I'll admit to putting off applying security patches because my laptop is a giant repository of state and it's a pita to bring it all back up again.
I have to remember how to start postgres, the erb command to make rails work, figure out how my upgrade to bash subtly broke virtualenvwrapper, make sure I didn't accidentally leave anything open in vim in some random window or go through the whining about vim swap files, and dump/restore all the state spread through 20+ browser windows, probably 30+ bash instances, and a couple repls. etc. It's really fucking annoying.
Another thing that ate 3 hours of my life recently: post reboot, zoom.us, which previously worked flawlessly, no longer works. I of course discover this as a meeting is starting.
But breaks stuff more often. Why would that be a net win? Probably a net loss because you incur the mental overhead of putting down whatever actual problems you are trying to solve and getting into a frame of mind to try to fix your computer, more frequently.
1) update rarely, fix a lot of breakage at a time
2) update daily, run into each breakage as soon as it happens and fix it
3) never update, no need to fix anything
Did I miss anything? Just saying 2) is what works for me. Takes 1-2 minutes every morning before I start working (updates to all systems have been checked and downloaded overnight).
The problem with #3 is that, while there are no updates to break things, there are also no updates to fix things, so whatever is broken will stay broken.
Most people in this situation are forced to reboot at some point because of critical failure (power, or something physically break). Thereafter they carefully reboot on a regular schedule to just ensure everything will come up after the unplanned emergency reboot.
While I totally agree with most of the article, it made me grind my teeth at the frequency the author changes the oil in their car... That poor valvetrain will be full of gummy carbon deposits. RIP whoever buys the car second-hand.
How would changing oil every 2 000 km be a problem? Modern oil should last 10x that long without any problem. I am used to having recommendations for changing oil once a year or every 15k km depending on what comes first. I remember recommendations to change 2 times a year 15-20 years ago because it back then was more normal to have different viscosity in summer and winter
Depends on oil. Many cars still run on semi-synth and those require the 15k/1yr routine, full synth can use longer kms and two years, or every year if you do under 5kkm/year
The sticky sludge that GP refers to is for the most part an emulsion of water, produced by combustion and/or from ambient humidity, and oil. If an engine is not run frequently or is used only for short intervals the water isn't driven off by engine heat. Engines that aren't run much must be maintained based on elapsed time, not duration of use, due to factors such as this.
The author's vehicle is a prime sludge candidate, although not as bad as people who go years w/o changes.
Loss of viscosity is a wholly separate form of degradation which along with accumulated Ph changes is responsible for the "every N miles" recommendation.
I always wondered how this oil-change scam still continues in the US. Modern cars with synthetic oils don't need an oil change every 3-4 months or 3000 miles.
I always changed the oil according to the (European or Japanese) manufacturer's schedule (once a year / every 10-15000km) and had no issues with the cars (even >10 years old cars).
Is there a difference though? If you see my other reply to the parent, I posted the US oil change schedule for a european car. It seems to be in line with what I experience in europe .
Is it a difference between european and US made cars?
I know nothing on the matter, but it seems to me that it's a cultural difference. Americans are just more used to changing the oil often, so oil companies can sell them more oil even if they don't need it.
"For years the accepted oil change interval (as per carmakers) had been every 3 months or 3000 miles, whichever comes first. And why was that? It was because oils of yesterday broke down when left in the crankcase environment for longer than the prescribed interval. The combination of heat, friction, and the oil oxidizing over time resulted in an unholy clothing of the engine's internal parts called sludge."
That's an absolutely insane number, and I have never heard of that (I have had cars for 20 years that always were 10k to 15k km intervals, and simply done on a yearly or 18 months service).
I have a tractor from 1939. The owners manual states that modern oils are much better than old oils and as a result practices of years past do not apply. (I don't recall the exact wording) If you ask collectors today there is a big debate on if it is safe to use modern detergent oils in those old tractors.
The manufacturers are pretty picky with what oils they recommend for their cars too. I mean if I need a "fully synthetic 10W-30" for my car for example (from the manual) - and the brands are all global megacorps (Shell, Havoline, Mobil, Castrol etc). So with the tight spec and global brands surely I must be getting the same thing in the bottle here as one gets in the US?
In my case, it's cheaper to use non-synthetic oil and change it quarterly, which generally aligns with other maintenance activity. (Tire rotation, filters, lights, etc)
Here in Europe engine oil is typically always multi-grade, and often fully synthetic. I believe this is because standards are stricter here (although I can't find anything confirming this), and synthetic oil often leads to engines lasting longer and lower emissions.
This is why manufacturers have different recommendations for Europe. The formula of the oil is completely different to the US, so it can stay in the engine a lot longer.
> synthetic oil often leads to engines lasting longer
This is true in theory, but for most people the rest of the car will fall apart before the engine dies. All the rubber and plastic is degraded by air and sunlight which has little to do with any use patterns. (there are of course uses that put more wear on them as well)
A good portion of the cars are junked with so little use that they would have made it that far without any oil changes at all. Of course the oil would be like syrup, and the engine would show other signs of the abuse, but it would still be running well enough to make that last trip before the car is destroyed for other reasons.
Modern Diesels in the US have a 15,000 mile interval for oil changes.. but they still put a sticker in your window at both dealerships and oil change locations to come back in 3-5,000 miles.
My 'free oil changes under warranty' deal when I bought my truck required me to come in every 5,000 miles. Their computer system was setup for cars, and they had no way to change it..
My Subaru Forester gets an effective oil change every 2500 miles. It burns a quart of synthetic in that time. Too little to get the recall work done (faulty rings). But at least the car gets a gradual oil change over the year!
It's interesting you say that... Synthetic oils don't necessarily last longer than mineral oils. The oil itself isn't the problem - its the detergents, friction modifiers and things that break down through heat cycles and over time. As a rule with my vehicles, they get new semi-synth every 7,000kms.
When I had my last car (Direct injection petrol turbo), I changed it every 5,000. And that oil was cooked each time.
Having pulled the tops off more engines than I could count, you can immediately tell an engine that gets regular oil flushes compared to one thats run on longer intervals.
Sure, the engine might last your whole ownership without having a serious issue... But inside it'll probably tell a different story.
It's funny how attitudes towards oil change are different in different countries. In Germany, I would never go to the garage just for an oil change. We do an oil change either at the scheduled inspection, or when we're at the garage anyway (for winter tires) and the mechanic says it's time :-) which is about once a year.
My last car was a bit of a beater, and we did the oil change more frequently. Reason was that it had a problem with the transmission (manual, as most cars here), and from time to time there would be little metal shavings in the oil! But apparantly this is less horrible than it sounds, and somewhat normal for old cars. There is even a little magnet in the oil vessel to collect the shavings. After a combination of learning this, and having lived in the states for a couple of years, I insisted on changing the oil more regularly.
Well, I had a German car (diesel engine) for 12 years and changed oil according to the board computer, it came out roughly on average around 32,000 km, i.e. a little bit more than 1 year (sold it with 300,000 km after 12 years).
But the oil of a manual transmission (and as well that of an automatic one) is normally separated from the engine oil, do I read it correctly that you had a car with those in common?
All cars have one or more magnets (for ferrous shavings)it is perfectly normal, but of course they are pretty much ineffective for other metals, and more generally other possible contaminants.
You must have missed that he drives 2000 not 20000km per year.
He says he changes oil around every year, so every 2000km (1200 miles).
If he changed every 6 months, he'd be changing every 1000km. That's insane!
I don't even refuel every 1000km!
Example oil change schedule for a modern car (Taken from Audi Q7 '09 service manual):
8.000km, 25.000km, 40.000km, 55.000km...
So ~15.000km between changes - or 7.5 years if you drive 2.000km per year. Now - obviously you wouldn't go 7 years without an oil and filter change but you should be completely fine changing every 1-2 years on a car that is driven so little (It's the aging of the oil that's the problem, not the distance driven)
Don't know the specifics of the Audi Q7 of the OP, but my Mercedes A 180 CDI 2012 model (Diesel engine) averages 53 mpg (4.4 l/100Km) for my conmute trip (35 KM total, including ~8 KM of mountain road, with a 300m elevation change). Given that my A180 has a 50 liter (13.21 US gallons) fuel tank according to the manufacturer, that means I could make up to ~1,100 Km with a single tank. In practice I refuel before that, but routinely make ~900 Km between refuels.
Audi A6 diesel (yes very much a dirty VW/Audi diesel-gate diesel, although supposedly "fixed" now). It does 40-45 mpg US (5.5-6L/100km) and the tank is just bottomless (17 gallons/65L or something?)
This is one of the most common mesofacts I run across. It's a typical example of something that has changed, but was drilled into people's brains as important a long time ago.
Another reason for the resistance is the pain of upgrading.
Anything that requires a reboot, for example, is likely to be held off for a long time. When people are at the computer, it's because they want to use it right now. They aren't there to watch it reboot a few times.
It's less painful once you scale past single systems. Insofar as you have distributed systems that tolerate some fraction of their instances disappearing and reappearing, you can roll an upgrade out across the system without users having to notice that it's happened. Where I work we do this very well -- we update and upgrade live production systems for CVEs and feature adjustments quite frequently without systems running on them being any the wiser.
Still remaining is the sheer amount of independent dependency tracks into any given software system. Operating system, multiple language ecosystems, multiple infrastructure systems ... once you include development dependencies and tools -- and you ought to -- the surface area to watch and manage goes up sharply.
My company's philosophy, for better or worse, is to have periodic scheduled restarts. It's nice knowing I won't be suddenly interrupted for an update, but it's also a PITA to have to restore my state every Monday morning.
Sure, but if we're in dreamland, then I think the second biggest factor will be making security upgrades invisible to users. If you could ensure that you can upgrade any part of your OS/application while it's running, without ever requiring a reboot, then the problem disappears as you can simply stop asking people when they want a security patch.
In terms of safety, not much good comes from asking people.
(That, or let's just all agree that CVEs are a threat to Internet security, and have a policy for always displaying a message like "Your device will reboot in 05:00 in order to install a critical security patch. Please save your work." and have the users take comfort that it's not just their computer, but half of the planet is rebooting with them. That solution is game-theoretically unstable, though.)
That's fine, but it would require companies to abide by the rule of security updates being just that, security updates and not ways to push their corporate agenda after the fact, accidentally bricking your computer (or maliciously), installing spyware, adware or or other junk.
Agreed that asking is largely counterproductive when it comes to security patches (as opposed to feature updates). To satisfy the concerns of folks like @jacquesm, it might be best to implement a soft form of the hidden update: Make them opt-out rather than opt-in.
Keeps everyone happy and dramatically improves compliance.
I'm with you on this. When we have self-driving cars I hope they take themselves in to get fixed for recalls. (But at a time of my choosing - not like how computers will update at inconvenient times now.)
Yes, the author's assertion that safety recalls "elicit prompt attention" doesn't reflect my [US] experience. Many people don't take the recalls seriously or just can't be bothered.
> counterproductive, in that presenting updates through the same channel tends to conflate them in the minds of users, with the result that critical security updates instead end up being given the lesser attention more appropriately due to a new feature update
Good article, but actually understates the problem. It's not just that nonsecurity updates are appropriately due lesser attention. It's that they are appropriately due negative attention. My computer works fine the way it is. I actively do not want it screwed around with. I make a necessary exception for security updates. To put anything other than a security update into that channel is an abuse of trust that was granted for an important purpose, and trust that is abused is liable to be lost.
All the items listed in the article are FUD to get you into the dealership so they can upsell you. Oil changes are necessary but not critical to be done unless long overdue. Cheaper to do them yourself or find a reliable shop.
Software updates: Is the car running right or is there an issue?
Safety recalls: We're obsessed with the dog-whistle safety, but like anything else, there are trade-offs. Our Honda got recalled for the airbag inflators and I waited a few months before eventually taking it in. Result: No one died. I also have one car that has no airbags at all. I feel fine.
60 comments
[ 3.3 ms ] story [ 120 ms ] threadI have to remember how to start postgres, the erb command to make rails work, figure out how my upgrade to bash subtly broke virtualenvwrapper, make sure I didn't accidentally leave anything open in vim in some random window or go through the whining about vim swap files, and dump/restore all the state spread through 20+ browser windows, probably 30+ bash instances, and a couple repls. etc. It's really fucking annoying.
Another thing that ate 3 hours of my life recently: post reboot, zoom.us, which previously worked flawlessly, no longer works. I of course discover this as a meeting is starting.
https://www.stavros.io/posts/provisioning-your-computer-one-...
The last lines for me read:
"The main script
First, let’s start with the main Ansible script:
~~~yaml This playbook is meant to provision a new computer. Run with: ansible-playbook -i , provision.yml"
EDIT: Phew, fixed, good catch, thanks. There was a triple dash in the article, which Lektor interpreted as the end of the post :/
At least you found the cause.
The author's vehicle is a prime sludge candidate, although not as bad as people who go years w/o changes.
Loss of viscosity is a wholly separate form of degradation which along with accumulated Ph changes is responsible for the "every N miles" recommendation.
I always changed the oil according to the (European or Japanese) manufacturer's schedule (once a year / every 10-15000km) and had no issues with the cars (even >10 years old cars).
Is it a difference between european and US made cars?
"For years the accepted oil change interval (as per carmakers) had been every 3 months or 3000 miles, whichever comes first. And why was that? It was because oils of yesterday broke down when left in the crankcase environment for longer than the prescribed interval. The combination of heat, friction, and the oil oxidizing over time resulted in an unholy clothing of the engine's internal parts called sludge."
That's an absolutely insane number, and I have never heard of that (I have had cars for 20 years that always were 10k to 15k km intervals, and simply done on a yearly or 18 months service).
The standards for oil are tightly regulated in the EU.
But many brands meet that spec. It's really only white label conventional bottles that don't.
In my case, it's cheaper to use non-synthetic oil and change it quarterly, which generally aligns with other maintenance activity. (Tire rotation, filters, lights, etc)
This is why manufacturers have different recommendations for Europe. The formula of the oil is completely different to the US, so it can stay in the engine a lot longer.
This is true in theory, but for most people the rest of the car will fall apart before the engine dies. All the rubber and plastic is degraded by air and sunlight which has little to do with any use patterns. (there are of course uses that put more wear on them as well)
A good portion of the cars are junked with so little use that they would have made it that far without any oil changes at all. Of course the oil would be like syrup, and the engine would show other signs of the abuse, but it would still be running well enough to make that last trip before the car is destroyed for other reasons.
My 'free oil changes under warranty' deal when I bought my truck required me to come in every 5,000 miles. Their computer system was setup for cars, and they had no way to change it..
Parts stores still push the 3k miles/3 months (or did until very recently). Dealers and manufacturers don't, and haven't for at least 10 years.
Having pulled the tops off more engines than I could count, you can immediately tell an engine that gets regular oil flushes compared to one thats run on longer intervals. Sure, the engine might last your whole ownership without having a serious issue... But inside it'll probably tell a different story.
My last car was a bit of a beater, and we did the oil change more frequently. Reason was that it had a problem with the transmission (manual, as most cars here), and from time to time there would be little metal shavings in the oil! But apparantly this is less horrible than it sounds, and somewhat normal for old cars. There is even a little magnet in the oil vessel to collect the shavings. After a combination of learning this, and having lived in the states for a couple of years, I insisted on changing the oil more regularly.
But the oil of a manual transmission (and as well that of an automatic one) is normally separated from the engine oil, do I read it correctly that you had a car with those in common?
All cars have one or more magnets (for ferrous shavings)it is perfectly normal, but of course they are pretty much ineffective for other metals, and more generally other possible contaminants.
If he changed every 6 months, he'd be changing every 1000km. That's insane! I don't even refuel every 1000km!
Example oil change schedule for a modern car (Taken from Audi Q7 '09 service manual): 8.000km, 25.000km, 40.000km, 55.000km...
So ~15.000km between changes - or 7.5 years if you drive 2.000km per year. Now - obviously you wouldn't go 7 years without an oil and filter change but you should be completely fine changing every 1-2 years on a car that is driven so little (It's the aging of the oil that's the problem, not the distance driven)
> I envy you.
Still envy me when I say a full tank is $100? :)
Or actually, the car will be perfectly fine. Nobody changes their cars every 3-4 months in this here parts, and the sky hasn't fallen.
Cars usually get scrapped for some problem that is uneconomical to fix due to the age of the car, long before the engine is worn out.
Anything that requires a reboot, for example, is likely to be held off for a long time. When people are at the computer, it's because they want to use it right now. They aren't there to watch it reboot a few times.
It's less painful once you scale past single systems. Insofar as you have distributed systems that tolerate some fraction of their instances disappearing and reappearing, you can roll an upgrade out across the system without users having to notice that it's happened. Where I work we do this very well -- we update and upgrade live production systems for CVEs and feature adjustments quite frequently without systems running on them being any the wiser.
Still remaining is the sheer amount of independent dependency tracks into any given software system. Operating system, multiple language ecosystems, multiple infrastructure systems ... once you include development dependencies and tools -- and you ought to -- the surface area to watch and manage goes up sharply.
In terms of safety, not much good comes from asking people.
(That, or let's just all agree that CVEs are a threat to Internet security, and have a policy for always displaying a message like "Your device will reboot in 05:00 in order to install a critical security patch. Please save your work." and have the users take comfort that it's not just their computer, but half of the planet is rebooting with them. That solution is game-theoretically unstable, though.)
Keeps everyone happy and dramatically improves compliance.
There have been news stories about this too, e.g. https://www.theguardian.com/money/2014/dec/17/us-drivers-str...
Good article, but actually understates the problem. It's not just that nonsecurity updates are appropriately due lesser attention. It's that they are appropriately due negative attention. My computer works fine the way it is. I actively do not want it screwed around with. I make a necessary exception for security updates. To put anything other than a security update into that channel is an abuse of trust that was granted for an important purpose, and trust that is abused is liable to be lost.
Software updates: Is the car running right or is there an issue?
Safety recalls: We're obsessed with the dog-whistle safety, but like anything else, there are trade-offs. Our Honda got recalled for the airbag inflators and I waited a few months before eventually taking it in. Result: No one died. I also have one car that has no airbags at all. I feel fine.