Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> Would you prefer if Windows worked like iOS where there are serious limitations on what apps can do in the background and where every application has to go through a central authority for vetting?

I've thought about this for a while, and honestly, for desktops/notebooks/tablets? Yes. Maybe not just one App Store like iOS, but at least sandbox all possible non-os code similarly to ChromeOS, in a way that's on by default and requires a boot-time flag to disable (and users should be allowed to do this, but OEMs shouldn't.)

Recently my dad bought a new $300 toshiba laptop because his old machine was just "slow", as in he had so much spyware on his computer that it was easier to just buy a new one than going through the hassle of cleaning up his old one. Even though his old laptop was perfectly good and of a recent hardware generation.

I'm 100% positive he's going to have the same issues on his new laptop, and his response was that he uses his iPad so much that it doesn't matter anyway.

The role of the modern day OS has changed immensely over the years. Nowadays there's simply no reason for legitimate applications to have the level of access to the underlying system that they used to have. Apps really don't need arbitrary filesystem access. They don't need to be able to overwrite core system files. They should be run in a sandbox or a container with as restricted of a set of permissions as possible.

For servers and development workstations the story is a little bit different, but those are exceptions to the rule, and with the proper release hatches like boot-time enabling of un-sandboxed code it's a good tradeoff IMO.



Right, so because Windows has been historically shitty creating this problem in the first place, we now willingly want Trusted Computing and thus losing control and ownership of our own devices, with the grandma reason given as rationale. Well, when this relationship with the likes of Apple and Microsoft goes toxic, and it's only a matter of time, I'll be one of those that will enjoy it, as I'll consider it a sort of social justice, as a lot of people told you so and you wouldn't listen.


Instead of this sanctimonious sermon, we would be better served attempting to solve the problem that locked-down platforms do, but in a more open way.

Because the fact remains - computers have been a difficult thing for many people to use and maintain reliably. The "Windows has been historically shitty" point might be a reason this problem has been worse than it needed to, but it's hardly a compelling excuse.

When people—and not just Grandma—point out that their iPads don't have the same problem, then we should take that on board, rather than telling them that they're stupid for not listening.


So I'm all for solving current problems, however if I'm sure of anything, I'm sure that companies like Microsoft or Apple won't deliver a solution that solves these problems without locking-in users into a trap which is worse than the problem that it's solving.

And on Windows, this is the oldest trick in the book - in order to sell something people don't want, you first need to create the demand for it.


> Nowadays there's simply no reason for legitimate applications to have the level of access to the underlying system that they used to have

I'm hearing this argument from time to time. But I'd like to know what exactly changed. I can think of a dozen use-cases for which a shared, system-wide filesystem would be absolutely necessary (yes, especially with computer-illiterate friends and relatives, as "files" is pretty much the only abstraction besides "web" that is widely understood even outside of "geek"/"power user" circles).

So, what exactly has changed between then and now that made those use-cases legitimate in the bast but not anymore now?


People neglected the fact that having $HOME access was already bad enough and lived in the illusion that running applications as normal user was safe.


Why is having a home directory bad?


Any application running under the user id has full access to his private data.

Add the capability to do network communication and suddenly the all wolrd has access to $ HOME.

This is why in the container model of mobile OS and Windows/Mac OS X sandboxes, applications only get to see file handles to files choosen by the user.


> he had so much spyware on his computer that it was easier > to just buy a new one than going through the hassle of > cleaning up his old one...

I don't understand. Why not just reinstall the operating system from blank media? If you get a new machine, you'll have to reinstall your user-applications and data anyway?


It's not something most people know how to do, and you have to pay for a blank media set. There is usually a "recovery partition" instead, but how well do those work?


You can download windows install files from Microsoft nowadays. There's a downloader that also includes functionality to burn a DVD image or create a bootable USB stick.

http://windows.microsoft.com/en-us/windows-8/create-reset-re...

As for the "it's not something most people know how to do": Buying a new computer also requires you to transfer files from the old to the new computer. And requires you to install your software. I'd say that both tasks aren't significantly harder to perform than clicking "next" on the Windows install dvd.


I like the way OS X handles it. Unsigned software won't run by default, but the user can override the limitation.


Yes, you can do this on Windows too, but the problem is this update disabler executable was signed, so it would have run anyway.


There are two problems: how on earth samsung may have the idea to produce this kind of software and who is signing software without giving a look at what he is signing. A problem of software development mentality (on windows) and a problem of responsability.


> who is signing software without giving a look at what he is signing.

Versign would by proxy, as one example. A certificate authorized to sign code was purchased from them. Samsung would directly, to prove that this shit software came from them.

Just like SSL/TLS. I could set up an SSL website that performs drive-by attacks, would Verisign sign that? Yes, yes they would.

Vericode isn't an gate keeper like Macs certificates are. It's designed to improve security: if I download a Samsung installer Windows will tell me that it is indeed from Samsung (during the UAC elevation) because the signature checks out. This means that I can be certain that unbeknownst malware won't be installed on my PC alongside the Samsung malware.

The story is different with WHDL (drivers), those are signed by Microsoft (in addition to yourself, I think).


Samsung signed the software, of course. Who else would?


when you are asking for companies to verify all the software that is signed by their certs, im quite certain you are not at all considering the consequences.


Signing software only proves that it came from you. It doesn't say anything about what the software does. That's what your reputation is for.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: