Well, they have certainly thrown another one at the wine folks. Now that they are "reasonably close" to having winapi done, they've got WinRT to deal with.
Some components are more critical than the others. What you do is have a gold set of parts, and then some substitutes, and you should in fact build it at least with all the substitutes put in all at once. Suppose that for all parts you have up to 3 substitutes. You build 3 different boards, and all three go through all the testing. You don't necessarily have to test all combinations unless you have a good reason to. Everything depends on volume and liability, of course.
Of course EMC testing doesn't ask for substitutes, but it's up to you as a design engineer to understand what parts may affect EMC performance and qualify them -- either through component-level testing, or through system-level testing. If you have a bunch of decoupling caps on the board, you'd probably want to run them and the substitutes on a component/network analyzer and make sure their as-shipped performance is in the same ballpark. You also probably want to ensure that by good luck the self-resonant frequency of "current" capacitors doesn't fall right on top of something nasty. The replacements may not give you such luck anymore. That's just one example. There are parts where the "special" requirements are not about RF, but, say, about leakage, or switching characteristics, etc. Some of them aren't even put down on paper: the circuit works fine with one part, you substitute something else (say "same" semiconductor but from a different vendor), and it turns out it doesn't work or works poorly or doesn't work over the entire temp. range, etc
Sure if you have rather rigorous design process, you make strides towards the circuit that does not depend on unspecified properties of the parts. But there's only so far you can go in that direction, there's a point where you have to test alternatives.
It's not good engineering to just stick to the minimum legal requirements. If it were so, lawyers could do the job:) There's a big chasm between minimum requirements and good engineering.
The deal is, it's not something that you can use while using the laptop as a laptop. As a tablet -- sure. Laptop with upright screen -- nope. Since laptops are, with scant few exceptions, much bulkier than tablets, I'd much rather have the sleek tablet as a separate device. That's me, though.
I don't know who you call a "buyer", but in any sane volume electronics manufacturing company, you'd actually have an engineer do part qualification prior to ordering anything that reaches the customer. I don't know what kind of a dysfunctional company your experience comes from, but man, if you design stuff without having a clue what exact parts will end up in the shipping product, you've got some serious organizational problems. Your production operations director or whatever that position is called should be fired ASAP.
Besides reliability, the obvious issue is that whatever compliance testing you've done on your prototypes is null and void if you do part substitutions outside of the engineering process. Usually a design subject to compliance testing is furnished with a list of acceptable substitute parts -- parts that should have been qualified for their purpose by the component engineering dept. If you deviate from that, you'll fail your UL audit/inspection.
It'd have been cheaper to put up a cylindrical screen and point a couple projectors at it. You can get a heck of a bright system for the cost of those LCD panels...
I know it's an anecdote, but my neighbor's display has developed a bunch of stuck pixels in 3 years or so. Annoying. I don't know if that's a common issue, though. It used TI's original chipset, I don't know if it was licensed to other parties though, so that may be a moot point.
Agreed. I've disassembled my neighbor's old one that had a bunch of stuck pixels. It's a fairly lightweight construction, the TV is pretty much an empty plastic box with a bit of hardware at the bottom. The center of mass is very low. I'd argue the model in question (I forget what it was) was a bit on the light side and was easy to tip over in spite of having this very low center of mass. It felt like 60lbs at most.
The screen brightness leaves a lot to be desired, IMHO. Personally, the only good thing ever to have come from those monsters is, well, monster Fresnel lenses. Good for melting concrete and such.
I've just installed Win 8 on a VM, and it doesn't boot any faster than Win 7, even after repeated boots. YMMV, of course.
The various gestures are a pain to execute if the real screen extends past what Win 8 imagines the screen to be -- namely the VM's window. I'd go as far as calling Metro interface's mouse gestures useless on a windowed VM because of that. For the desktop mode, the number of applications people typically use is very small anyway, so you might as well throw the common shortcuts on the desktop and be done. For other things, keyboard shortcuts are OK.
Win 8 seems rather unpolished. The settings are haphazardly scattered between the Settings App and Control panel. I don't mind the apps, they look nice and fluid, but they won't even let me have their own kool-aid if I have to go to desktop mode just to do the basics. I was expecting that every application that came with Windows would be ported to Metro. That MS hasn't done that pretty much dismisses the whole Metro exercise in my mind. Fucking stick to it or go home, MS, mmkay?
I don't know what kind of a laptop do you have in mind, but on what I consider a laptop a touchscreen would be an ergonomic nightmare. Are you sitting comfortably in an ergonomically sound position? Hands at the level of your elbows, resting on the keyboard or your desk? Now fucking raise your entire forearm 4-5 inches up, as would be necessary to point at something on the screen without elevating hands above the elbows, and keep it there. Tell me how it feels after you've kept them there for 15 minutes. I hope it clears any illusions.
It's not about Apple, Microsoft, or anyone else. It's about biomechanics and ergonomics of the situation. It's about the same reasons lightpens didn't pan out. It truly sucks to have to use them, and the problem wasn't their weight -- I've had extremely lightweigh ones, and they pretty much made you wish to embed the monitor in the desk because your arms would start killing you after an hour of use.
Yes, if you're standing up, touchscreens are a different ballgame. If MS wants to pander to people who work in fast food restaurants, well, good for them. Yes, I'm exaggerating, but the touchscreen market is pretty much limited to standing jobs.
Let's see. Cable modem/router, NAS, TV, two MacBooks, an iMac, two iPods, washer, dryer, microwave, fridge, and a bunch of toys with processors in them. Out of those, 8 have raster displays, and 9 can be considered general purpose since you have at least console-style access to them and can at least run scripts of some kind. Adding two cars to the mix, that's probably close to a 100 CPUs in one household of four people. Yes, I know a microprocessor doesn't really make a general purpose computer, but who said GP computers:)
So, obviously, VPS providers normally only serve users that don't really use the capacity they paid for? WTF? If I pay for certain bandwidth, I better get it.
Why does making the decision about data storage backends is something that's touted as being good in any shape or form? Ultimately the query optimizer has all the information that's needed to make such a decision. If I want an SQL database, that's what I want, not any particular backend. I shouldn't need to worry about it.
For truly install-and-forget embedded uses, Firebird fails because it has nowhere near the kind of testing that sqlite gets. Sqlite is tested to ensure data integrity that lets it be used in avionics, for example. It's the only open source project out there that I know of that could even begin to claim that. In the recent years, if there's a problem with sqlite database corruption, I treat it like a failed memtest run: it means the hardware needs to be replaced.
Slashdot Mirror
Heck, some people are actually allergic to certain vaccine ingredients and can't be vaccinated because of that.
Above post should go as a Community FAQ question/answer pair onto electronics stack exchange ;) Well done.
Well, they have certainly thrown another one at the wine folks. Now that they are "reasonably close" to having winapi done, they've got WinRT to deal with.
Some components are more critical than the others. What you do is have a gold set of parts, and then some substitutes, and you should in fact build it at least with all the substitutes put in all at once. Suppose that for all parts you have up to 3 substitutes. You build 3 different boards, and all three go through all the testing. You don't necessarily have to test all combinations unless you have a good reason to. Everything depends on volume and liability, of course.
Of course EMC testing doesn't ask for substitutes, but it's up to you as a design engineer to understand what parts may affect EMC performance and qualify them -- either through component-level testing, or through system-level testing. If you have a bunch of decoupling caps on the board, you'd probably want to run them and the substitutes on a component/network analyzer and make sure their as-shipped performance is in the same ballpark. You also probably want to ensure that by good luck the self-resonant frequency of "current" capacitors doesn't fall right on top of something nasty. The replacements may not give you such luck anymore. That's just one example. There are parts where the "special" requirements are not about RF, but, say, about leakage, or switching characteristics, etc. Some of them aren't even put down on paper: the circuit works fine with one part, you substitute something else (say "same" semiconductor but from a different vendor), and it turns out it doesn't work or works poorly or doesn't work over the entire temp. range, etc
Sure if you have rather rigorous design process, you make strides towards the circuit that does not depend on unspecified properties of the parts. But there's only so far you can go in that direction, there's a point where you have to test alternatives.
It's not good engineering to just stick to the minimum legal requirements. If it were so, lawyers could do the job :) There's a big chasm between minimum requirements and good engineering.
The deal is, it's not something that you can use while using the laptop as a laptop. As a tablet -- sure. Laptop with upright screen -- nope. Since laptops are, with scant few exceptions, much bulkier than tablets, I'd much rather have the sleek tablet as a separate device. That's me, though.
$3k for a 75 inch screen is expensive?!
I don't know who you call a "buyer", but in any sane volume electronics manufacturing company, you'd actually have an engineer do part qualification prior to ordering anything that reaches the customer. I don't know what kind of a dysfunctional company your experience comes from, but man, if you design stuff without having a clue what exact parts will end up in the shipping product, you've got some serious organizational problems. Your production operations director or whatever that position is called should be fired ASAP.
Besides reliability, the obvious issue is that whatever compliance testing you've done on your prototypes is null and void if you do part substitutions outside of the engineering process. Usually a design subject to compliance testing is furnished with a list of acceptable substitute parts -- parts that should have been qualified for their purpose by the component engineering dept. If you deviate from that, you'll fail your UL audit/inspection.
It'd have been cheaper to put up a cylindrical screen and point a couple projectors at it. You can get a heck of a bright system for the cost of those LCD panels...
Change the bulb and you'll know what you're missing brightness-wise :)
I know it's an anecdote, but my neighbor's display has developed a bunch of stuck pixels in 3 years or so. Annoying. I don't know if that's a common issue, though. It used TI's original chipset, I don't know if it was licensed to other parties though, so that may be a moot point.
Agreed. I've disassembled my neighbor's old one that had a bunch of stuck pixels. It's a fairly lightweight construction, the TV is pretty much an empty plastic box with a bit of hardware at the bottom. The center of mass is very low. I'd argue the model in question (I forget what it was) was a bit on the light side and was easy to tip over in spite of having this very low center of mass. It felt like 60lbs at most.
The screen brightness leaves a lot to be desired, IMHO. Personally, the only good thing ever to have come from those monsters is, well, monster Fresnel lenses. Good for melting concrete and such.
You only think that because you don't have to use the damn touchscreen continuously while sitting down.
IOW: I'll check if I can return the damn system builder DVD for a refund, it's that kind of a letdown.
I've just installed Win 8 on a VM, and it doesn't boot any faster than Win 7, even after repeated boots. YMMV, of course.
The various gestures are a pain to execute if the real screen extends past what Win 8 imagines the screen to be -- namely the VM's window. I'd go as far as calling Metro interface's mouse gestures useless on a windowed VM because of that. For the desktop mode, the number of applications people typically use is very small anyway, so you might as well throw the common shortcuts on the desktop and be done. For other things, keyboard shortcuts are OK.
Win 8 seems rather unpolished. The settings are haphazardly scattered between the Settings App and Control panel. I don't mind the apps, they look nice and fluid, but they won't even let me have their own kool-aid if I have to go to desktop mode just to do the basics. I was expecting that every application that came with Windows would be ported to Metro. That MS hasn't done that pretty much dismisses the whole Metro exercise in my mind. Fucking stick to it or go home, MS, mmkay?
I don't know what kind of a laptop do you have in mind, but on what I consider a laptop a touchscreen would be an ergonomic nightmare. Are you sitting comfortably in an ergonomically sound position? Hands at the level of your elbows, resting on the keyboard or your desk? Now fucking raise your entire forearm 4-5 inches up, as would be necessary to point at something on the screen without elevating hands above the elbows, and keep it there. Tell me how it feels after you've kept them there for 15 minutes. I hope it clears any illusions.
It's not about Apple, Microsoft, or anyone else. It's about biomechanics and ergonomics of the situation. It's about the same reasons lightpens didn't pan out. It truly sucks to have to use them, and the problem wasn't their weight -- I've had extremely lightweigh ones, and they pretty much made you wish to embed the monitor in the desk because your arms would start killing you after an hour of use.
Yes, if you're standing up, touchscreens are a different ballgame. If MS wants to pander to people who work in fast food restaurants, well, good for them. Yes, I'm exaggerating, but the touchscreen market is pretty much limited to standing jobs.
Like, um, duh?
Let's see. Cable modem/router, NAS, TV, two MacBooks, an iMac, two iPods, washer, dryer, microwave, fridge, and a bunch of toys with processors in them. Out of those, 8 have raster displays, and 9 can be considered general purpose since you have at least console-style access to them and can at least run scripts of some kind. Adding two cars to the mix, that's probably close to a 100 CPUs in one household of four people. Yes, I know a microprocessor doesn't really make a general purpose computer, but who said GP computers :)
My car has probably as many on board computers than that. It'd only be fishy if the car ended up in a lake.
So, obviously, VPS providers normally only serve users that don't really use the capacity they paid for? WTF? If I pay for certain bandwidth, I better get it.
+1 Informative
-1 Whoosh
Europe: they give you a house next to the brewery, and fresh, free beer for life. :/
The US: they give you an alias.
So, what other major SQL database products offer this choice? Why, or why not? There's your answer I think.
Why does making the decision about data storage backends is something that's touted as being good in any shape or form? Ultimately the query optimizer has all the information that's needed to make such a decision. If I want an SQL database, that's what I want, not any particular backend. I shouldn't need to worry about it.
For truly install-and-forget embedded uses, Firebird fails because it has nowhere near the kind of testing that sqlite gets. Sqlite is tested to ensure data integrity that lets it be used in avionics, for example. It's the only open source project out there that I know of that could even begin to claim that. In the recent years, if there's a problem with sqlite database corruption, I treat it like a failed memtest run: it means the hardware needs to be replaced.