Which is done because solder is thinner/lighter than sockets, and therefore you can get a thinner/lighter laptop. There are real engineering tradeoffs all over the place here that are being elided over in order to make "things should be easier/cheaper to repair" seem obvious rather than a complicated set of tradeoffs.
You shifted the argument (which nobody was making) to being "modern is better" using your car example. I'm bringing it back to user-serviceable is _better_.
>Which is done because solder is thinner/lighter than sockets, and therefore you can get a thinner/lighter laptop.
So let them figure out another way to make it thinner/lighter. I don't want them to do it at the expense of upgrading common parts. We're not asking for every single chip to be replaceable. Just the obvious parts that have an established 10+ year history of being serviceable. I bet welding every single part on a car makes it lighter too, but I don't think people want to replace the entire car if their fuel pump breaks.
>There are real engineering tradeoffs all over the place here that are being elided
Huh? You switched from purely cosmetic factors - thinner/lighter, to engineering. You will have to explain that one to me...
Thinner and lighter are not purely cosmetic (although I think Apple overvalues them) -- laptops are meant to be carried and therefore being smaller/lighter are actual benefits. Even if they were purely cosmetic, however, there is actual engineering required to do them. The laws of the universe don't go, oh, you only want to make a cosmetic change, therefore you can get X into this space/thermal profile without any other tradeoffs.
User-serviceable parts are better, all things being equal. But all things aren't equal. You're trying to mandate tradeoffs for everyone else, and I don't think you understand everyone else's tradeoffs well enough to do that.
I think at some point you're going to have to present an argument that user-serviceable RAM makes a perceptible difference in making the laptop thin/light going forward. Several assumptions - The RAM itself can't shrink or be made to be more energy efficient, or socket design is static, can be challenged. However a manufacturer without a mandate/incentive will probably never invest their efforts to making that happen. They're going to optimize for the commercial aspect of selling the product in a way that maximizes their bottom line via market segmentation and price gouging on storage. Obviously, making money is not something that I'm against, but I simply would like to see an acceptable compromise made between making money and making it repair-friendly. Like you said, everything is a trade-off.
You want me to present an argument that stuff that doesn't exist isn't better than stuff that exists?
And "acceptable compromise" -- acceptable to who? Because the current compromise is acceptable to people who buy these laptops with soldered-on RAM. And it's not acceptable to those who buy other laptops that don't have soldered RAM. The compromise has been accepted by the market. "Well, the market got it wrong and should do it better," you might think. Sure, okay, but history is rife with examples of people telling markets to do things better and things not getting better, and it's helpful to think about why markets haven't gotten to better on their own.
Because you're the one dismissing user serviceability as an impeding factor to reducing the overall package size?
>And "acceptable compromise" -- acceptable to who?
To me, and anyone else who shares my values.
>Because the current compromise is acceptable to people who buy these laptops with soldered-on RAM.
Seems like you can layer-on any argument you want simply by citing sales. I suppose people think Windows security is the best since they overwhelmingly choose to buy Windows machines over Apple or Linux desktops. But that is a false statement, because not everyone is aware of the intricacies of operating system design or that different options exist, or how they differ etc etc. Extrapolating _just_ from commercial success to use as data in an argument is a fools errand because of the thousands of variables involved in a large diverse group making a purchasing decision.
>Sure, okay, but history is rife with examples of people telling markets to do things better and things not getting better,
Hmm I'm curious.. can you provide a sampling of those incidents that apply to our discussion here?
>and it's helpful to think about why markets haven't gotten to better on their own.
They almost never do. Systems typically evolve to coalesce around a local maxima. All around the world (in democracies anyway), citizens have always installed regulations to make the market serve us, rather than the other way around. For e.g. in the US - child labor, minimum wage, consumer product safety, etc, antitrust laws, etc etc. Before each of those laws got passed, I'm sure people made arguments against it with "well the market has accepted it". (You seem like a decent bloke, so I'm not suggesting that you personally would be against those regulations, just making a general point)
Introducing a constraint like this, however, motivates the manufacturers to innovate making user-replaceable parts better instead of discarding the entire concept just because they can.
