (Apple, I’m looking at you) which is why a way out is important.
darknexus,
“Sandboxing doesn’t help in this situation. Even if a piece of software can’t get outside the sandbox, if you voluntarily run it inside of your browser, it has access to whichever features the parent process does.”
That’s probably the heart of the disagreement right there. It’s not really the case that a sandboxed extension has to have the same level of access as the parent process.
Not the same level of access. But, given the tasks of most browser extensions (to alter your browser environment or your displayed web page in some way) I’d like you to explain how a browser extension could operate without having access to the browser’s dom or the ability to use the network. Those two abilities are the only things modern malware need to steal any information you enter while online.
It can be a challenge to come up with good permissions. And sometimes legitimate code will require access to functionality that would be exploitable by malware. To make things even more confusing, sometimes the same code can be both legitimate and malware depending on how it is used. How do you classify such things? My own opinion is that technology should err on the side of freedom, so long as users are properly informed of the risks.
I must acknowledge there are differing opinions like yours out there. But I’d like you to elaborate how the walled garden is any better at pre-emptive security when software channels aren’t officially being vetted? To repeat an earlier point, I don’t think the walled garden offers any pre-emptive security at all on it’s own.
As for reactive security, of course malware can be pulled from a google repo, but then it can be blacklisted from other sources too. So what security advantage is there to making users download from a centralised repo?
The centralised repo will only be more secure if google starts to police the submissions, and that brings us full circle to the points made earlier about the consequential loss of freedom. Even so, a better, far less controversial solution to this whole ordeal of sideloading would be for google to maintain white/blacklists that could be enabled by default in the browser. This way everyone could benefit from warnings that something is known to be explicit malware, which is much more useful than assuming that anything not loaded from google’s repo *might* be malware.
Edited 2012-06-13 03:33 UTC
Metadata can be faked. This method ensures that only people tech-savy enough to know how not to break their browser has enough control to break their browser.
Laurence,
“Metadata can be faked. This method ensures that only people tech-savy enough to know how not to break their browser has enough control to break their browser.”
Can be faked to do what? Any metadata can be faked. But if the requested permissions are enforced by the sandbox and software attempts to escalate it’s access above that specified in metadata, then it should be killed automatically. Furthermore the default max permissions should be restrictive enough such that the user needs to explicitly ok dangerous calls before the software will run.
The sandbox gives us much more security than we normally have when running extensions under blind faith. Although this could improve security for all extensions, I’d be open to removing sandbox restrictions from extensions that have already been vetted by google.
they also don’t let you install apps on your android device unless you use their store! google says SUCK IT!!!!
Since when? I’ve known of some carrier ROMs that have this ability taken out either by the phone manufacturer or your carrier, but I’ve not seen any stock Android rom impose such a limitation.
MMM, I’ve got an Android Phone and Tablet, I’ve installed a few apps “not” from the Andriod Store. You know like Amazon’s store? You might wanna look at your settings on your device and check “Unknown Sources”
the same as “Allow apps from outside the Marketplace” that exists on Android.
You can toggle this off with a switch. Nothing to see here.
And trust me, I chomp at the bit to diss Google
1
Browser: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9.2.8) Gecko/20100722 Firefox/3.6.8 (.NET CLR 3.5.30729)