Net margin? It costs several billion to get drugs to market. All that FDA safety and efficacy stuff is about $310 million, from lab rat studies over 3 years to a 3-year follow-up with patients to document side-effects and efficacy once the drug is out in the market. Only about 10% of new drugs make it to market, and a lot goes into the pipeline and doesn't even make it to being a "new drug". The sunk research costs are $3-$5 billion per drug.
For these drugs, they have to recover $3-$5 billion to break even. With a small target market, the gross margin is going to be a lot bigger to make that up. This is one of the reasons cognitive enhancement is medically ethical: something like Modafinil treats narcolepsy, but would be universally-useful to practically everyone due to its safety and efficacy as a non-addictive eugerogic (let you pull an all-nighter now and then) or study drug. The market becomes enormous, and the amount of money you can make on 60 cents per pill is ludicrous (it only costs about 11 cents to manufacture a 100mg dose). As it stands, it might cost you $12 a pill in America, although I can get it out-of-pocket for $240 for a 90-day supply from my PBM without insurance coverage.
Drugs with broad uses tend to become cheap and generic.
No, all evidence shows that when an individual human unit gets wealthier, it slows its rapid breeding. Humans don't need to produce more children when fewer die out, either, as a society.
As a population, a human society includes a gradient of wealth. The expansion of that society causes scarcity pressures, which eventually limit that expansion. Those limits are felt at different levels in different ways.
Think about food. If you have fertile land in good climate to produce food for 10,000,000 people, and have a population of 5,000,000, what happens if you raise your population by 20%? You add 1,000,000 people. You've still got the same specifications for making food: if you have to expend 10% of your population to make food, then before you had 500,000 people working on food production (farmers, fertilizer chemists, shipping, tractor makers, tool makers, etc.), and now you have an additional 100,000.
With that expansion, there's 20% more food, 20% more people, and 20% more hours of each type of labor going into making the food. The cost per unit of food required for each person is unchanged.
Now what happens if you have 10,000,000 people and bump by 20%?
You now have 2,000,000 more people to feed, and you have to farm on less-viable land. You need more irrigation and more fertilizer. You get lower yield, so need to farm over a wider span of land. That means more farm hands, more seeding, and an increase in all inputs (e.g. you need more fuel for the tractors, more water irrigating that whole span, twice the fertilizer to handle twice the land area, etc.).
Up to now, 2 million people required 200,000 laborers to make food. Now it requires 400,000 for these additional 2 million mouths to feed. That means the marginal cost of food is higher--in total, 11.67% instead of 10% of your population works on making food.
That means 1.67% of your income which was spent on other things is instead spent on food production. Those other things aren't made (less wealth) because they can't be bought. For rich people, this is essentially-unimportant: food requirements are generally constant, and rich people buy more-expensive food (pay for additional luxury) and so have both flexibility and an existing deep investment in luxury--and they still pay a very small portion of their income for food, even if they're eating caviar and lobster.
As you get into middle-class and poor, this increase in the cost of food reduces wealth substantially. The middle-class feel poorer; the poorest can't afford to eat. Because of this pressure, they also don't have the capacity to rear families, and will tend to slow down population growth.
As shown by history, resolving this scarcity pressure causes a population increase. This has been demonstrated as recently as 2006, where the recession caused slowed population growth in the United States, and the reduction of unemployment lead to a notable but small increase in population growth (see 2008-2012 versus the employment-population ratio and the unemployment rate as indicators of factors impacting how Americans perceive their access to financial stability). Back in the early 1900s, scarcity of food in particular lead to development of new fertilizers and intensive farming techniques; since the 1950s, world population has been on a sharper upward trend. It keeps happening.
A sharp increase in the sense of stability among family-minded Americans who would like to start a family or have a larger family but who don't feel they can afford it right now will lead to the obvious: sudden financial stability, the perceived capacity to enlarge their family as they've always dreamed, and more children. So it is, so it's always been.
Rich people take vacations and private jet flights. Middle-classers just up-size their house from 982sqft (1950 average new single-family home size) to 2,300sqft (2000 size) and try to figure on when they can play all their XBox games.
Generally, we've gotten bigger houses and apartments to store our computers, tablets, kitchen appliances, washing machines, 6 TVs (one in every room), guitars, bicycles, fancy lamps, and so forth. We've also started to eat at McDonalds a lot, because who has time to cook? If we cooked, we couldn't watch all that anime on Netflix.
"The jobs are going away, folks. The jobs are going away. *waves pinched fingers* *shakes titties*" -- Donald J. Trump
I still want to transition to 32-hour work weeks, or maybe as low as 28. That's one of the side goals of my Universal Social Security plan: the efficiency improvements would normally lead to sharp population expansion, but I'd rather stall wealth growth in favor of shorter working hours. At a point, being wealthier doesn't really help you, because you buy all these toys but you work all the time and don't have any leisure hours to play with them.
XBox 360, XBox 360 Elite, XBox 360 Arcade, XBox 360 S [250/4/320]GB models, XBox 360 Pro, XBox 360 Super Elite, XBox 360 Core. The Arcade has 256MB or 512MB of on-board memory; some models don't handle HDMI; and the low-storage models have to read from disc which, for some games, takes a minute or two to load large sets of assets instead of 10-15 seconds from hard drive.
That's the console's fault? If you want better performance, the power requirement is higher.
The Switch eats under 20 watts. The Wii U eats under 10 watts at full load. The Playstation 3 uses 170+ watts to play a movie. All of these consoles had the same resolution output and FPS within their generation as competitors. The Nintendo consoles tend to be more Japanese-centric, and target less photorealism via the use of brighter colors; that's not to say they haven't gone that route on the Wii U, although Playstation 3 titles used more-detailed textures instead of largely-airbrushed stuff. The production difference to accomplish the same visual style is minor; the technical capability is close, although the PS3 will be able to do more real-time lighting and shader effects than the Wii U, among other things.
What's left is basically gameplay. The games work on either console; one just works while eating a shitload of electricity. Some consoles overheat to the point of backing out their GPUs.
Try turning it off.
You mean unplugging it? "Off" on modern consoles means "talk to the Internet to get alerts and updates, and install things in the background." Even the damned Switch consumes 9 watts when powered off. The Wii U consumes 0.4 watts while off; the XBox One consumes 12.9W; the XBox 360 (Wii U generation) consumed nearly 3 watts. That means the XBox One actually consumes more power than a Wii U under full load when the console is "turned off"; and Nintendo decided this was okay with the Switch since nobody complained to Sony and Microsoft.
My point was that modern titles with OpenGL shaders, real-time lighting, and the like run at around 30fps for many games, notably games written in Unity 3D; while you'll hit games with hardly anything technically-impressive (e.g. Hollow Knight) that can't run at low-resolution without straining to get more than 4fps during action scenes. As a general rule, a Unity 3D game will run with much less resource consumption and a higher frame rate to provide similar visual effects of technical complexity to a really slow, choppy, resource-intensive Unreal Engine game.
It's not even just Unity. Things with custom engines, things on Cardboard or SDL, things that aren't UE seem to just run fine. Granted, anything that's got huge amounts of visual effects, highly-complex shaders, particle fields, and 16x anti-aliasing is going to go way, way past what my Intel HD graphics can handle; I just expect that to be the tone. It shouldn't be luck whether or not a game with comparable visual effects requires a $6,000 graphics card or not.
That sounds so much better than clicking on the task bar to select a running application
One of those 1cm-wide blocks that expands into a menu of things? Oh, right, every desktop stopped having task bar buttons and moved to just showing an icon that may or may not expand a menu.
on the shortcuts to start a new instance of something I use regularly
I actually have Chrome, Xchat-gnome, and the terminal on the left side in the Activities view; the dock isn't on my screen normally, because it's a waste of space. Thing is, it's faster for me to just tap the Windows key, type, and hit enter than to reach over, grab the mouse, move to a dock pinned to the left of the screen (Unity), and click an icon. If I'm just logging in and want to start 6 applications across 4 desktops, I have to go into the Activities view anyway, and can click-click-click the shortcut icons if I want to bother.
or quickly finding it in the start menu if it's not something I
I didn't even know I had a DVD burner. I typed "DVD" and it showed me Thoggin.
I wasn't asking you if you used Windows 8 there, I was pointing out that Windows 8 is no longer current, having been supplanted with a version of Windows, Windows 10, that uses a classic desktop,
Windows 10 is less like the "classic" desktop than Windows 7. People here with Windows 10 are re-configuring it to try and look more like older Windows, and people with 7 have been downloading Classic Shell. Yes, people are running away from the Windows 7 and Windows 10 interfaces.
so making comments about Classic Desktop declining in popularity as evidence that non-desktop environments rule the roost is, well, not smart
I was pointing out they're a loud minority. They're a small sample who are malcontent and so noisy, while most people who are fat and happy don't care to bother arguing with idiots to explain that what they're getting without putting in any effort is great as far as they can tell. People who want to bitch are noisy.
When you have a million happy users and thirty who are pissed off, you have a forum flooded with whining and wargablers. It looks like the whole world is having a nuclear meltdown.
you've found half a dozen keyboard shortcuts that work great in GNOME 3 and make it almost usable
I've found "Windows Key", "Typing" (as in text, like "Firefox" or "Web browser"), and "Ctrl+Alt+arrow" (the same key combination that switches desktops in Gnome 2 back in 2001). I'm sure you'd be bewildered if I told you about this secret, magical spell called "Alt+Tab" too--another keyboard shortcut that nobody knows about, except us smart people who dig through the source code.
You can adapt patterns. Typically, if the tool doesn't work, you're solving the problem wrong; frequently, the problem benefits from an adapted approach; and sometimes, you have a problem so novel there isn't a specialized tool that fits it, and wedging everything else into it is painful.
Think like how you can open a paint can with a screwdriver, but they also make painter's tools with a lever to open a paint can. A screwdriver works; a painter's tool works with less effort, more control, and no visible damage to the can and lid. Often times we've invented general-purpose tools as such to replace the screwdriver, and programmers will try to break from the pattern because they think they can engineer an even-more-specialized paint lid remover. That's a bloody waste of time; that problem's already optimized in a way substantially-similar to the proposal. We minimize the number of tools to the set that provides the fewest trade-offs for the most functionality.
There's a reason software quality did not drastically improve with the "invention" of design patterns
Actually, modern software is practically-impossible to write without the modern design patterns people have developed over time. Back in the day, we had 64KB of working memory; now we have hundreds of loaded libraries and eighty-million-line programs with tens of thousands of moving parts. It quickly becomes an unmaintainable mess when everything is directly calling unorganized functions and exposing unique interfaces.
Because we can, we do: software quality was just fine in 1991, and we can write much-more-complex programs with just as much bugginess and instability as Windows 3.1. Nobody thinks software quality is unacceptable; they think they want to write more-impressive software. This is also why CPUs went from a small heat sink to a giant fan and aluminum radiator assembly the size of your fist, even though running a Core i7 at 90% of its rated speed can get you nearly fanless (I used to run an Athlon 64 1.9GHz that would overheat and shut down in under 10 seconds due to a dying fan and dust-clogged heat sink; I ran it at 1.8GHz and it was never over 60C).
In programming, we have design patterns. Object-oriented programming design patterns are, at current state-of-the-art, largely the GoF patterns originally described in Java, but now used in PHP, C#, and so forth. These are building blocks to engineer complex programs, based on foundational concepts like encapsulation and polymorphism. In essence, design patterns are the kinds of architectural components which repeatedly arise when people try to find the best way to do something, and allow you to enclose pieces of programs from other pieces in ways which make functional sense, such that you have an interface which must do a logical thing (and so you can assume--fixing the innards if it doesn't) and a body of code sufficiently-small to actually make correct in a meaningful way (you plug hundreds of these together to make a large thing which should be more-correct than one giant pile of complex, intertwined muck with no visible borders).
The field of programming is full of people who actively claim design patterns are just bullshit and restrict creativity, and that "any good programmer" can write correct code without all that crap. These are also the same people who don't seem to grasp that there are several distinct types of languages, data structures, databases (hierarchical, relational, document, node/graph, etc.), and so forth, and try to one-size-fits-all their favorite language and library into everything.
I'm sure, as an electrician, you can figure on what happens when a programmer just does things "creatively" instead of holding those discussions with highly-skilled programmers and mathemeticians at large. It's the same thing that happens when an electrician gets an idea on how to wire a 4-way switch circuit using only 3-way switches (hint: set it up so that the "off"configurations have hot-to-hot wiring into the light socket so no current flows). Engineers discuss these things and decide what's reasonable and what would work but is absolute horse shit and should never be done.
There are a lot of people who are seen as good programmers because they know and understand a whole code base, so they didn't have to architect that code any good. What you have is a load of garbage nobody can understand, barely-maintainable by one particular person, that appears to work but is an absolute nightmare.
Yeah I liked the springy, wobbly windows and all. It went away in favor of still using massive 3D resources, but not doing anything worth a damn except technically compositing on 3D flip buffers with 3D-generated shadows to present exactly the same visual effects as ever with inflated hardware requirements.
If you needed this for verification of Facebook chat not being private, you should probably get a refund for the two defective braincells rattling around in your head. You can see your entire chat history from any device; how do you think that works?
I've got a 4790k Intel Core i7 with HD Graphics. It plays 3D games alright, but Hollow Knight struggles. That is to say: Unity 3D games with 2D or 3D graphics and complex, shiny effects can get 24fps at 1900x1080; Unreal Engine games with high-resolution sprites and few real-time effects struggle and stutter at 1200x800 resolution, making platforming nigh-impossible and rendering games unplayable. Not that Hollow Knight doesn't look as good as any Unity 3D game--the art style is fantastic--it's that Hollow Knight doesn't rely on fancy, high-end graphics effects, while many Unity 3D games use complex shaders and real-time lighting and all kinds of heavy shit and still get higher performance and a high-end look.
I'll never understand the Unreal Engine fetish. Maybe it's just that Unity 3D is easy and a lot of people pick it up and make quick-and-dirty games that don't have a professional feel, while UE has fewer first-time indie devs making Flappy Bird clones and so showcases more recognizable, high-end games. People think they'll be able to do a professional job if only they use the same tool, even though they're about as skilled as the guy who made a half-working, crashy Flappy Bird clone.
There was a time when you paid $200 to get a console because it was $200 and games worked. You could play games on PC, except you had to futz around with autoexec.bat and config.sys to exclude drivers so you had 770k of conventional RAM or DOOM wouldn't run. Eventually, you could play games on PC, but the games cost $60 and you had to spend $400 on upgrading your video card every 4-6 months, so you spent a lot just to play games. $200 for a 10-watt console playing $20-$50 games for the next 3-5 years was just good sense: it worked and it didn't cost assloads for non-game bits all the time.
Now you spend $400 on a console that has 6 separate models, so your version might not play the game without long-ass load times. Then they release a new revision of the console, so you have to spend $600 more for the new one. Then they make it modular, so you have to keep buying RAM and GPU modules to upgrade for $200 every several months. The console also eats 245W of power, so playing for 4 hours per day is like an extra 15% on your electricity bill. Then the console idles at 30W all the time, and kicks on full-power for about 20 hours per week to download updates, and it ends up costing you $12/month just to have it.
Dual Channel gives less than a 5% performance increase over single-channel RAM. It's almost always better to skip the DC kit in favor of just more RAM.
.if you want it to do a fairly limited subset of things, mostly involving a full-screen only desktop.
I want the DE out of my way. You talk like you want to waste your time playing with the DE instead of using it as a vehicle to get your hands on the applications you want to work with.
Cinnamon is a traditional desktop. It can have as many virtual desktops as you want
Yes, it can have as many of a fixed number of virtual desktops as you want. Oh, you have 3x3 but you need 10? Better re-configure your virtual desktop setup to be 3x4 so you have 12. In Gnome 3, you just start using a new virtual desktop.
If by "Activities" you mean "Applications", they're either a single click or a categorized start menu away. They're not buried in massive full-screen giant icon panels.
No, in Gnome 3, there's an "Activities" view. If you touch the top-left corner or press that Windows key Microsoft got stuck on all keyboards nowadays, it suddenly takes all your windows on the current desktop and shows them in what's essentially a MacOS Expose view. If you then start typing, it starts searching through the command names, icon labels, and keywords associated with your applications, showing you the matching applications; if you hit the Enter key, it executes the first one, which is often the most-recent one you've run that's in the results.
That means I can get Gnome Terminal up by tapping Meta with my pinky and typing "ter" and hitting enter. I can get Xchat-gnome running the same way. Applications aren't buried in menu trees or whatever you're babbling about with giant icon panels; they're at the beck and call of a fleeting thought and a brief twitch of my little finger. Gnome 3 quite nearly simply responds to my will and desire, and would do exactly that if I could eliminate the 3-4 keystrokes and 20-30 milliseconds required to actually express what I'm thinking.
You can switch between your open windows by clicking them in this view. You can drag them to another desktop--they're all shown on the right side of the Activities view. There's a new desktop at the bottom of all that; or you can stick a window between two existing desktops and it'll insert a new, empty desktop there. Ctrl+Alt+up/down moves you through desktops--Activities view or not.
Microsoft and replaced with Windows 10, which you should try out some time - it essentially reverts back to the traditional desktop, complete with start menu
Actually, we don't have Windows 8 here; we're running Windows 7, and upgrading to Windows 10 soon. Windows 7 users have been using an open-source software called Classic Shell to get a Windows-XP or Windows-2000/95 start menu in Windows 7.
Windows 10 has a nice feature where you press Win+Tab and it gives you an Expose view like Gnome Shell. You can add (horizontal) desktops, although you can't create a new desktop between two existing ones, and need to manually add a new desktop to the end. Ctrl+Win+Left/Right moves between desktops in either view.
The Windows 10 start menu is just an alphabetical list of applications in a rolling menu, too. There are some folders, some controls, and a Windows-8-style panel area on the right side. Microsoft has a Cortana search bar next to the start menu which functions like Gnome Shell or Unity Desktop search, kind of.
That means Windows 10 requires a lot of mouse work to use the virtual desktops and to start applications; Gnome 3 allows that, but will let you get things running with just the keyboard, up to and including the creation of new desktops. If you want to move windows between desktops or create desktops between other desktops, you need to pull out the mouse.
It sounds exactly like you've never seen Gnome 3 or Windows 10. Maybe you're just full of shit.
The problem is "passionate nerds" are basically people who write things like XML::Bare; the folks who write actual useful shit are both interested in their field and well-trained, either by a college education or by an independent formal study. The people living the delusion that the best tech workers are self-trained all think they're programmers because they've read the PHP documentation online and gotten Apache to belch out a Web page; they typically disdain books and formal practices because "it's all obvious if you've got actual talent" and prattle on about how programming is a specific genetic trait you're born with.
Programmers need to be able to solve discrete problems, but they don't necessarily need to be high-level software architects. They do need to be engineers.
Also these are H1-C Visas: instead of a $50k IIT graduate, we have a $30k high school graduate. Once you OJT them, they're just as good; and you can give them slow raises, 2% or so, instead of bumping them by $30k. Since they don't have degrees, you can hire somebody else's trained monkey for peanuts, too.
GNOME 3 has a faster workflow than Cinnamon or Unity Desktop. You can tap the top-left corner or press the Meta key and get a view of all your windows on the current desktop; create a new desktop by dragging a window between any desktops; and pull up an application by bringing up the Activities view and just typing. It pretty much gets out of your way.
Unity is basically 4x4 with zoomed-out view of all desktops at once, cluttered together, with a dock and an extra click to pull out a search bar. Cinnamon has a fixed virtual desktop arrangement and (last I checked) no Activities view, so you're back to spending 6 seconds navigating (well-arranged) menus instead of 1 second to open whatever application you were thinking.
The world is moving away from Windows 95 desktop interfaces. There's a loud minority installing "Classic Shell" on Windows, and they get way too much attention.
Well public transit is usually slow and limited because one driver can transport many passengers. What's the efficiency of Uber over, well, Uber?
Well, hold on, my original argument spoke of time spent on the bus or rail system.
A single light rail ride here ranges from having a car containing 3 passengers to three cars containing 40 passengers. Obviously, the light rail using a driver to move 3 passengers is slow, clunky, and inefficient: it takes 49 minutes of driver time to move those 3 people, when the driver could spend those 49 minutes moving 120 passengers.
What happens when we include the passengers, though?
A 3-passenger, 49-minute trip involves a driver driving for 49 minutes and three passengers riding for 49 minutes; they could also end up waiting 25 minutes for the train to appear, but let's ignore that because it's complex and involves position in time rather than length of time (it's more risk of missing the train than direct time cost, so the cost is hard to compute). For a 17-minute drive, that's 32 additional minutes lost per passenger--a productivity loss of 96 minutes--plus the 49 minutes of the driver, totaling 125 minutes.
This compares to Uber by requiring 17 minutes of driver time plus 17 minutes of passenger time. With sufficient Uber drivers and sufficient passenger load, the Uber driver would (optimally) find another local hail near his drop-off point. Call that 3 minutes of loss per trip, and you have an optimum time cost of 0 minutes passenger (who is home and can continue to perform minor housework for 3 minutes) and 20 minutes driver, although we can assume the passenger dawdles for the 3 minutes wait and call it 23 minutes. For three trips, that's 69 minutes.
When we get to a fully-loaded, 120-passenger cars, we're talking about 49 minutes of driver time to carry 120 passengers losing 32 minutes each. That's 3,889 minutes of lost productivity for our light rail system, at a minimum--49 minutes driver and 3,840 passenger.
Uber compares to this with a total productivity cost of 2,760 minutes compared to the passenger driving himself. That's 2,400 of driver time and 360 minutes of passenger time spent waiting for the Uber.
Basically it seems like a scheme to pay taxes to pay Uber instead of paying Uber to provide the exact same taxi service.
Public transit usually carries a taxpayer cost unreflected by operational pricing. Baltimore City's light rail draws $15.6 million of revenue, while the light rail transit cost is $38.6 million. The MTA bus system in the Baltimore regional area similarly draws about $166 million, bringing a cost of $302 million per year.
The MTA Regular Bus service in the Baltimore region operates roughly 21.6 million revenue-miles at a cost of, roughly, $14.015 per mile on which zero or more passengers are driven (the full costs include miles in which passengers are not driven--not an empty bus that just has no one, but a bus out of service going back to staging). The per-mile cost of my 13-mile drive to work via Uber at $17 is $1.30.
The bus system actually has a ridership of 3 million boarding per year (a round-trip is two boardings). With an average commute distance of 19 miles, that's $74 million per year for Baltimore City to cover the full fare of Uber rather than to use MTA City Bus. To restrict tickets to $1.60, it's $69 million.
So maybe they've got a scheme to pay $75,000 to Uber instead of paying $300,000 for buses?
The whole idea of taxes for transit is people making $6,000/year don't pay $30/day to take public transit; they pay $3/day.
Username logged into the machine, XBox IDs (that Microsoft already has), and command lines (of applications usually run through GUIs that are trying to elevate privileges).
People are going to continue screaming about Microsoft constantly recording your microphone, cameras, keystrokes, mouse movements, screen captures, and uploading all of your files and gmail to their analytics server, because of course. For now, they'll try to squeeze something out about usernames and XBox IDs being extra-secret.
Wait until they realize people are logging onto Windows 10 using an outlook.com user account, and log-on events cause the machine to contact Microsoft to validate your password.
In Baltimore, it's US $3.60 for a round-trip, US $1.80 for a one-way, and US $3.95 for a day pass on all public transit.
Here's the thing: generally, public transit is slow and limited. The light rail is on a rail, and takes three times as long as driving if you include the 15-45 minute wait (sometimes trains come 40 minutes in between--and, hell, sometimes they alternate so every other train goes all the way, so the next train may stop short of your destination and require you to get off and wait another 25 minutes to continue). Buses can take 3 hours to make a 25-minute car trip.
Think about the lost wages and lost productivity riding around for 3 hours when you could drive for 20 minutes.
Now imagine if, for lower tax costs, you could just get a single-rider, direct transport to your destination. A car ride. My ride from home to work is 49 minutes by light rail, right now, for $1.80; it's $17-$23 by Uber or Lyft, and only 17 minutes. The east-west Light Rail Red Line is estimated at $2.1 trillion installation cost, and so was to be installed in segments over 20 years, with the cheap parts (on-grade, rather than underground tunnels or above-grade rail) built first to start generating economic benefit on a hundred billion or so per year; we couldn't get the Federal funding.
So Canada thinks they can spend $75,000 per year of money they take out of your income so you can further pay $3.50 for a one-way trip that takes 20 minutes, instead of paying $20 for that same trip. Their alternative is spending several billions for a bus system or hundreds of billions for a rail system with long transit times. Seems legit.
Inflation over 10 years should be 2% per year, or 21%. That $20,000 should be $24,379. With technical progress, the cost of education should come down (the share of education fees associated with tuition will reflect class size and professor salaries; the share associated with facilities management overhead should decrease due to improved technology reducing the amount of labor required and, thus, the total salary paid per student-year to the collection of people who aren't professors). That means the $20,000 should be somewhat less than $24,379.
Because the costs should come down for the entire cost of tuition, the proportion of tuition taken as debt should also shrink for students with an active income during their vocational training. This further reduces the expected exit debt.
Some landlords with 6-12 units have said a unit that stays empty an extra month is their entire year's profit down the drain. That's why they're aggressive and make you keep paying if you move before your lease, and recruit exiting tenants to find a new tenant. They'll start looking for a new tenant 3 months before your lease is up, and kick you straight out if you don't renew right away when your lease is up.
Slashdot Mirror
Gross margin? Huge.
Net margin? It costs several billion to get drugs to market. All that FDA safety and efficacy stuff is about $310 million, from lab rat studies over 3 years to a 3-year follow-up with patients to document side-effects and efficacy once the drug is out in the market. Only about 10% of new drugs make it to market, and a lot goes into the pipeline and doesn't even make it to being a "new drug". The sunk research costs are $3-$5 billion per drug.
For these drugs, they have to recover $3-$5 billion to break even. With a small target market, the gross margin is going to be a lot bigger to make that up. This is one of the reasons cognitive enhancement is medically ethical: something like Modafinil treats narcolepsy, but would be universally-useful to practically everyone due to its safety and efficacy as a non-addictive eugerogic (let you pull an all-nighter now and then) or study drug. The market becomes enormous, and the amount of money you can make on 60 cents per pill is ludicrous (it only costs about 11 cents to manufacture a 100mg dose). As it stands, it might cost you $12 a pill in America, although I can get it out-of-pocket for $240 for a 90-day supply from my PBM without insurance coverage.
Drugs with broad uses tend to become cheap and generic.
No, all evidence shows that when an individual human unit gets wealthier, it slows its rapid breeding. Humans don't need to produce more children when fewer die out, either, as a society.
As a population, a human society includes a gradient of wealth. The expansion of that society causes scarcity pressures, which eventually limit that expansion. Those limits are felt at different levels in different ways.
Think about food. If you have fertile land in good climate to produce food for 10,000,000 people, and have a population of 5,000,000, what happens if you raise your population by 20%? You add 1,000,000 people. You've still got the same specifications for making food: if you have to expend 10% of your population to make food, then before you had 500,000 people working on food production (farmers, fertilizer chemists, shipping, tractor makers, tool makers, etc.), and now you have an additional 100,000.
With that expansion, there's 20% more food, 20% more people, and 20% more hours of each type of labor going into making the food. The cost per unit of food required for each person is unchanged.
Now what happens if you have 10,000,000 people and bump by 20%?
You now have 2,000,000 more people to feed, and you have to farm on less-viable land. You need more irrigation and more fertilizer. You get lower yield, so need to farm over a wider span of land. That means more farm hands, more seeding, and an increase in all inputs (e.g. you need more fuel for the tractors, more water irrigating that whole span, twice the fertilizer to handle twice the land area, etc.).
Up to now, 2 million people required 200,000 laborers to make food. Now it requires 400,000 for these additional 2 million mouths to feed. That means the marginal cost of food is higher--in total, 11.67% instead of 10% of your population works on making food.
That means 1.67% of your income which was spent on other things is instead spent on food production. Those other things aren't made (less wealth) because they can't be bought. For rich people, this is essentially-unimportant: food requirements are generally constant, and rich people buy more-expensive food (pay for additional luxury) and so have both flexibility and an existing deep investment in luxury--and they still pay a very small portion of their income for food, even if they're eating caviar and lobster.
As you get into middle-class and poor, this increase in the cost of food reduces wealth substantially. The middle-class feel poorer; the poorest can't afford to eat. Because of this pressure, they also don't have the capacity to rear families, and will tend to slow down population growth.
As shown by history, resolving this scarcity pressure causes a population increase. This has been demonstrated as recently as 2006, where the recession caused slowed population growth in the United States, and the reduction of unemployment lead to a notable but small increase in population growth (see 2008-2012 versus the employment-population ratio and the unemployment rate as indicators of factors impacting how Americans perceive their access to financial stability). Back in the early 1900s, scarcity of food in particular lead to development of new fertilizers and intensive farming techniques; since the 1950s, world population has been on a sharper upward trend. It keeps happening.
A sharp increase in the sense of stability among family-minded Americans who would like to start a family or have a larger family but who don't feel they can afford it right now will lead to the obvious: sudden financial stability, the perceived capacity to enlarge their family as they've always dreamed, and more children. So it is, so it's always been.
Rich people take vacations and private jet flights. Middle-classers just up-size their house from 982sqft (1950 average new single-family home size) to 2,300sqft (2000 size) and try to figure on when they can play all their XBox games.
Generally, we've gotten bigger houses and apartments to store our computers, tablets, kitchen appliances, washing machines, 6 TVs (one in every room), guitars, bicycles, fancy lamps, and so forth. We've also started to eat at McDonalds a lot, because who has time to cook? If we cooked, we couldn't watch all that anime on Netflix.
"The jobs are going away, folks. The jobs are going away. *waves pinched fingers* *shakes titties*" -- Donald J. Trump
I still want to transition to 32-hour work weeks, or maybe as low as 28. That's one of the side goals of my Universal Social Security plan: the efficiency improvements would normally lead to sharp population expansion, but I'd rather stall wealth growth in favor of shorter working hours. At a point, being wealthier doesn't really help you, because you buy all these toys but you work all the time and don't have any leisure hours to play with them.
2 models.
XBox 360, XBox 360 Elite, XBox 360 Arcade, XBox 360 S [250/4/320]GB models, XBox 360 Pro, XBox 360 Super Elite, XBox 360 Core. The Arcade has 256MB or 512MB of on-board memory; some models don't handle HDMI; and the low-storage models have to read from disc which, for some games, takes a minute or two to load large sets of assets instead of 10-15 seconds from hard drive.
That's the console's fault? If you want better performance, the power requirement is higher.
The Switch eats under 20 watts. The Wii U eats under 10 watts at full load. The Playstation 3 uses 170+ watts to play a movie. All of these consoles had the same resolution output and FPS within their generation as competitors. The Nintendo consoles tend to be more Japanese-centric, and target less photorealism via the use of brighter colors; that's not to say they haven't gone that route on the Wii U, although Playstation 3 titles used more-detailed textures instead of largely-airbrushed stuff. The production difference to accomplish the same visual style is minor; the technical capability is close, although the PS3 will be able to do more real-time lighting and shader effects than the Wii U, among other things.
What's left is basically gameplay. The games work on either console; one just works while eating a shitload of electricity. Some consoles overheat to the point of backing out their GPUs.
Try turning it off.
You mean unplugging it? "Off" on modern consoles means "talk to the Internet to get alerts and updates, and install things in the background." Even the damned Switch consumes 9 watts when powered off. The Wii U consumes 0.4 watts while off; the XBox One consumes 12.9W; the XBox 360 (Wii U generation) consumed nearly 3 watts. That means the XBox One actually consumes more power than a Wii U under full load when the console is "turned off"; and Nintendo decided this was okay with the Switch since nobody complained to Sony and Microsoft.
My point was that modern titles with OpenGL shaders, real-time lighting, and the like run at around 30fps for many games, notably games written in Unity 3D; while you'll hit games with hardly anything technically-impressive (e.g. Hollow Knight) that can't run at low-resolution without straining to get more than 4fps during action scenes. As a general rule, a Unity 3D game will run with much less resource consumption and a higher frame rate to provide similar visual effects of technical complexity to a really slow, choppy, resource-intensive Unreal Engine game.
It's not even just Unity. Things with custom engines, things on Cardboard or SDL, things that aren't UE seem to just run fine. Granted, anything that's got huge amounts of visual effects, highly-complex shaders, particle fields, and 16x anti-aliasing is going to go way, way past what my Intel HD graphics can handle; I just expect that to be the tone. It shouldn't be luck whether or not a game with comparable visual effects requires a $6,000 graphics card or not.
That sounds so much better than clicking on the task bar to select a running application
One of those 1cm-wide blocks that expands into a menu of things? Oh, right, every desktop stopped having task bar buttons and moved to just showing an icon that may or may not expand a menu.
on the shortcuts to start a new instance of something I use regularly
I actually have Chrome, Xchat-gnome, and the terminal on the left side in the Activities view; the dock isn't on my screen normally, because it's a waste of space. Thing is, it's faster for me to just tap the Windows key, type, and hit enter than to reach over, grab the mouse, move to a dock pinned to the left of the screen (Unity), and click an icon. If I'm just logging in and want to start 6 applications across 4 desktops, I have to go into the Activities view anyway, and can click-click-click the shortcut icons if I want to bother.
or quickly finding it in the start menu if it's not something I
I didn't even know I had a DVD burner. I typed "DVD" and it showed me Thoggin.
I wasn't asking you if you used Windows 8 there, I was pointing out that Windows 8 is no longer current, having been supplanted with a version of Windows, Windows 10, that uses a classic desktop,
Windows 10 is less like the "classic" desktop than Windows 7. People here with Windows 10 are re-configuring it to try and look more like older Windows, and people with 7 have been downloading Classic Shell. Yes, people are running away from the Windows 7 and Windows 10 interfaces.
so making comments about Classic Desktop declining in popularity as evidence that non-desktop environments rule the roost is, well, not smart
I was pointing out they're a loud minority. They're a small sample who are malcontent and so noisy, while most people who are fat and happy don't care to bother arguing with idiots to explain that what they're getting without putting in any effort is great as far as they can tell. People who want to bitch are noisy.
When you have a million happy users and thirty who are pissed off, you have a forum flooded with whining and wargablers. It looks like the whole world is having a nuclear meltdown.
you've found half a dozen keyboard shortcuts that work great in GNOME 3 and make it almost usable
I've found "Windows Key", "Typing" (as in text, like "Firefox" or "Web browser"), and "Ctrl+Alt+arrow" (the same key combination that switches desktops in Gnome 2 back in 2001). I'm sure you'd be bewildered if I told you about this secret, magical spell called "Alt+Tab" too--another keyboard shortcut that nobody knows about, except us smart people who dig through the source code.
You can adapt patterns. Typically, if the tool doesn't work, you're solving the problem wrong; frequently, the problem benefits from an adapted approach; and sometimes, you have a problem so novel there isn't a specialized tool that fits it, and wedging everything else into it is painful.
Think like how you can open a paint can with a screwdriver, but they also make painter's tools with a lever to open a paint can. A screwdriver works; a painter's tool works with less effort, more control, and no visible damage to the can and lid. Often times we've invented general-purpose tools as such to replace the screwdriver, and programmers will try to break from the pattern because they think they can engineer an even-more-specialized paint lid remover. That's a bloody waste of time; that problem's already optimized in a way substantially-similar to the proposal. We minimize the number of tools to the set that provides the fewest trade-offs for the most functionality.
There's a reason software quality did not drastically improve with the "invention" of design patterns
Actually, modern software is practically-impossible to write without the modern design patterns people have developed over time. Back in the day, we had 64KB of working memory; now we have hundreds of loaded libraries and eighty-million-line programs with tens of thousands of moving parts. It quickly becomes an unmaintainable mess when everything is directly calling unorganized functions and exposing unique interfaces.
Because we can, we do: software quality was just fine in 1991, and we can write much-more-complex programs with just as much bugginess and instability as Windows 3.1. Nobody thinks software quality is unacceptable; they think they want to write more-impressive software. This is also why CPUs went from a small heat sink to a giant fan and aluminum radiator assembly the size of your fist, even though running a Core i7 at 90% of its rated speed can get you nearly fanless (I used to run an Athlon 64 1.9GHz that would overheat and shut down in under 10 seconds due to a dying fan and dust-clogged heat sink; I ran it at 1.8GHz and it was never over 60C).
I built a system to run a programming environment, and it happened to run games fairly well.
In programming, we have design patterns. Object-oriented programming design patterns are, at current state-of-the-art, largely the GoF patterns originally described in Java, but now used in PHP, C#, and so forth. These are building blocks to engineer complex programs, based on foundational concepts like encapsulation and polymorphism. In essence, design patterns are the kinds of architectural components which repeatedly arise when people try to find the best way to do something, and allow you to enclose pieces of programs from other pieces in ways which make functional sense, such that you have an interface which must do a logical thing (and so you can assume--fixing the innards if it doesn't) and a body of code sufficiently-small to actually make correct in a meaningful way (you plug hundreds of these together to make a large thing which should be more-correct than one giant pile of complex, intertwined muck with no visible borders).
The field of programming is full of people who actively claim design patterns are just bullshit and restrict creativity, and that "any good programmer" can write correct code without all that crap. These are also the same people who don't seem to grasp that there are several distinct types of languages, data structures, databases (hierarchical, relational, document, node/graph, etc.), and so forth, and try to one-size-fits-all their favorite language and library into everything.
I'm sure, as an electrician, you can figure on what happens when a programmer just does things "creatively" instead of holding those discussions with highly-skilled programmers and mathemeticians at large. It's the same thing that happens when an electrician gets an idea on how to wire a 4-way switch circuit using only 3-way switches (hint: set it up so that the "off"configurations have hot-to-hot wiring into the light socket so no current flows). Engineers discuss these things and decide what's reasonable and what would work but is absolute horse shit and should never be done.
There are a lot of people who are seen as good programmers because they know and understand a whole code base, so they didn't have to architect that code any good. What you have is a load of garbage nobody can understand, barely-maintainable by one particular person, that appears to work but is an absolute nightmare.
Yeah I liked the springy, wobbly windows and all. It went away in favor of still using massive 3D resources, but not doing anything worth a damn except technically compositing on 3D flip buffers with 3D-generated shadows to present exactly the same visual effects as ever with inflated hardware requirements.
If you needed this for verification of Facebook chat not being private, you should probably get a refund for the two defective braincells rattling around in your head. You can see your entire chat history from any device; how do you think that works?
I've got a 4790k Intel Core i7 with HD Graphics. It plays 3D games alright, but Hollow Knight struggles. That is to say: Unity 3D games with 2D or 3D graphics and complex, shiny effects can get 24fps at 1900x1080; Unreal Engine games with high-resolution sprites and few real-time effects struggle and stutter at 1200x800 resolution, making platforming nigh-impossible and rendering games unplayable. Not that Hollow Knight doesn't look as good as any Unity 3D game--the art style is fantastic--it's that Hollow Knight doesn't rely on fancy, high-end graphics effects, while many Unity 3D games use complex shaders and real-time lighting and all kinds of heavy shit and still get higher performance and a high-end look.
I'll never understand the Unreal Engine fetish. Maybe it's just that Unity 3D is easy and a lot of people pick it up and make quick-and-dirty games that don't have a professional feel, while UE has fewer first-time indie devs making Flappy Bird clones and so showcases more recognizable, high-end games. People think they'll be able to do a professional job if only they use the same tool, even though they're about as skilled as the guy who made a half-working, crashy Flappy Bird clone.
There was a time when you paid $200 to get a console because it was $200 and games worked. You could play games on PC, except you had to futz around with autoexec.bat and config.sys to exclude drivers so you had 770k of conventional RAM or DOOM wouldn't run. Eventually, you could play games on PC, but the games cost $60 and you had to spend $400 on upgrading your video card every 4-6 months, so you spent a lot just to play games. $200 for a 10-watt console playing $20-$50 games for the next 3-5 years was just good sense: it worked and it didn't cost assloads for non-game bits all the time.
Now you spend $400 on a console that has 6 separate models, so your version might not play the game without long-ass load times. Then they release a new revision of the console, so you have to spend $600 more for the new one. Then they make it modular, so you have to keep buying RAM and GPU modules to upgrade for $200 every several months. The console also eats 245W of power, so playing for 4 hours per day is like an extra 15% on your electricity bill. Then the console idles at 30W all the time, and kicks on full-power for about 20 hours per week to download updates, and it ends up costing you $12/month just to have it.
People talk about hardware and not about games.
Dual Channel gives less than a 5% performance increase over single-channel RAM. It's almost always better to skip the DC kit in favor of just more RAM.
.if you want it to do a fairly limited subset of things, mostly involving a full-screen only desktop.
I want the DE out of my way. You talk like you want to waste your time playing with the DE instead of using it as a vehicle to get your hands on the applications you want to work with.
Cinnamon is a traditional desktop. It can have as many virtual desktops as you want
Yes, it can have as many of a fixed number of virtual desktops as you want. Oh, you have 3x3 but you need 10? Better re-configure your virtual desktop setup to be 3x4 so you have 12. In Gnome 3, you just start using a new virtual desktop.
If by "Activities" you mean "Applications", they're either a single click or a categorized start menu away. They're not buried in massive full-screen giant icon panels.
No, in Gnome 3, there's an "Activities" view. If you touch the top-left corner or press that Windows key Microsoft got stuck on all keyboards nowadays, it suddenly takes all your windows on the current desktop and shows them in what's essentially a MacOS Expose view. If you then start typing, it starts searching through the command names, icon labels, and keywords associated with your applications, showing you the matching applications; if you hit the Enter key, it executes the first one, which is often the most-recent one you've run that's in the results.
That means I can get Gnome Terminal up by tapping Meta with my pinky and typing "ter" and hitting enter. I can get Xchat-gnome running the same way. Applications aren't buried in menu trees or whatever you're babbling about with giant icon panels; they're at the beck and call of a fleeting thought and a brief twitch of my little finger. Gnome 3 quite nearly simply responds to my will and desire, and would do exactly that if I could eliminate the 3-4 keystrokes and 20-30 milliseconds required to actually express what I'm thinking.
You can switch between your open windows by clicking them in this view. You can drag them to another desktop--they're all shown on the right side of the Activities view. There's a new desktop at the bottom of all that; or you can stick a window between two existing desktops and it'll insert a new, empty desktop there. Ctrl+Alt+up/down moves you through desktops--Activities view or not.
Microsoft and replaced with Windows 10, which you should try out some time - it essentially reverts back to the traditional desktop, complete with start menu
Actually, we don't have Windows 8 here; we're running Windows 7, and upgrading to Windows 10 soon. Windows 7 users have been using an open-source software called Classic Shell to get a Windows-XP or Windows-2000/95 start menu in Windows 7.
Windows 10 has a nice feature where you press Win+Tab and it gives you an Expose view like Gnome Shell. You can add (horizontal) desktops, although you can't create a new desktop between two existing ones, and need to manually add a new desktop to the end. Ctrl+Win+Left/Right moves between desktops in either view.
The Windows 10 start menu is just an alphabetical list of applications in a rolling menu, too. There are some folders, some controls, and a Windows-8-style panel area on the right side. Microsoft has a Cortana search bar next to the start menu which functions like Gnome Shell or Unity Desktop search, kind of.
That means Windows 10 requires a lot of mouse work to use the virtual desktops and to start applications; Gnome 3 allows that, but will let you get things running with just the keyboard, up to and including the creation of new desktops. If you want to move windows between desktops or create desktops between other desktops, you need to pull out the mouse.
It sounds exactly like you've never seen Gnome 3 or Windows 10. Maybe you're just full of shit.
The problem is "passionate nerds" are basically people who write things like XML::Bare; the folks who write actual useful shit are both interested in their field and well-trained, either by a college education or by an independent formal study. The people living the delusion that the best tech workers are self-trained all think they're programmers because they've read the PHP documentation online and gotten Apache to belch out a Web page; they typically disdain books and formal practices because "it's all obvious if you've got actual talent" and prattle on about how programming is a specific genetic trait you're born with.
Programmers need to be able to solve discrete problems, but they don't necessarily need to be high-level software architects. They do need to be engineers.
Also these are H1-C Visas: instead of a $50k IIT graduate, we have a $30k high school graduate. Once you OJT them, they're just as good; and you can give them slow raises, 2% or so, instead of bumping them by $30k. Since they don't have degrees, you can hire somebody else's trained monkey for peanuts, too.
GNOME 3 has a faster workflow than Cinnamon or Unity Desktop. You can tap the top-left corner or press the Meta key and get a view of all your windows on the current desktop; create a new desktop by dragging a window between any desktops; and pull up an application by bringing up the Activities view and just typing. It pretty much gets out of your way.
Unity is basically 4x4 with zoomed-out view of all desktops at once, cluttered together, with a dock and an extra click to pull out a search bar. Cinnamon has a fixed virtual desktop arrangement and (last I checked) no Activities view, so you're back to spending 6 seconds navigating (well-arranged) menus instead of 1 second to open whatever application you were thinking.
The world is moving away from Windows 95 desktop interfaces. There's a loud minority installing "Classic Shell" on Windows, and they get way too much attention.
Well public transit is usually slow and limited because one driver can transport many passengers. What's the efficiency of Uber over, well, Uber?
Well, hold on, my original argument spoke of time spent on the bus or rail system.
A single light rail ride here ranges from having a car containing 3 passengers to three cars containing 40 passengers. Obviously, the light rail using a driver to move 3 passengers is slow, clunky, and inefficient: it takes 49 minutes of driver time to move those 3 people, when the driver could spend those 49 minutes moving 120 passengers.
What happens when we include the passengers, though?
A 3-passenger, 49-minute trip involves a driver driving for 49 minutes and three passengers riding for 49 minutes; they could also end up waiting 25 minutes for the train to appear, but let's ignore that because it's complex and involves position in time rather than length of time (it's more risk of missing the train than direct time cost, so the cost is hard to compute). For a 17-minute drive, that's 32 additional minutes lost per passenger--a productivity loss of 96 minutes--plus the 49 minutes of the driver, totaling 125 minutes.
This compares to Uber by requiring 17 minutes of driver time plus 17 minutes of passenger time. With sufficient Uber drivers and sufficient passenger load, the Uber driver would (optimally) find another local hail near his drop-off point. Call that 3 minutes of loss per trip, and you have an optimum time cost of 0 minutes passenger (who is home and can continue to perform minor housework for 3 minutes) and 20 minutes driver, although we can assume the passenger dawdles for the 3 minutes wait and call it 23 minutes. For three trips, that's 69 minutes.
When we get to a fully-loaded, 120-passenger cars, we're talking about 49 minutes of driver time to carry 120 passengers losing 32 minutes each. That's 3,889 minutes of lost productivity for our light rail system, at a minimum--49 minutes driver and 3,840 passenger.
Uber compares to this with a total productivity cost of 2,760 minutes compared to the passenger driving himself. That's 2,400 of driver time and 360 minutes of passenger time spent waiting for the Uber.
Basically it seems like a scheme to pay taxes to pay Uber instead of paying Uber to provide the exact same taxi service.
Public transit usually carries a taxpayer cost unreflected by operational pricing. Baltimore City's light rail draws $15.6 million of revenue, while the light rail transit cost is $38.6 million. The MTA bus system in the Baltimore regional area similarly draws about $166 million, bringing a cost of $302 million per year.
The MTA Regular Bus service in the Baltimore region operates roughly 21.6 million revenue-miles at a cost of, roughly, $14.015 per mile on which zero or more passengers are driven (the full costs include miles in which passengers are not driven--not an empty bus that just has no one, but a bus out of service going back to staging). The per-mile cost of my 13-mile drive to work via Uber at $17 is $1.30.
The bus system actually has a ridership of 3 million boarding per year (a round-trip is two boardings). With an average commute distance of 19 miles, that's $74 million per year for Baltimore City to cover the full fare of Uber rather than to use MTA City Bus. To restrict tickets to $1.60, it's $69 million.
So maybe they've got a scheme to pay $75,000 to Uber instead of paying $300,000 for buses?
The whole idea of taxes for transit is people making $6,000/year don't pay $30/day to take public transit; they pay $3/day.
Username logged into the machine, XBox IDs (that Microsoft already has), and command lines (of applications usually run through GUIs that are trying to elevate privileges).
People are going to continue screaming about Microsoft constantly recording your microphone, cameras, keystrokes, mouse movements, screen captures, and uploading all of your files and gmail to their analytics server, because of course. For now, they'll try to squeeze something out about usernames and XBox IDs being extra-secret.
Wait until they realize people are logging onto Windows 10 using an outlook.com user account, and log-on events cause the machine to contact Microsoft to validate your password.
In Baltimore, it's US $3.60 for a round-trip, US $1.80 for a one-way, and US $3.95 for a day pass on all public transit.
Here's the thing: generally, public transit is slow and limited. The light rail is on a rail, and takes three times as long as driving if you include the 15-45 minute wait (sometimes trains come 40 minutes in between--and, hell, sometimes they alternate so every other train goes all the way, so the next train may stop short of your destination and require you to get off and wait another 25 minutes to continue). Buses can take 3 hours to make a 25-minute car trip.
Think about the lost wages and lost productivity riding around for 3 hours when you could drive for 20 minutes.
Now imagine if, for lower tax costs, you could just get a single-rider, direct transport to your destination. A car ride. My ride from home to work is 49 minutes by light rail, right now, for $1.80; it's $17-$23 by Uber or Lyft, and only 17 minutes. The east-west Light Rail Red Line is estimated at $2.1 trillion installation cost, and so was to be installed in segments over 20 years, with the cheap parts (on-grade, rather than underground tunnels or above-grade rail) built first to start generating economic benefit on a hundred billion or so per year; we couldn't get the Federal funding.
So Canada thinks they can spend $75,000 per year of money they take out of your income so you can further pay $3.50 for a one-way trip that takes 20 minutes, instead of paying $20 for that same trip. Their alternative is spending several billions for a bus system or hundreds of billions for a rail system with long transit times. Seems legit.
It's a romantic list. They're having a romantic fantasy about the good old days of the cinema.
Inflation over 10 years should be 2% per year, or 21%. That $20,000 should be $24,379. With technical progress, the cost of education should come down (the share of education fees associated with tuition will reflect class size and professor salaries; the share associated with facilities management overhead should decrease due to improved technology reducing the amount of labor required and, thus, the total salary paid per student-year to the collection of people who aren't professors). That means the $20,000 should be somewhat less than $24,379.
Because the costs should come down for the entire cost of tuition, the proportion of tuition taken as debt should also shrink for students with an active income during their vocational training. This further reduces the expected exit debt.
Some landlords with 6-12 units have said a unit that stays empty an extra month is their entire year's profit down the drain. That's why they're aggressive and make you keep paying if you move before your lease, and recruit exiting tenants to find a new tenant. They'll start looking for a new tenant 3 months before your lease is up, and kick you straight out if you don't renew right away when your lease is up.