As far as replacing parts is concerned, you can always pop a new battery or memory card (at least on non-iOS devices). By the time a non-replaceable part is gone, chances are that other parts are also getting old and you probably need a new phone anyway. If you want to spare the money buy a ~$400+ phone, and if you don't, you can get a ~$100 model that does everything your old phone did, and use the old phone for a simpler purpose just like you would do with an old laptop. If you still insist on replacing the part, you can always go to a "mobile hospital" (there are literally hundreds here in India, and I've seen one or two in pretty much every major US city) and get the screen, camera, or charger port replaced. Spares are available for everything. If you did the same replacement on a Fairphone, the most you would be saving on is the labor that the mobile repair guy would be charging. And even that savings would probably get lost because the Fairphone modules would be costly because of their low volumes, proprietary physical interface, and 'ethical manufacturing'. Sure, it's ethical for the big picture and Gaia would be so happy, but it's probably cruel to my communications budget which is already getting bled dry paying ridiculous prices for mobile data. $10 for 3 long youtube videos over your data limit? It's highway robbery out there.
As far as actual customization goes, the smartphone OS and hardware ecosystem pretty much has you locked in place. You can't really remove any features. Want a phone without a camera? you'll either have a half-functional phone (no QR codes, etc.), or you'll have a buggy piece of shit. Maybe the developer of your favorite app did the right thing and added a check for a camera, but most likely he just programmed it to go straight to the camera and getImage(). Want a phone without a GPS to go with your tinfoil hat? It's most likely coupled with some other useful part (like a modem) in a module, and even if you could take it out, the Google WiFi SSID-based location system's already got your exact location within 3 feet as soon as you turn on the WiFi. Want a phone with a bigger camera? Why not get the extra processing power that it would end up needing anyways and get the newest and biggest Samsung, HTC or Xiaomi. Wanna upgrade GPU, CPU, RAM or root storage? Fuggedaboutit - It's probably soldered to the base phone and/or inside an SoC chip. Want a new OS? Good luck finding anything that is worth switching to and will run reliably unless your phone sold at least a 100,000 copies. Want a bigger battery or a wireless charging system? Most good and recent-ish phones can be easily fitted with one.
You can't compare the IBM PC platform to these SoC phones, and here's why:
1. The choice of internal components is very limited, and a lot of stuff is combined into one chip. There are 3-4 major SoC vendors out there. Same goes for the Camera CCDs, sensors, and all the chips that go in there. Compare that to the ~10 major graphics card and motherboard manufacturers in the market. Or the ~40 different types of CPUs you can install on a given motherboard socket.
2. Unlike a desktop OS with replaceable drivers, an Android OS image has to already have all the drivers it needs installed in the image. With a desktop OS you get an installer that lets you install it on any compatible hardware. With a smartphone OS, the install takes a team of 40 engineers 6 months of work, day and night, after which they release the OS image (which still has fucking bugs in it, mind you, till they iron half of them out with every +.001 version). Or some smart kid makes the install using Cyanogen or something in one all nighter, but it ends up having bugs that never go away because the kid got bored of the project or got hired to that team of 40 engineers.
3. The choice of manufacturers and models at different price points is simply fucking crazy. There are easily hundreds of smartphone/tablet manufacturers out there if you know how to look beyond what Best Buy/Verizon will sell you. You can pretty much customize by simply choosing the right phone for what you need.
4. Unless your phone usage is very different from the average user, there's almost no difference in the experience of using any of these devices. No matter how good your phone is, it's gonna get slow as fuck anyways once you actually start using it because you'll keep installing shit until it slows down, and then try to optimize it. 5. When a good Samsung falls out of your pocket, you get a few scratches and your Gorilla glass screen survives the drop. When a modular phone does the same, the floor is scattered with modules and pieces of your dignity. Oh, and that 100 megapixel camera module that you got for $400? nowhere to be found - probably fell into the hole with the cockroaches in it.
Don't get me wrong - these problems could all be ironed out and we may indeed have a truly awesome modular phone in our hands someday. It's just going to a take an industry-wide paradigm shift (and a mature and widely deployed version of Project Ara) for it to be actually useful. Until then, the most a Fairphone will do for you is make you the star of one coffee break where people finally notice your Code Monkey ringtone. If you're really desperate and work at Facebook or some shit, maybe, just maybe it might be worth it for that.
I don't see why this represents a serious objection. You would obviously have the vehicle software / sensor stack optimized for local conditions. Indian designers could rig it so that lights flashed and horns honked. Algorithms could be designed so that you could simulate bluff charges / random aggressive behavior / whatnot.
Really, from what I've seen of third world driving, a simple pseudo random number generator along with five or so stock behaviors (go, stop, go faster, swerve, swerve more) should do just fine.
You present the problem better than I do. In third world driving, the road is full of random actors like that, and in close enough proximity that any automated driving device/"stack" safe enough to be made road legal would keep the car stopped the whole time. I have a reverse sensor that pings if there's an obstruction some x feet away from the car. If I'm reversing out of parking into busy mixed vehicle/pedestrian traffic (and you have to in places), it just beeps constantly - if I was to go strictly by the sensor I'd probably have to wait hours. Bad third world traffic requires a human driver who also acts as a negotiator for the right of way, and has a sensor grid complex enough to sense every inch of the car, and has the human intelligence required to predict the actions of these random actors quick enough. I'm not saying it's impossible for self-driving to one day be good enough to use here. It's just that the state of the art is not as advanced right now. No amount of million dollar cars that can barely traverse the I-90 by itself will convince me that self-driving cars are going to be a thing any year soon. All kinds of automation is already used for safety features and assists and that will keep getting advanced, but autonomous driving throughout the world is still pretty far from real.
Actually, if you combine all different versions it beats even the 64bit integer. Techsmartly made a fancy pivot chart of it a while back: http://techsmartly.net/freePS3...
Well played, sir. Well played
"Alcoholism is a disease, but it's the only one you can get yelled at for having. Goddamn it Otto, you are an alcoholic! Goddamn it Otto, you have Lupus! One of those two doesn't sound right."
There are a number of STDs that people get yelled at, too. But your point is well taken. Until recently, alcoholics were considered to just have weak character and were very badly mistreated. Recognizing addiction as a disease helped to change that stigma. Like alcoholics, for many, obesity is not a simple matter of mind over matter and a lack of willpower.
I partly agree with you. Morbid obesity that renders a person dysfunctional should be considered a bona fide disease. Ideally, the government should provide free proper rehabilitation for anyone morbidly obese to recover. But, the line should not be drawn at the BMI mark for Obese (which is what is implied), and here's why:
1. It would encourage indulgence by rewarding it, and lead to possible misuse.
2. It would increase hostility in the workplace against the Obese.
3. As as someone who almost touched the Obese line himself once, I think my weight and how I look is none of your business, or that of anyone that I engage with on a professional or social basis. Many people would be offended by being offered the Obesity benefit based on bad hair days. I don't think most Obese people want another possible label that just about anyone can throw at them.
Yelling is not the best cure, but not yelling also has major downsides. Sometimes people need to be alerted.
"Let's show this prehistoric bitch how we do things downtown!" -- The Ghostbusters