3 - Maintaining a data path forward is tricky enough. Demanding that users be able to downgrade at anytime would be a very tall ask if user data has to survive the downgrade.
4 - This seems outlandishly expensive to do. This effectively reads “nobody can ever deprecate an api on anything”. This also seems to be broadly incompatible with fixing certain security vulnerabilities - would everybody have to maintain TLS 1.1 or plaintext api endpoints for old clients? Would a social media network have to maintain api endpoints that leaked more data than users were comfortable with?
This also seems to be broadly incompatible with fixing certain security vulnerabilities - would everybody have to maintain TLS 1.1 or plaintext api endpoints for old clients?
Or they would forced to produce an update that doesn't do anything other than e.g. upgrade the TLS version --- and has absolutely nothing else.
I don't see how an automatic update setting is incompatible with 2. If a user says "go ahead and install updates as needed" that is the user expressing their desire to receive updates.
I also think the phrasing in 1 is a little needlessly aggressive though I believe it comes from a place of frustration. The difference in my mind between saying "this thing updated behind my back" and "this thing updated automatically for me" is whether the user has registered the update as being beneficial or not, and depending on the device, that's a WIDE spectrum. I know my smart outlets update their firmware all the time, and an extremely small handful of times I do notice, because sometimes they end up not reconnecting to the wifi quite right and need to be reconnected. However if they updated and, for example, broke HomeKit support and no longer worked, I'd be angry the next time I tried to use them.
> 3 - Maintaining a data path forward is tricky enough. Demanding that users be able to downgrade at anytime would be a very tall ask if user data has to survive the downgrade.
I mean, this is just an engineering problem pure and simple. Most of the time, in my experience, graceful downgrade just isn't prioritized because, well, who can even do it for starters? Installing old software oftentimes means you need to do some really intense stuff, like wiping whatever device entirely, so the retention of data is moot.
If this was mandated I see no problem with getting it done in my industry. It's simply a matter of making it a priority IMO.
> 4 - This seems outlandishly expensive to do. This effectively reads “nobody can ever deprecate an api on anything”.
With certain products I can definitely see it being an advantage, and the first place my mind goes to is again, smart home products and appliances, automotive hardware, that sort of thing. Large, expensive items that incorporate software that the user interacts with can be an absolute nightmare when the OEM randomly decides that the way something's worked for years and years for you is now just not an option, or worse still, locks it behind a paywall. And what are your options here? Buy a new car or dishwasher? Or eat shit and pay them $20 a year that they have not earned and are providing no value for?
This is why the newest car I have is a 2018 Corvette, because I know all it's software and have access to it, and there's no system that's going to lock my heated seats behind a Chevrolet Premiere+ subscription where I have to give chevy money to permit my car to engage a damn relay for me.
Continuous updates continue to permeate, including into things that are still surprisingly connected to the internet in the first place.
I think that in time, forced updates will cause enough trouble that people will become more conscious of and dislike them. For some, one bad update is all it will take.
So, I think it's worth waiting to see if anti-update competitors appear before regulating this.
Were you keeping your personal files in the usual "Documents" and "Videos" and such laid out by Microsoft? Or somewhere else?
For example, support for downgrades means you a security vulnerability can be reintroduced by a malicious user which may not be desirable. Writing software that’s backwards and forwards compatible across all releases can be extremely expensive to impossible (eg a feature in your application that requires a new OS or you need to use a now removed API when running on older releases).
There are difficult technical issues involved and trying to legislate specifics may not be the best idea vs other approaches that improve real freedom (eg you have to release sufficient details to your customers that they can write their own software for your hardware).
And there are regulated industries where a software update could be the fulfillment of a recall.
Actually, yes they are. I'm not a big fan of legislation, but the upgrade crazyness has to stop at some point!
> For example, support for downgrades means you a security vulnerability can be reintroduced by a malicious user which may not be desirable
What if "I, the user" deem it "desirable"?
I'm holding to bios with known vulnerabilities so I can work around "security features" that are "for my own protection" like 1) preventing me from underclocking (to keep the security features of the now-dead SGX) 2) using any M2 WWAN or NVMe that I want
It's gone to a point where it's not desirable for me to upgrade, and to prefer the risks that come with an exploit as at least I know my freedom to use my hardware the way I want will not suddenly become limited.
Another example: getting root on android with mediatek was considered a "bug" and work a mandatory "upgrade" that prevent users from being able to get root that way.
But I want to be root!
Only if said operator signed a contract. No contract=no liability.
Here in the US, the Supreme Court has made it clear that law enforcement agencies are not required to provide protection to the citizens cf https://www.nytimes.com/2005/06/28/politics/justices-rule-po...
If even the police isn't liable, why should I be liable or have any kind of duty to protect your system?
Your system, your problem.
You are using loaded words to 1) imply I would support some questionable actions where you assume intent and 2) refer to negative externalities, but I'll suppose you are not trolling interact in good faith with you.
Both the actions you define are ignoring property rights: your well, your parking lot = I can't do that, unless you allow me (with a contract!)
My well, my parking lot = I can do that, and you don't get to say what I do with my property, unless we have signed a contract which creates liability.
Many people seem to have a strong desire to be able to force ME to update MY browser/operating system/bios/whatever else to be up to THEIR standards, because it has consequences on THEIR liability.
I care a bit about them, but I care way more about MY freedom: I do what I what with MY computers.
Note that intent matters: I have no desire to cause bad things to other people. I'd be very sad if my computer was used as "part of a botnet" like someone else said. I might even try to avoid that - but only as far as it puts my freedom first, and there's not even a requirement that I try (because I might have better things to do lol)
Should bad things like botnets happen, 1) it wasn't my intent, as preserving my freedom was my intent 2) the negative externality is sad, but it's not my liability: you should secure your property, or said differently "your problems aren't my problem"
This whole interaction feels very strange to me. By any chance, are you European? Europeans seem to have very different concepts of freedom and liability than we do.
But we're talking about you deliberately not securing (or even actively reducing the security of) your property in a manner that could reasonably be predicted to lead to harm to other's property, and that harm occuring, through no fault of the harmed party. You're not supposed to store loaded guns unsecured on your front porch (not around here anyways).
>By any chance, are you European? Europeans seem to have very different concepts of freedom and liability than we do.
No, I'm from the US.
>You are using loaded words to 1) imply I would support some questionable actions where you assume intent and 2) refer to negative externalities, but I'll suppose you are not trolling interact in good faith with you.
I'm asking whether you would, not suggesting that you do. But yes, the acts in question were significantly more questionable than those in my previous comment, since your answer to that was more extreme than I expected. So how about the middle ground: Do you believe that the EPA should impose restrictions on companies' or citizens' right to dump whatever toxic waste they want into rivers? If you think the physical commons should be protected from predictable harm by negligence or reckless disregard, why not the digital?
We have different preferences about what's the right security/freedom ratio.
> You're not supposed to store loaded guns unsecured on your front porch (not around here anyways).
I don't think the government or anyone has any say about where or how I may keep my guns.
People believe they might have a say, so there are laws on the books, but they're frequently taken down by the courts.
> So how about the middle ground: Do you believe that the EPA should impose restrictions on companies' or citizens' right to dump whatever toxic waste they want into rivers?
I believe it shouldn't, but that's just my opinion.
You may not believe it, but in terms of efficiency and keeping the environment, it doesn't matter (see below)
> If you think the physical commons should be protected from predictable harm by negligence or reckless disregard, why not the digital?
I don't believe in in the physical word, and I don't believe it in the digital world either.
Regardless of my beliefs, there's a nobel prize winner who's shown that it doesn't matter how the rights are initially assigned, as long as parties can negotiate with no transaction costs.
Check the Coase theorem.
The EPA imposing restrictions create transaction costs, so I think we're better off without them
Even the author of the “theorem” you cite indicated that he didn’t believe it to be practical.
Backports exist because of this reason. Just added them as a requirement to the list of rules above.
Make them a bit terser, and maybe “Right to repair” will heave them out of the science fiction tarpit.
> 2. In particular, the user determines exactly when an upgrade is performed.
Haha, at last, yes !! Take that stupid windows XP countdown to reboot !!
For number 3 if we allowed for people to roll back any update that can include Teslas where they are doing OTA safety updates so that wouldn’t work out.
It took 7 years though:
Maybe there should be a law that says:
1. Upgrades may be performed but never behind the user's back.
2. In particular, the user determines exactly when an upgrade is performed.
3. The user may roll back any update at any time.
4. Any services which the software depends on should be compatible with all versions of the updated software.
EDIT: 5. Security backports should be made available. However, the user should always be in control over whether they are installed. Sometimes working code is more important than 100% secure code. Also this rule will prevent companies from quickly forcing an update and sweeping security breaches under the rug.