How can I trust Firefox?

Discussion of general topics about Mozilla Firefox
User avatar
BenBasson
Moderator
Posts: 13671
Joined: February 13th, 2004, 5:49 am
Location: London, UK
Contact:

Post by BenBasson »

scratch wrote:keep in mind he was installing under virtual pc. this could be the cause.

Could be, but it's too easy to ignore. I'll look into it if necessary, but I'm sure others here are already equipped to do so.

scratch wrote:so they shouldn't be allowed to use mirrors? many other sites do so, including big name download sites. i think microsoft even does.

If Mozilla.org has lots of official mirrors, potentially they could obtain subdomains of mozilla.org and map them to external nameservers. At least then it looks like an official distribution channel.

scratch wrote:it doesn't now? sure, all the extensions out there are unsigned, but just the fact that it says this seems to indicate that it isn't always the case.

It doesn't appear to, since it's in the Firefox 2.0 Download Manager Wiki.

scratch wrote:and remember, the vast vast majority of apps out there are unsigned.

You're correct, but the vast majority of websites aren't standards compliant either. Even if other applications don't use these things, I see no reason why we shouldn't.

shakey_snake wrote:Explain to me how we're soooooooooo immature. And even if we are, how that invalidates the arguement that this article is a piece of crap.

Oh, the irony.
Last edited by BenBasson on December 20th, 2004, 6:00 pm, edited 1 time in total.
User avatar
CableModem
Posts: 291
Joined: August 23rd, 2004, 3:32 pm
Location: Palo Alto CA

Post by CableModem »

He mentioned several times the phrase "VPC", which I take to mean he was running Virtual PC on a Mac, which explains many of the errors. Next time you want to critisize Firefox, try it on a true Windows box, not Windows running on a Mac.
User avatar
scratch
Posts: 4942
Joined: November 6th, 2002, 1:27 am
Location: Massachusetts

Post by scratch »

...or try the mac version. but i'm not sure your assumption is correct. i think what he meant was he was so paranoid about firefox being insecure that he ran it on a virtual windows pc on windows (there is now a version of virtual pc for windows), so that if it did something nasty, it wouldn't infect his main system.
User avatar
rfrangioni77
Posts: 1510
Joined: January 5th, 2004, 1:48 am
Location: Bumdumbourge, near Totalslava
Contact:

Post by rfrangioni77 »

Personally, I don't "trust" any application. I trust myself...but personal responsiblity isn't as popular a topic as blaming other things, I guess... *shrug*
Lost User 36785
Posts: 0
Joined: December 31st, 1969, 5:00 pm

Post by Lost User 36785 »

rfrangioni77 wrote:Personally, I don't "trust" any application. I trust myself...but personal responsiblity isn't as popular a topic as blaming other things, I guess... *shrug*


Yep, it all boils down to common sense--another not-so-popular thing these days.
crlorentzen
Posts: 83
Joined: January 14th, 2004, 10:22 am
Location: USA

Post by crlorentzen »

Peter Torr wrote:(Secunia currently lists three unpatched vulnerabilities, for example).


one of which is in all web browsers.

and this is so bad compared to 20 (my count may be wrong looking through the 74 advisories) unpatch vunerabilities listen for IE? http://secunia.com/product/11/
Firefox: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2b5) Gecko/20091204 Firefox/3.6b5
Thunderbird: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1.7pre) Gecko/20091221 Lightning/1.0b2pre Shredder/3.0.1pre
moza
Posts: 414
Joined: October 22nd, 2003, 2:00 am

Post by moza »

I know this is ment to be on the serious side of security, But I just cant help going back to this little snippet from the blog..
(Always remember the Ten Immutable Laws of Security, and in particular Law #1: If a bad guy can persuade you to run his program on your computer, it's not your computer any more.)

Urrm is he refering to our Honourable Bill Gates and the Windowz & Office range of products (Ever fully read the EULA) ???
xexagon
Posts: 407
Joined: March 24th, 2004, 2:52 pm
Location: Eastern flatlands, UK
Contact:

Post by xexagon »

Cusser wrote:Bug: Looks like there's installer errors... I use zip builds exclusively, so I wouldn't know, but this should be investigated if it's true, and I seem to remember a problem like this before... could be a regression.


Well it's either a bug, or it's not. Which is it? I thought you were going to check?

Cusser wrote:Point: Mirror URL is unrelated to getfirefox.com or mozilla.org


Agreed, as I already said. This could easily be remedied by the site offering a choice of 'trusted' mirrors to download from.

Cusser wrote:Point: Binary is unsigned, triggering WinXP SP2s warning.


Am I missing something here? I'm trying to think of one program that I use regularly that is signed. Just because Windows says the download isn't signed doesn't mean it's untrustworthy (as the article states). Are you proposing that all software should be approved by Microsoft before it can be downloaded? Why use a cross-platform, free, open source browser if that's the case?

Cusser wrote:Point: Extension system should support signing
Note: It will before Firefox 2.0


I don't know much about this, to be honest. Will Mozilla (rather than VeriSign) be signing the extensions? If so, that's great.

Cusser wrote:Point: If you want to make users more security aware, giving them secure tools (such as Firefox) isn't the only solution. Several steps of the processes outlined could directly lead to false trust in unsigned binaries and lead users to ignore otherwise useful warnings.


Is the warning useful, though? All it means is that someone has paid a few hundred dollars for a certificate. I'd trust/value it more if the process was free.

Cusser wrote:An inexperienced user following the same process would get the following facts:
1) It's ok to install from unidentified sites
2) It's ok to install unsigned binaries, despite Windows warning me that it could be anything.
3) It's ok to install extensions that are unsigned.

To an extent, yes, it's fine if you do those things with Firefox and get extensions from update.mozilla.org, but in essence it's training users to ignore security measures.


1 - This is easily remedied if Mozilla announces it's using a mirror.
2 - Depends on how much you value the current signing system.
3 - Only if you white list a site, read the warning, wait for a few seconds while you consider your action, then press install. Signing things sort of goes against the spirit of open source (it's a trust/community thing). Still, if Mozilla is signing the extensions, then that's fine by me. Will they charge to sign, though?

Cusser wrote:Not everyone who critisises Firefox does it in a frankly ridiculous manner with little basis to their arguments. Firefox might be more secure, and the article may not be correct in all of it's assertions (installer errors?) but it's basis is spot on.


So it's got a some of its facts wrong, shifts from the particular to the general, scaremongers, is selective (e.g. it says FF has x Secunia advisories, but doesn't say how many IE has, or of what type) but the gist of it's right? Hmmmm.

If an article is well-argued, objective etc. then I am more than happy to take it on board, but this struck me as some sort of scattershot rant. Still, it does indirectly raise questions about the direction of Firefox as it becomes more popular. For example, community products rely on trust a lot; for how long will Firefox maintain this approach?
User avatar
BenBasson
Moderator
Posts: 13671
Joined: February 13th, 2004, 5:49 am
Location: London, UK
Contact:

Post by BenBasson »

xexagon wrote:Well it's either a bug, or it's not. Which is it? I thought you were going to check?

Cry me a river. A regression is still a bug, and I'm not figuring it out when it's past midnight. You're free to do so... I might have a go later.

xexagon wrote:Am I missing something here? I'm trying to think of one program that I use regularly that is signed. Just because Windows says the download isn't signed doesn't mean it's untrustworthy (as the article states). Are you proposing that all software should be approved by Microsoft before it can be downloaded? Why use a cross-platform, free, open source browser if that's the case?

I don't think it has to be signed by Microsoft, and in any case, we should have Firefox signed. If, as of Windows XP SP2, a user is going to see a warning for every unsigned package, I can't honestly think what the motivation would be behind deliberately not signing it, so I'm suggesting this is done.

xexagon wrote:I don't know much about this, to be honest. Will Mozilla (rather than VeriSign) be signing the extensions? If so, that's great.

I don't know about the details. I read the Download Manager Wiki which states that signed XPIs is a prerequisite for Firefox 2.0

xexagon wrote:Is the warning useful, though? All it means is that someone has paid a few hundred dollars for a certificate. I'd trust/value it more if the process was free.

The point is that you know who signed it. The point is consistency. You go to mozilla.org, your download comes from a subdomain of mozilla.org (regardless of where that domain is mapped to on the Internet), you run the file, it says it's signed by Mozilla.org.
David H
Posts: 2254
Joined: June 13th, 2003, 5:17 am
Location: Japan

Post by David H »

What a sad, sad attempt. I didn't see one thing that makes me even consider the idea that there is anything misleading or untrustworthy about Firefox. I'm going to go through the article and comment as I go along.

Not only does this software come from a completely random university server, but I have no way of checking if it is the authentic Firefox install or some maliciously altered copy.


Redirections are done every day by thousands of sites. The link is posted directly from Mozilla.org. Isn't that trustworthy enough? Besides, this is blatently wrong. If you're really that paranoid, get the md5 or sha1 sum from the site and use them verify your download. Not easy, sure, but possible.

Oops, my network connection died.


Blatant fearmongering and an attempt to unfairly cast doubt on the integrity of Firefox by associating it with a random negative event.

...this time coming from -- I kid you not! -- a numeric IP address, the bastion of spammers and phishers and all manner of other digital rogues


Blatently disregarding that there are plenty of legitimate uses for numeric IP addresses.

a completely blank MessageBox.


Which most likely comes from the fact that he's using a virtual machine and not a supported platform. And if he's so paranoid, why continue at this point?

It dutifully tells me the extension isn't signed (good), but makes the default choice Install Now (bad). This is the opposite of what Internet Explorer decided to default to when it detected unsigned code (ref: above). Now tell me again, which is the more secure browser?


A very, very minor point, and one that ignores the fact that IE will often install things without asking you at all (or at least did until SP2).

[regarding Flash] That's probably a good move for most users, although personally I tend to click Run inside IE because I know it will warn me about unsigned programs.


Very trusting of IE I see. Why so suspicious here though? Isn't getting it from the Macromedia site enough?

I just get the usual "This could be a virus; do you want to run it anyway?" dialog. But without any evidence to base my trust decision on (where it came from, who the publisher was, etc.), what should I do?


Didn't you just download it from the Macromedia site? I think the point of the dialog was to make you think about fishy forums and such, not about legitimate sources of popular software.

What's really frightening though is that there is a "Don't ask me again" option in this dialog...


...Didn't that originate with IE?

How do I disable Flash inside Firefox? Good question. I don't see any menu items or Tools -> Options settings,


He obviously just didn't look hard enough. Options>downloads>plug-ins. In addition, you can just remove the plugin from you plugins folder.

According to Google, I have to download yet another unsigned extension to enable the blocking of Flash content.


According to Google? What does that mean? And since when does IE even offer that option?

How do I know I didn't just install some terrible malware from a compromised web server?


How do you know you didn't just install some Malware the last time you downloaded an IE "helper" toolbar? Indeed, which browser is more likely to install things undetected? You're supposed to use your own best judgement. Certificates mean next to nothing in terms of security.

Law #1: If a bad guy can persuade you to run his program on your computer, it's not your computer any more.


Which means that, in the end, security is up to you, not the software. He's twisting this quote to try to pin the blame on Firefox for something that is the user's responsibility.

To continue my benevolent fairness, I actually think Firefox is a nice browser.


Yeah, a real balanced job he's doing here.

But just because it doesn't currently have any unpatched security vulnerabilities talked about in the press doesn't mean they don't exist.


Generally true of all software. It's the overall track record that is most telling. And in the Firefox vs. IE race, it's no contest.

Mozilla keeps their security bugs hidden from the public (just like Microsoft does)


Mozilla keeps it's security-related bugs under wraps until it can fix them, which is generally a short time depending on the severity of the bug. The security-sensitive flag is for that brief period of time between when a bug is found and it's fixed, so that it doesn't exacerbate a problem that's already known. M$ however apparently keeps security bugs under wraps so that it doesn't have to fix them. (And sometimes doesn't even fix them even when they've been made public).

But the thing that makes me really not trust the browser is that it doesn't matter how secure the original code is if the typical usage pattern of the browser requires users to perform insecure actions.


...with a rehash of the rather innocuous or dubious complaints mentioned in the article. It also completely ignores everything Firefox does to protect users from doing insecure actions, such as requiring you to whitelist a site before installing extensions, not allowing you to automatically install executables, reasonable default security settings, and more. Plus, there isn't the inherently dangerous ActiveX and integration with the OS that IE has.

I personally don't care if people choose to run Firefox or Linux or any other software on their computers -- it's their computer, after all -- but we'll never get past the spyware / adware problem if people continue to think that installing unsigned code from random web sites is A Good Idea.


Which seems to be deliberately implying that Firefox is an insecure application, and that it will only harm anyone who uses it.

Overall, this is one of the most sickeningly biased bits of writing I've ever seen. It simply oozes with phrases designed to make you question the integrity of the program. I'd say it was a masterpiece of propeganda if it wasn't so ham-handed about it. And the comments following it are just as biased. Looks like a classic astroturfing attempt to me.
michaell522
Posts: 2417
Joined: November 4th, 2002, 4:47 pm
Location: London, UK
Contact:

Post by michaell522 »

David H wrote:
Not only does this software come from a completely random university server, but I have no way of checking if it is the authentic Firefox install or some maliciously altered copy.


Redirections are done every day by thousands of sites. The link is posted directly from Mozilla.org. Isn't that trustworthy enough? Besides, this is blatently wrong. If you're really that paranoid, get the md5 or sha1 sum from the site and use them verify your download. Not easy, sure, but possible.

I think he has a point, and it applies even more to extensions.

The fact that the link is from mozilla.org doesn't mean anything much - mozilla.org has no direct control over the stuff that's on a bunch of mirrors managed by third parties. As for the md5 or sha1 sums - I'm not sure there are any on the mozilla.org site, and having them on the mirrors doesn't help to assure anyone of anything. PGP signatures are better, but you're still almost certainly going to be in a position of trusting a key from a keyserver without knowing who it actually belongs to.

With ActiveX vs XPI - they're both equally "dangerous" in the fact that they mean running software on your computer without any kind of "sandbox". Of course, IE has had a bunch of flaws in the past, many before Firefox even existed, which meant things got installed without the user being prompted. But as of IE in XP SP2, installing ActiveX has much the same level of whitelist/prompting as installing an XPI in Firefox. It has been explicitly said by the Mozilla security folks several times that the Firefox XPI whitelist is not a security measure as such - it's to prevent users being confused/inconvenienced by random sites attempting to install XPIs that the user doesn't want, in the same way that the popup blocker blocks unwanted popups. ActiveX has the advantage that legit stuff is signed - XPI stuff isn't. Firefox isn't really ahead in that race any more, and if we're going to keep marketing Firefox on the basis of security, I don't see why it's a good thing that XPIs aren't signed.

Overall, this is one of the most sickeningly biased bits of writing I've ever seen. It simply oozes with phrases designed to make you question the integrity of the program.

I've seen stuff just as bad on "our" side - Asa's blog, for example. And much of the stuff in these forums...
User avatar
DuncanL
Posts: 235
Joined: August 1st, 2003, 7:41 am

Post by DuncanL »

michaell wrote:ActiveX has the advantage that legit stuff is signed - XPI stuff isn't. Firefox isn't really ahead in that race any more, and if we're going to keep marketing Firefox on the basis of security, I don't see why it's a good thing that XPIs aren't signed.

But anything can be signed simply by paying the fee. There are no tests involved; a signature only proves that it was signed by the person it says signed it (and even that could be faked by using a stolen credit card). Plenty of spyware is signed.
There are no magic solutions (though Firefox's whitelist seems a sensible route to me - it places the burden of trust on the end user); a signature does not prove the intent of the signed code.
Current Firefox Version: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-GB; rv:1.9.1b99) Gecko/20090605 Firefox/3.5b99
xexagon
Posts: 407
Joined: March 24th, 2004, 2:52 pm
Location: Eastern flatlands, UK
Contact:

Post by xexagon »

Cusser wrote:
xexagon wrote:Well it's either a bug, or it's not. Which is it? I thought you were going to check?

Cry me a river. A regression is still a bug, and I'm not figuring it out when it's past midnight. You're free to do so... I might have a go later.


Well, you did say:

Cusser wrote:Excuse me while I go and do something constructive with my time relating to this post, like ensuring that appropriate bugs are already on file.


so I assumed you were so taken by the writer's argument that you didn't want to spend one more minute arguing with us 'immature' types, just in case we'd all downloaded a trojan-infested copy of Firefox. My mistake.

I thought you were in Reading (about 120 miles from me). It's certainly not just past midnight here in Colchester!

Cusser wrote: I can't honestly think what the motivation would be behind deliberately not signing it, so I'm suggesting this is done. . . . I don't know about the details. I read the Download Manager Wiki which states that signed XPIs is a prerequisite for Firefox 2.0


Incidentally, SP2 didn't flag anything when I downloaded and installed Firefox 1.0. If it does, then I guess it is worth shelling out a few hundred dollars on a certificate. Not because it makes Firefox any more secure, but because it avoids needless scaremongering and confusion.

The idea of a centralised, paid for, catch all certificate system strikes me as doomed and opposed to the whole idea of the internet. However, assuming Mozilla signs certificates relating to Mozilla products, then that's great. I mean, why should Mozilla get Microsoft to sign its products?

Anyway, nothing has convinced me that the original article was anything more than a scaremongering piece of blatant propaganda.
User avatar
happyweasel
Posts: 2
Joined: December 21st, 2004, 7:57 am

Post by happyweasel »

He is basicly saying that if you are stupid the software is insecure. Isn't that true of all software?
schapel
Posts: 3483
Joined: November 4th, 2002, 10:47 pm
Location: Ann Arbor, Michigan
Contact:

Post by schapel »

happyweasel wrote:He is basicly saying that if you are stupid the software is insecure. Isn't that true of all software?


No, you're not getting the main point. The point is that someone can alter the copy of Firefox on one of the mirror sites, and you wouldn't know because you have no way of checking whether the copy you're downloading is an official copy of Mozilla or not. With an MD5sum posted on the mozilla.org site, you can compare the MD5sum of your downloaded copy to that and help ensure you have an untampered copy.
Post Reply