What you're suggesting is nothing new. There are many parts in a PC. Motherboard, 2-4 DIMMs, powersupply, case, DVD drive, hard drive, monitor, keyboard and mouse. That's 10-12 parts for a simple PC.
While you'd be able to claim "I got it all from one place, so keeping track of where I got things is not a problem". You still need to consider when you bought the specific item. This means keeping a detailed database or spreadsheet for all thousand machine with service record for accounting purposes.
Additionally, if you're going to build small form factor systems (which means less materials to build and less to throw away) and you want to stay at least somewhat "green", you'll need to spend a great deal of time researching and identifying reliable components (less throwing away) which use minimal power. It's not good enough to just buy what's cheap. You need to consider what the environmental cost of each component is. When you toss a beige box, can you reliably have it recycled? Consider the energy costs in melting things down. How much epoxy is in the system etc...
Then you have accountability. With Dell, you have an American organization (using all Chinese parts, but that's besides the point) which supports their equipment in America and build machines as a whole. Instead of dumping them, you should instead demand you get more for your money. Before buying a new series of machines, be sure to demand they provide detailed environmental data regarding the machines you're buying. Make them work for the money.
Now, if you're going to go the cheap way. Small tossable machines are best. Look for machines like the Shuttle XS35GT (http://www.newegg.com/Product/Product.aspx?Item=N82E16856101099), it's a simple little machine that works well for Windows 7 and Office. For another $150 you can get a 22" screen. Total cost $500 for the purchase + about $300 or so for the software (if you're lucky, Microsoft licensing shouldn't be too bad).
So, when all is said and done, you've save $200 a machine and now, you have to support each machine yourself. Dell will come and fix the computers for you. Now you've got to toss it and replace it. I have no idea what the failure rates on the machine is, but you'll find out quick enough.
Somehow $1000 per machine from Dell actually sounds like a good deal now. And I hate those guys. Out of the 30 Xeon based workstations we've bought in the past year, 10 have had major problems with running high performance software. Dell has now spent a month and sent 4 different guys several times to repair one of them. It's pathetic. The last guy, we almost died laughing at since we saw him take the new motherboard out of a box and hold it in one hand and actually pick his nose and scratch his ass with the other while staring at it for a while.
Qt is amazing. To this day I see it as "The Standard C++ Library" as the one which is part of the standard is utter crap.
Qt suffers some major drawbacks. First of all, because of lack of something similar to delegates as you'd find in C# (though there are some interesting "clones" of it), Qt still uses moc. I have a personal library which practically clones the structure of Qt but is based on inheritance instead of signals and slots. It works really well and Nokia/Qt should consider switching to that model as well.
Second, Qt draws everything itself. It does not make use of native system form controls which has always been a huge point of contention in the Mac world. Qt does a great job, but because of the piss poor support from Apple for making accessibility features easily implementable, DIY form controls are a bit of a nightmare. It was a hair better in Carbon, but in Cocoa, it's no fun.
Now, you're in the right direction for iPhone development. WebKit is based on a clone of Qt Core which is really the best part of Qt anyway. You can do almost all your coding on another platform and then port back to Mac with little effort thanks to Qt core.
If you're making a cool program instead of some form controls based crap, then in reality, you're probably writing the entire UI yourself anyway. For that, you can easily make use of Qt and QGLWidget as a base. QGLWidget is insanely easy to port to NS classes.
I write all my iPhone stuff by doing all the development in Windows in Visual Studio (2008, but will move to 2010 now that Qt supports it). I write everything using OpenGL ES. I use the really great tools from NVidia and AMD (as well as others like Unity) to develop shaders which I use for entirely inappropriate purposes such as "because I can". Then I use Window CE based phones running a Samsung ARM processor to optimize my code for iPhone. When I'm done, I load up the project in XCode, toss it into a skeleton application, compile and debug for iPhone.
Best part is, for multiple developers you can use a single Mac Mini running iRapp to integrate Mac OS X into your Windows desktop and you never even have to code on Mac (which I never liked unless it's from vi on the command line). I guess you could also use a hackintosh as well, but a Mac Mini is OK for an occasional compile and a little debugging. A bit on the pricey side though.
The next "nasty nasty" is how to support Android like this, I do this in a similar fashion, but then I don't even need the mac to do it. Just code everything in Visual Studio, then create make files to compile the native code for Android. Then use whatever crapheap Java tool you want to hack the Android UI together, make some JNI calls to start the library and you're good to go. Expect to have at least 20 Android devices near by to test on though, you can't write native code for Android and expect it to function properly on different devices.
As for CrapBerry, you're basically screwed altogether. There are some piss-poor solutions to use GCC to compile for a Java back-end, but it's crud. The only thing I can recommend in this situation is that if you're starting a new app, keep the code as "portable as possible" meaning so that you can either preprocess your C++ code to Java and polish it or make your C++ code look exactly like Java for the most part.
Qt is an awesome platform, but it looks like it will be a long time if ever before there's such a thing as a good cross platform solution for phone development. Hell, I have problems writing code that will work on two different Android devices.
These are developers making freebie apps based on web standards.
Personally, I use iPhone myself, but I have Android and am looking forward to getting Windows Phone 7. When I write apps, what's interesting to me more than anything else is my ability to port the basics and write the app around them. On a Java based platform, my hair and fingernails itch just thinking about porting tens of thousands on lines of C++ code to a platform which doesn't guarantee a processor type. I don't mind screen resolution as I do as much as possible with OpenGL as I can, but screen format is a problem. I don't like the aspect ratio problem.
I know my cost of development for iPhone is SUBSTANTIALLY less than for Android as I have all models of iPhone and will probably only need to buy one per year to support it. Testing on all models is pretty easy to do. Android on the other hand is more of a hackers platform where you write it for the phone you have and if it works for someone else... great. If it doesn't, well, you'll need to get another phone to test with. These phones unlocked range from $200-$1000 and for any decent level of quality control, you'd need at least 20 of them today with people testing on all of them. In another year, you'll need a lot more. When Android is shipping on x86 hardware, then you'll need to make your code compile for both x86 and ARM which isn't that big of a deal except for alignment issues, but still then you're maintaining two executables and you have to buy even more devices.
Android doesn't guarantee :
- Screen resolution
- Graphics processor
- CPU type/generation/feature set (vector unit)
- Audio chip
- Audio latency
- Input method
- Multi-touch vs. Single touch
- Keyboard type.
- OS interaction with running applications
So, for anything more advanced than the basics webby kinda stuff, Android is a nightmare.
Cost of development for Android (with some QA) is MUCH higher than for iPhone.
The key is to write the application for iPhone, Android AND Windows 7 and leave it up to your paying customers to decided which one you'll focus the most QA on because of the best profit stream.
If you're starting from scratch on a new app, I guess it doesn't matter that much, you can work around it by using MonoDroid, MonoTouch and Windows Phone 7 SDK. With some work you can port most of your libraries to a CIL assembly. C++ doesn't port so smoothly there, but it's not that bad either. I am hoping that Mono eventually becomes the champion of phone development. It seems well on its way.
Webkit was a special case. Lars Knoll (last I heard who now works at the company formerly known as Trolltech) was the author of it. Smart guy, very nice too.
WebKit was very little more than :
- A) Stripping as much of Qt as possible out of konquerer.
- B) Rewriting what was left of Qt from "compiled... oops error, need new function".
- C) Wrapping QWidget as an NSWidget
When people started complaining, Apple took the code base and isn't giving back. Then Apple tried giving back, but seemed to do a terrible job of it. But because of customer demand, they kept adding to and fixing webkit. Eventually, everyone just started using WebKit... including Trolltech and Lars Knoll:)
I worked with Apple on browsers before the webkit decision was made and shortly after Microsoft forgot about IE for Mac. Trust me when I say, they wanted two things from a browser. Some control over the code base, the ability to not have to screw too much with open source licenses and headaches. Their WebKit project gave them just that. Given the importance of the browser to any platform, they made a great move even though it looked awful at first.
Now thanks to Google and Chrome being ported to Mac, it looks like there might even be a good, stable browser on Mac someday. As an occasional Mac lover, I look forward to getting more stable software on OS X. Bells and whistles are nice, but I really would rather them focus on getting the basics stable. Maybe start using progress bars or status messages to suggest that the computer is working and isn't actually locked up.
Boot time really needs improvement too. Pretty bad that Windows 7 restarts twice as fast on my MacBook as OS X does:(
As a developer you'll be far more valuable if you're able to deliver and app no matter what the platform is.
The idea is, the platform, the language etc... is entirely irrelevant for an engineer. My personal definition of an engineer is either "A person who can produce the most complex solution for the simplest problem" or more applicably to the discussion "A person who can identify a problem, break it down into smaller easier problems, posit solutions for each one, design and/or implement each solution therefore solving the problem".
Writing an application is a big problem. You need to identify "Do we need it?", "What should it do?", "Desktop, mobile, specialized device?", etc... once you've identified that you're interested in writing a mobile application, you can then identify whether it should run on just one platform or be platform neutral. If it's platform neutral, can a "core functionality" be identified and written as a library which will compile for any device of interest. Then you can implement the UI for each device as needed.
Problem is, unless you're using Adobe tools or maybe Unity3D (for the moment, which means you'll have to reinvent some major wheels at times) there isn't a common programming language for the mobile platforms. Apple has C/C++ support (then stick it in a ObjC wrapper), Android has C++ support... with effort, then make a Java wrapper. Windows Phone 7 tries to pretend like C/C++ doesn't exist, but I've heard of success of using it through managed extensions. Symbian if you actually consider that platform worth bothering with has C/C++ with Qt which is kinda nice. And Blackberry (another annoying waste of time) will require a bunch of rewriting everything for Java. (I hear you can use C/C++ but at your own risk). Brew doesn't even count, that platform is a programming disaster.
People are trying to make cross platform solutions for phones, but Apple, Google, RIM, Nokia etc... are making this as complicated as possible. MonoTouch is okish and Air is also okish, but you'll have limits on them both as opposed to programming natively for each.
What I have been doing is to write all my core libraries in C#, but "typedef" all your types to something native to your code. Then for strings, subclass String and add functionality to it to make it work like the one from Qt. C# is great because it borrows so heavily from C++ AND Java. If you code like this, then you can make a simple script that will convert your C# to either C++ or Java without any intervention and then porting is EASY!!!!
Intel will have an offering which provides equal performance for approximately the same price. I love AMD, thanks to them, it makes it possible for me to buy new Intel chips for less money all the time.
Let's not forget the "do all your shopping in your pajamas" effect. While I can buy most of my stuff online these days, and I can even save enough money that I can buy three different models and toss the two I like less because of it, I still get off my ass and go to the store... while I can still get off my fat ass that is.
If you need to go out and shop around to find what you're looking for, there is walking involved. There's bending over and picking things up. Sure, this doesn't sound like a lot of exercise, but 2 years ago, I went from 200lbs to 152lbs in 4 months by doing a little of this and a little of that.
I found simply being out and about made me eat less and eating is something a lot of people do when "I'm bored" sets in. The exercise from walking alone isn't just a matter of exercise, but it's a means of doing something other than eating.
Shopping in a mall probably wastes gas in the car and pumps carbon into the atmosphere. But shopping from your couch adds wrappers to the trash bin, not to mention all the other paper involved. As you're shopping from your couch for things like radios and ipods, eventually your ass gets big enough that you start shopping for pants and shirt because you're embarrassed to go into the Big and Fat shop. Of course, if you find yourself sitting next to a skinny little hottie on a bench while she's nipping on a yogurt with muesli and you're chopping down on a triple whopper with extra bacon, you should take the hint then and there.
The environmental impact is much much worse from shopping online than the article leads. People stop moving, they get fat, they buy bigger cars to haul their fat asses around and they buy bigger articles of clothing and guess what... they order more stuff online because moving their fat asses is even harder than before.
Now that I'm "in shape" I loaded up a bag 50lbs of weight (about the same as I lost, 23ish KG for the rest of the world) and I went for a 5km walk with it on. I damn near died from that. I have far more muscle mass now than I did then, but carrying that weight and walking a decent speed was bordering on impossible. If I had that weight back for real, I'd lay in bed and order everything online and say screw it. I'd have a house completely filled with blister packs and cardboard because I wouldn't even bother moving my ass to bring it to the trash can.
Should be ban shopping online because it's killing the planet? Hell no, I love shopping online and it's keeping the guy who has to move my 100lb boxes around from getting a fat ass. But we certainly should consider a way we can get fat assed people out of their couches as well.
I have purchased approximately a thousand DVDs over the past 14 years. I have a room in my house damn near consumed by them. They take tons of space. Before that, I bought tons of VHS (possibly literally, they were heavy)
What I hate is :
1) buying on VHS then having to pay again when I want DVD, then again when I want Bluray, then again when I want downloadable, etc... I have pirated films that I have already purchased on earlier formats. If the industry wants a proof of purchase for every film in my library and a serial number for it, cool... where do I send it. Then when they release it in a new format, let me download it.
2) If I can't buy a film downloadable in another country, I simply download it elsewhere. I'm perfectly happy to pay for films I watch. But I sure as shit won't be treated like a 3rd rate citizen because they decided where I live isn't convenient to them to sell in (Norway is too small to interest many companies).
3) If they are price gouging because of my location (very common, BluRay in U.S. $14, in Norway $45) and they won't let me download it at U.S. prices + Norwegian taxes, then I'll pirate it today and wait for it to land on the $15 rack in clearance down the line.
As I said, I've bought about 1000 DVD's at an average price of $22 (have a spreadsheet) a disc. That's $22,000 U.S., if they want to call me a pirate for trying to give them money but refusing to pay a 200% premium for my location, then send the cops my way:) Oh.. this is Norway, wouldn't matter.
I actually work in the industry as well. Used to develop software at a post production shop that produced DVDs for the Scandinavian market for WB, BV, etc... Disney used to ask me to rip DVDs for them because they couldn't get their hands on their own films due to licensing limitations. So to make local trailers for films, they'd send my pre-release DVDs and ask me to rip them for them.
I often pirate video games, can't find them in the local stores and I won't pay $15 shipping for a $10 disc... I use cracks constantly, I buy games and use cracks so I don't need to leave the disc in.
Oh... I have all 1400 discs ripped to my server so I don't need to screw with finding discs. Also, I'm bloody tired of paying over and over for the same film if my kid scratches the disc. So, both my Wii's are cracked and play from hard disc, my playstation 2 too.
At the moment, I can't hook my DVD player's HDMI up to my projector since it's a cheap projector with only VGA on it. It has RCA also, but when I use it, the speakers in the room start buzzing as the chasis of the DVD player isn't properly grounded. The HDCP crack will solve this. HDFury is a half assed solution, this will allow real solutions instead.
2 needs a brief revision. It wasn't specifically about cheap embedded devices, it was entirely about small and simple.
The issue was that they wanted to encourage (at pain of death) the vendors to embed the HDCP encoding hardware directly into the video decoding chips. The HDCP encoder needed to be tiny enough that it would take only a minuscule portion of the FPGA/ASIC containing the video decoder. This made it so vendors would agree to do it. If they would have implemented a more complex system similar to ZRTP (one of the more complex, yet secure streaming encryption systems) first of all, there would be tons of compatibility problems. Vendors of video IP would have tons of problems implementing it and the differing interpretations of the spec would guarantee uselessness of the spec.
Additionally, it should be able to be implemented using pure logic, arithmetic is expensive. A full adder adds a huge layer of gate depth into a chip. Each bit adds at least one layer of depth. This adds latency etc... a logical algorithm can perform an entire step in the encoding/decoding chain in one level of depth. Also, logic operations are REALLY simple to route in ASICs and FPGAs. Non-power-of-2 multiplications and divisions would add loss, not only damaging the quality of the signal, but possibly the control codes, rendering the output undecodable.
By placing the HDCP stuff directly within the ASIC/FPGAs of the decoders (the primary goal as earlier noted), it made it so that while the algorithm itself was generally weak to begin with, the keys would be much harder to obtain as they would be coded directly inside the ASIC. Also, it makes it impossible to intercept the decoded signal between the decoder and the encrypter.
DVD CSS is an entirely separate interesting case. I was attempting to license the DVD specs to make a licensed DVD player back before DeCSS was released. I was reading through the requirements for licensing and find it amazing that almost all the interesting bits have disappeared from the face of the planet. Here's some things I vaguely remember from it, they may be imprecise :
1) The spec was made by Intel (a pretty smart group of guys normally)
2) The spec itself cost $5,000 + delivery via a secure physical method.
3) The spec was licensed to a company, but only after meeting certain security requirements.
4) Only "cleared individuals" would be permitted access to the spec.
5) The spec was to exist ONLY within a sealed room, no windows, heavily locked door.
6) The door would only be able to be opened by individuals cleared for access by the DVD consortium.
7) When the cleared individuals were not in the room and at any time the door was open, the spec would have to be stored in a fireproof safe which limited access only to individuals (the ones cleared to be in the room).
8) All source code for implementing the spec would exist ONLY on a removable medium (floppy, removable hard disk, zip disk, this was before thumb drives).
9) All binary implementations of the code would be heavily encrypted and could not be decrypted or deciphered without the use of a multi-million dollar laboratory environment.
Oops! That's where the failure was. So long as code can be run, it has to be in a decrypted form. So once the code is decrypted to be executed, then ANY debugger which can function on an application in memory could be used to read the code in assembler form. The complexity of the algorithm is irrelevant, all that matters is the amount of time a person has to sit and sort it out. The film industry wanted this clause in, but I'm entirely convinced that the guys who made the spec knew before hand it was entirely unrealistic.
You'll note that they focused a great deal of effort on the protection of the physical specification. I'd actually be amazed if the formal specification even exists in electronic form anymore. I'd imagine that it's strictly in print and is photocopied for each new licensee if there are actually any.
Seriously, I would be a huge fan of cutting back violent sport for anyone that has enough testosterone to believe that yelling at an inanimate device such as a TV set would have any impact positive or negative to the outcome of the event which they're watching.
How do you ban selling violent games to kids that have parents who will simply buy them for them? What defines a violent game? Super Smash Bros is EXTREMELY violent. Should that be banned? For tht fact, I remember "Nintendo Duck Hunt" having been quite violent as the only object of the game was to kill animals with a pistol. On the other hand, my 8-year son has friends who rush home each day to play Call of Duty. Their parents apparently didn't believe they were violent. Some of the 10 year olds in my neighborhood play Grand Theft Auto, their dads are sure they know that it's not real.
Why do we waste time on laws that are a waste of time outside of the theory. We're making laws that to try and do the things that parents should be doing instead. If the parent lets their kid go out and buy a violent game, and the kid almost definitely got the money for the game from mom and dad, how hard is it to get their mom and dad to simply buy the game for them? Of course, most of those kids can't get the money from mom and dad for the game, so they just pirate it instead. So what you're really saying is, to the few kids that are honest enough to actually buy a copy of the game, they should pirate it instead?
Tell you what... when you figure out a way to keep 11 year old kids from looking at porn, I'll support your bill.
Apple's model for the telephone market is almost definitely better for application developers. An application developer can buy 3 models of phones (and iPad if you care about that thing), test on each and be sure that everything works. The fact that iOS is such a closed platform is fantastic and makes it so that we developers can be more confident of what we ship to the public. It also means that we can optimize code to run well on all the phones which run that operating system.
Android on the other hand is more like the next step of Symbian... with slightly better design and control. Symbian was a heap of shit for developers. The API was a nightmare, content delivery worked only sometimes. Their package management system was a tinker toy. Additionally, their memory model was designed with a 25 year old PDA in mind, and their argument for it was that it needs to work with GCC 2.91. They implemented an ad-hoc exception model with a "clean-up stack" which was a lame excuse for auto-pointers as 2.91 didn't have good template support.
Android on the other hand has a relatively simple development model and it seems as if application development (so long as native code isn't important) is really quite easy. You can code in their Java like language (I do this to help with the law suit to differentiate and call it something else) and make an app and get it running quickly. Unfortunately, it runs on about a billion different processors (there are tons of ARMs out there) with a gazillion (quite cool that word is in the spell checker) graphics subsystems out there (nVidia, frame buffer, TI, etc...) and there are a multitude of different types of touch screens (single touch, multi touch, hi-resolution, low resolution, no-touch, just joypad, high latency, low latency). There are a pile of audio subsystems, I won't even begin to cover the massive number of those, it's mind boggling.
Writing simple cook book and business apps for Andoid is a charm. Takes far less time than on iOS, almost as little time as on Windows Phone 7 (which is WAY EASY) and can be tested more or less in an emulator without any problems. The only issue is the touch screen input which can be averted by making the buttons all a little bigger.
Anything requiring high response rates, fancy input methods, real-time audio, etc... is a nightmare on the platform. It's even worse than on Windows. There are just too many methods of input.
Android is a pretty neat touch screen platform that allows absolutely any manufacturer out there to make a full blown smart phone for almost nothing. Chinese vendors are already pumping these things out by the truckload and it's only a matter of time before it's possible to buy full smartphones for $50 or less.
You can buy an after market iPhone screen and touch panel from China for $20 (free shipping). And they are pretty good replacements. This means that they can get them for less than half that. Cheap system on a chip ARM processors can be bought for less the same. It's entirely possible that you can get ALL the parts required to make a full Android phone in China for probably $30. The specs will be pathetic, but will improve rapidly over time. The result, an Android phone containing the bare minimum memory required to run the phone, the bare minimum CPU required to run a telephone call, the bare minimum audio quality required to hear the other person, probably not even enough specs to download an application.
Of course, noone would buy these phones right? Well, probably not more than 100,000 of each model (which is the target Nokia sets for their mid-range smart phones). Remember there are a shit load of Asian people buy Chinese knockoffs of all these things. And what's best is, these aren't even knock offs. Thanks to the open source nature of Android, it's 100% legitimate to make these things. Of course, no westerners would buy these things. Umm... or would they. DealExtreme.com will sell tens of thousands of these. They'll be sold all over the Mediterranean and Caribbean islands to t
Seriously, who gives a shit about a telephone that can do more than an iPhone or whatever can now. 720p 30fps video is certainly good enough for a mobile phone. 1080p is only useful when we start using the phone to replace PCs. For that we need a PC operating system.
When will Mac OS X run ALL Mac software on the iPhone? Or when will <insert vendor's name here>'s phone start running Windows 7 64-bit. I want a phone with dual core, 1.0+ Ghz or better x86 or something with a dynamic translator that can run Mac OS X or Windows 7 64-bit. It needs wireless USB, wireless HDMI and my office desk chair should have an inductive charger in it so I can leave it in my pocket while I'm working.
Until we have that much, then who gives a crap about dual core processors on telephones? Last I checked my Mario DS runs just fine on a 67Mhz ARM9 CPU with a cheap crap GPU. Need for Speed runs great on a 412Mhz ARM 11 on my old iPhone 3G etc...
If they want to speed up the H.264 CODEC to handle 1080p, that's actually quite nice, then I can use it for high def movie playback, but unless they add enough storage space to make 20+ Mbps practical, I think I'll stick to 9Mbps 720p. Why more than double the pixels without at least doubling the bit rate to support it. Good quality scaled 720p is much better than crap quality 1080p any day. In fact, I'd rather have 1080i to get the higher field rate.
I think I'll wait patiently for Atom based phones before bothering with dual core in a hand held device. All this crap in-between is a hold me over for the real thing.
I don't actually have an X-Box 360, haven't figured out what I would need one for. It doesn't play Blu-Ray, almost every 360 game worth playing will also be released for Wii or PC, so why have yet another box.
On the other hand, I have 5, YES 5 360 wireless controllers. Two on my game machine hooked to the projector. Two on my 8 year old son's PC and one on my daughter's PC. The X-Box 360 controller is the best thing ever to have to gaming on the PC. Finally a standard PC joystick. It only took a billion years. Last time there was a standard game controller for the PC was the original two button, analog axis controller in the 80's that you needed a separate A/D card (there was no sound cards yet) to connect to the machine. It was the biggest piece of crap ever, but the 360 controller is AWESOME. A standard console grade controller for the PC.
Well that said, once I got these controllers, I bought all the Lego games (Star Wars, Indiana Jones, Batman and Harry Potter), I bought Tomb Raider and piles of other games. After all, the PC just became the most powerful console and the best part is, it plays games from nearly every old console and it also plays Blu-ray etc...
Well, my kids LOVE playing games on the PC with their controllers. When the game says "press the green button", they press the green button. In the room with the projector (a cheap little 100 lumen BenQ Joybee) you can see the colors, but not the letters on the buttons. Also, since the projector typically is downscaling 1280x1024 to 800x600, it is really quite hard to read letters on the screen (even when playing Wii). So, the colors are GREAT.
Here comes Microsoft to tell us that if any of our controllers break, we'll be replacing them with controllers of lower usability. Oh well... good thing couldn't last forever could it? Let's just hope that Logitech and others don't change.
Yep, was thinking that this will be a big boon to the GPS tracking community. Unfortunately, that cops will probably be able to use AGPS as well, which means you'd have to block your cell phone signal. If you're actually in need of a device like this, then there's a really good chance that you probably need to use your phone (probably a tosser) to arrange where you're trying to get away to, book plane tickets, etc... so, you're kinda screwed then. And as a thought, the tracker is probably GSM based anyway, so AGPS would be a much cheaper solution for cops tracking more than one "perp".
Besides, the link you pointed us to is kinda cool, but the prices are obscene. GPS is really easy to block, so it shouldn't take more than a $20 frequency generator with a shit load of bleed. An old CB radio almost certainly can be modified to do this and that would cost nearly nothing.
libc assumes and operating system. An operating system requires system resources, many of the systems at hand actually don't have room for that. For example, the systems you would use in a DVR that run on 100 milliwatts to control the power of the main DVR. These often function as the real time clock for the system as well. I won't suggest you're using a real time clock dependent leap second precision in a DVR environment, however I will say that there is date and time with second level accuracy. These systems are either written in C WITHOUT libc or they're written in assembler.
I often find myself writing Atmel assembler for tiny little chips that I use for configuring 3rd party ASICs. They're really useful as you can use a 2 wire interface to control a device that typically requires 8-16 wires to configure. Atmel chips that are too small for C cost about 20 cents in bulk or over a dollar for the same functionality but C programmable. Since we at least one of these on each daughter board in our devices, it does add up (there are sometimes 5 boards). And since we're making hundreds of thousands of them, we DO save a ton of money that can't be accounted for as "the development time to code and maintain assembler would make up for the additional chip cost".
One of those chips... is a real time clock with network clock synchronization and 0.1us precision. Don't assume that just because this is the 21st century that everyone has the tools available that you consider to be "bare minimum requirements". Hell, if resources were readily available, I would port all our code to C# and write a TI DSP optimized tracing JIT back end for the Mono ILM. But frankly, that's just not the way the cookie crumbles.
Non-idiot programmers have these awesome computing devices with operating systems and tons of ram and hard drive. Sadly, many of us idiot programmers work on devices that have just enough code to run the initial system check and then if you want something you need to compile it and do it yourself.
Of course, we could use libraries, but the boss doesn't want to buy a library for each thing and the open source library has a GPL license or a "not for use in commercial software" or whatever license. Or there's a "just give me credit in the program somewhere" license... umm there's no UI... where can I do that? Engrave it on the cover?
Or we could just open source everything we do, then the first company in China that decides it would be easy to build now that the software is done can make it and sell it on dealextreme for 1/20th the price.
Nope, we write it over and over and over. We like feeding out families.
Let's face it, in we've found copies of the Epic of Gilgamesh dating to 4500B.C., the old testament (which is really more a monotheistic superset of gilgamesh) is from 1800B.C.ish. The real reason only "scholars" got the whole math thing is because they were the only ones who knew why the were counting down instead of up.
Nobo has a point... but it would make it so that the hardware engineers would suffer instead of the software ones. 1/86,4000 of a day = 1 second could be a fair solution. All we would need to do then would be to come up with a new atomic clock which allows for the alteration and then come up with computer crystals that are accurate to the new system (hey, let's get ones that are accurate to begin with, that would be great).
But, since respectable companies tend to run their own SNTP servers and they themselves adjust against national servers (we hope), it could simply be a good idea to ditch the leap second in favor of fixing all the clocks.
But, I think the real issue of the article is the occasions where "17:59:60" is a valid time. I think for presentation (and databases), it would in fact have been better to simply prolong 17:59:59 or progressively added a millisecond for the next 1000 seconds for example. Although it might through off scientific calculations during that period, the impact would be less critical.
The proper solution is to make programmers aware of leap seconds. There are 86400 seconds in a normal day, however there is an additional second added once or twice a year to adjust for solar time.
Wikipedia documents it quite well and programmers in modern times should be heading to wikipedia almost constantly anyway. The real problem occurs when the date/time is given in seconds since an "event" such as Jan 1, 1970. Most programmers don't know about leap seconds and I must admit, I don't generally bother calculating for them. But if it were necessary, it would be relatively trivial to do so.
We shouldn't remove fixes to the clock just because programmers are undereducated. I'm quite convinced that just posting this on Slashdot will raise awareness across a high percentage of the programming world.
I also always wondered why undergraduate studies for computer science didn't make time a relevant issue. It seems as if it's one of the more complex topics and yet, we don't pay any attention to it. Last formal education I had on time (not talking about physic related, but calendar) was in primary school. There are so many time systems out there that we should pay more attention to educating programmers on it.
Audio carried via the SPDIF spec either over copper or optical are obviously immune to classic interference style degradation of audio since it is digital media and you can tell if the audio has errors or not.
That said, if you're running cables in RF noisy environments, the chances of bit errors over copper are infinitely higher than over optical. There is truly a clear benefit to using optical over copper in this case. Same even goes for Ethernet cabling, when making a nice long run across a building where we couldn't avoid RF interference, we could not achieve full gigabit speeds over copper, so we switched to fiber and all was good.
A better comparison is using a $1.99 10 foot optical cable vs. a $100 10 foot monster cable optical cable. While the monster cable might be better manufactured and could theoretically survive a shallower turn radius, in reality, the quality difference between Monster and Radio Shack brown packs is zero.
The only time I spend extra money on cables anymore is for the ones I keep in my laptop bag. I prefer a more durable molded cable which bends easier (you know those nice soft ones that feel almost like microfiber on the insulation?) so that they don't get destroyed while I'm bouncing around my almost never well packed notebook bag.
sounds promising on the track expansion, was wondering though, can modern high speed trains operate over these tracks or would they require new ones. And if they make new ones, then would the new tracks be usable for cargo transportation? Still sounds terribly silly to me that from my understanding the railroads themselves are generally privately owned.
Ans also getting back to the topic, what's involved with noise abatement? Would it not be satisfactory to recess the tracks, piling up the moved earth and stone to the sides of the rail ways... obviously, drainage will be an issue, but I know very little about modern drainage techniques. I'd have thought the greatest cost would be the massive amount of stone that would have moved and laid for the foundation of the railway.
Also, somehow 4-5 days for Washington to New York sounds painfully long to me. I took an amtrak from Spokane to DC one time and I remember it taking just under 3 days. Are cargo trains much slower?
Sorry for all the "what if"s and "what about"s and such, but you seem to have some good information, I wanted to exploit you:)
The #1 key to a strong economy is the ability to move from point A to point B. We currently use trucks FAR more than we should to move around freight that should be moved by efficient trains instead. Problem is, it's too expensive. Additionally, with budget airlines out there, taking a plane is cheaper than Greyhound most of the time. That's just STUPID.
The biggest shortcoming of trains is their dependence on rails. There's no way around this. After all, trains are... well trains. But, let's be realistic here, a single mega-train with hundreds of cars can move around substantially more than a fleet of airplanes... but they're slow and for shipping, it's more efficient to move things by plane. Trains are far smarter than planes many times because they often eliminate the need for hotel rooms for business travelers. You can take an overnight train to your business destination and take a shower in the morning and be ready for a meeting when you get to where you're going. Additionally, if trains sold their rooms in a similar fashion to hotels, so you could bring your whole family by train on a vacation and just rent a room on the train instead of paying per person, then using a 200Mph train, you could transport a family from NY to CA in 18 hours and because the trains don't have a million security checks before boarding, it would actually take only a little longer than using an airplane (it definitely works this way on shorter trips, tried in Germany). Additionally, the "lost money" from the ticket sales can be made up for on food, drink, etc... use RyanJet as a model for trains. Pretty sure they're going to start charging for using the bathroom pretty soon.
Problem is, you can't possibly put a train system like this together in the U.S. using normal government contractors for $8 billion since they'll just piss away all the money on the big-wigs. This is a great opportunity to invest in small business. Have one firm design the rail, it's not that complex (compared to things like making a space shuttle or airplane), then find small subcontractors that are used to performing services for fixed rates and eating the losses if they fall short. Give them enough seed money to expand and buy construction vehicles and get them each to build a portion of the track. The result would be creating lots of new businesses and business opportunities. Making companies across the country that are ready to rebuild the parts of America that are falling to pieces (Detroit anyone?) and the return on that investment alone will pick up a huge amount of the business.
If the larger government contractors start failing and having to sell off chunks of their business, these new businesses which are established upon the basis of getting jobs done for what they quoted in the first place will step in and pick up after them. The large government contractors are big, fat and sloppy and spill more money than they use productively into fat men's pockets. If they can't perform competitively in a modern market where the little guy can, then let them fall and the little guy will take over instead.
As for the trains, well don't reinvent it. Germans, Chinese, Japanese etc... all have them already. Just buy what you need from them or buy the designs. They've done all the work already. Just build the rails to meet the needs of their trains.
Unlike the current rail business which is a disaster from end to end, it's time the U.S. simply owned its rail system. After all, the U.S. owns the interstate highway system, how is this any different? If the rail companies can't make it work, then maybe they just need to fade away.
The end result is, the U.S. has a new rail infrastructure. Train owners can rent tracks from the government and therefore new companies can become involved without being stonewalled by greedy railway owners. A bunch of new little big construction companies are established and willing to do big work for reasonable rates without the bureaucracy, people have an alternative to flying and hotels, and the cargo busi
Slashdot Mirror
What you're suggesting is nothing new. There are many parts in a PC. Motherboard, 2-4 DIMMs, powersupply, case, DVD drive, hard drive, monitor, keyboard and mouse. That's 10-12 parts for a simple PC.
While you'd be able to claim "I got it all from one place, so keeping track of where I got things is not a problem". You still need to consider when you bought the specific item. This means keeping a detailed database or spreadsheet for all thousand machine with service record for accounting purposes.
Additionally, if you're going to build small form factor systems (which means less materials to build and less to throw away) and you want to stay at least somewhat "green", you'll need to spend a great deal of time researching and identifying reliable components (less throwing away) which use minimal power. It's not good enough to just buy what's cheap. You need to consider what the environmental cost of each component is. When you toss a beige box, can you reliably have it recycled? Consider the energy costs in melting things down. How much epoxy is in the system etc...
Then you have accountability. With Dell, you have an American organization (using all Chinese parts, but that's besides the point) which supports their equipment in America and build machines as a whole. Instead of dumping them, you should instead demand you get more for your money. Before buying a new series of machines, be sure to demand they provide detailed environmental data regarding the machines you're buying. Make them work for the money.
Now, if you're going to go the cheap way. Small tossable machines are best. Look for machines like the Shuttle XS35GT (http://www.newegg.com/Product/Product.aspx?Item=N82E16856101099), it's a simple little machine that works well for Windows 7 and Office. For another $150 you can get a 22" screen. Total cost $500 for the purchase + about $300 or so for the software (if you're lucky, Microsoft licensing shouldn't be too bad).
So, when all is said and done, you've save $200 a machine and now, you have to support each machine yourself. Dell will come and fix the computers for you. Now you've got to toss it and replace it. I have no idea what the failure rates on the machine is, but you'll find out quick enough.
Somehow $1000 per machine from Dell actually sounds like a good deal now. And I hate those guys. Out of the 30 Xeon based workstations we've bought in the past year, 10 have had major problems with running high performance software. Dell has now spent a month and sent 4 different guys several times to repair one of them. It's pathetic. The last guy, we almost died laughing at since we saw him take the new motherboard out of a box and hold it in one hand and actually pick his nose and scratch his ass with the other while staring at it for a while.
Seriously, don't tell a guy to get behind sexually abused children until he's be cleared for that. There are databases for that.
Wait... maybe it's those specific databases that has the Pope scared for the church?
Qt is amazing. To this day I see it as "The Standard C++ Library" as the one which is part of the standard is utter crap.
Qt suffers some major drawbacks. First of all, because of lack of something similar to delegates as you'd find in C# (though there are some interesting "clones" of it), Qt still uses moc. I have a personal library which practically clones the structure of Qt but is based on inheritance instead of signals and slots. It works really well and Nokia/Qt should consider switching to that model as well.
Second, Qt draws everything itself. It does not make use of native system form controls which has always been a huge point of contention in the Mac world. Qt does a great job, but because of the piss poor support from Apple for making accessibility features easily implementable, DIY form controls are a bit of a nightmare. It was a hair better in Carbon, but in Cocoa, it's no fun.
Now, you're in the right direction for iPhone development. WebKit is based on a clone of Qt Core which is really the best part of Qt anyway. You can do almost all your coding on another platform and then port back to Mac with little effort thanks to Qt core.
If you're making a cool program instead of some form controls based crap, then in reality, you're probably writing the entire UI yourself anyway. For that, you can easily make use of Qt and QGLWidget as a base. QGLWidget is insanely easy to port to NS classes.
I write all my iPhone stuff by doing all the development in Windows in Visual Studio (2008, but will move to 2010 now that Qt supports it). I write everything using OpenGL ES. I use the really great tools from NVidia and AMD (as well as others like Unity) to develop shaders which I use for entirely inappropriate purposes such as "because I can". Then I use Window CE based phones running a Samsung ARM processor to optimize my code for iPhone. When I'm done, I load up the project in XCode, toss it into a skeleton application, compile and debug for iPhone.
Best part is, for multiple developers you can use a single Mac Mini running iRapp to integrate Mac OS X into your Windows desktop and you never even have to code on Mac (which I never liked unless it's from vi on the command line). I guess you could also use a hackintosh as well, but a Mac Mini is OK for an occasional compile and a little debugging. A bit on the pricey side though.
The next "nasty nasty" is how to support Android like this, I do this in a similar fashion, but then I don't even need the mac to do it. Just code everything in Visual Studio, then create make files to compile the native code for Android. Then use whatever crapheap Java tool you want to hack the Android UI together, make some JNI calls to start the library and you're good to go. Expect to have at least 20 Android devices near by to test on though, you can't write native code for Android and expect it to function properly on different devices.
As for CrapBerry, you're basically screwed altogether. There are some piss-poor solutions to use GCC to compile for a Java back-end, but it's crud. The only thing I can recommend in this situation is that if you're starting a new app, keep the code as "portable as possible" meaning so that you can either preprocess your C++ code to Java and polish it or make your C++ code look exactly like Java for the most part.
Qt is an awesome platform, but it looks like it will be a long time if ever before there's such a thing as a good cross platform solution for phone development. Hell, I have problems writing code that will work on two different Android devices.
These are developers making freebie apps based on web standards.
Personally, I use iPhone myself, but I have Android and am looking forward to getting Windows Phone 7. When I write apps, what's interesting to me more than anything else is my ability to port the basics and write the app around them. On a Java based platform, my hair and fingernails itch just thinking about porting tens of thousands on lines of C++ code to a platform which doesn't guarantee a processor type. I don't mind screen resolution as I do as much as possible with OpenGL as I can, but screen format is a problem. I don't like the aspect ratio problem.
I know my cost of development for iPhone is SUBSTANTIALLY less than for Android as I have all models of iPhone and will probably only need to buy one per year to support it. Testing on all models is pretty easy to do. Android on the other hand is more of a hackers platform where you write it for the phone you have and if it works for someone else... great. If it doesn't, well, you'll need to get another phone to test with. These phones unlocked range from $200-$1000 and for any decent level of quality control, you'd need at least 20 of them today with people testing on all of them. In another year, you'll need a lot more. When Android is shipping on x86 hardware, then you'll need to make your code compile for both x86 and ARM which isn't that big of a deal except for alignment issues, but still then you're maintaining two executables and you have to buy even more devices.
Android doesn't guarantee :
- Screen resolution
- Graphics processor
- CPU type/generation/feature set (vector unit)
- Audio chip
- Audio latency
- Input method
- Multi-touch vs. Single touch
- Keyboard type.
- OS interaction with running applications
So, for anything more advanced than the basics webby kinda stuff, Android is a nightmare.
Cost of development for Android (with some QA) is MUCH higher than for iPhone.
The key is to write the application for iPhone, Android AND Windows 7 and leave it up to your paying customers to decided which one you'll focus the most QA on because of the best profit stream.
If you're starting from scratch on a new app, I guess it doesn't matter that much, you can work around it by using MonoDroid, MonoTouch and Windows Phone 7 SDK. With some work you can port most of your libraries to a CIL assembly. C++ doesn't port so smoothly there, but it's not that bad either. I am hoping that Mono eventually becomes the champion of phone development. It seems well on its way.
Webkit was a special case. Lars Knoll (last I heard who now works at the company formerly known as Trolltech) was the author of it. Smart guy, very nice too.
:)
:(
WebKit was very little more than :
- A) Stripping as much of Qt as possible out of konquerer.
- B) Rewriting what was left of Qt from "compiled... oops error, need new function".
- C) Wrapping QWidget as an NSWidget
When people started complaining, Apple took the code base and isn't giving back. Then Apple tried giving back, but seemed to do a terrible job of it. But because of customer demand, they kept adding to and fixing webkit. Eventually, everyone just started using WebKit... including Trolltech and Lars Knoll
I worked with Apple on browsers before the webkit decision was made and shortly after Microsoft forgot about IE for Mac. Trust me when I say, they wanted two things from a browser. Some control over the code base, the ability to not have to screw too much with open source licenses and headaches. Their WebKit project gave them just that. Given the importance of the browser to any platform, they made a great move even though it looked awful at first.
Now thanks to Google and Chrome being ported to Mac, it looks like there might even be a good, stable browser on Mac someday. As an occasional Mac lover, I look forward to getting more stable software on OS X. Bells and whistles are nice, but I really would rather them focus on getting the basics stable. Maybe start using progress bars or status messages to suggest that the computer is working and isn't actually locked up.
Boot time really needs improvement too. Pretty bad that Windows 7 restarts twice as fast on my MacBook as OS X does
As a developer you'll be far more valuable if you're able to deliver and app no matter what the platform is.
The idea is, the platform, the language etc... is entirely irrelevant for an engineer. My personal definition of an engineer is either "A person who can produce the most complex solution for the simplest problem" or more applicably to the discussion "A person who can identify a problem, break it down into smaller easier problems, posit solutions for each one, design and/or implement each solution therefore solving the problem".
Writing an application is a big problem. You need to identify "Do we need it?", "What should it do?", "Desktop, mobile, specialized device?", etc... once you've identified that you're interested in writing a mobile application, you can then identify whether it should run on just one platform or be platform neutral. If it's platform neutral, can a "core functionality" be identified and written as a library which will compile for any device of interest. Then you can implement the UI for each device as needed.
Problem is, unless you're using Adobe tools or maybe Unity3D (for the moment, which means you'll have to reinvent some major wheels at times) there isn't a common programming language for the mobile platforms. Apple has C/C++ support (then stick it in a ObjC wrapper), Android has C++ support... with effort, then make a Java wrapper. Windows Phone 7 tries to pretend like C/C++ doesn't exist, but I've heard of success of using it through managed extensions. Symbian if you actually consider that platform worth bothering with has C/C++ with Qt which is kinda nice. And Blackberry (another annoying waste of time) will require a bunch of rewriting everything for Java. (I hear you can use C/C++ but at your own risk). Brew doesn't even count, that platform is a programming disaster.
People are trying to make cross platform solutions for phones, but Apple, Google, RIM, Nokia etc... are making this as complicated as possible. MonoTouch is okish and Air is also okish, but you'll have limits on them both as opposed to programming natively for each.
What I have been doing is to write all my core libraries in C#, but "typedef" all your types to something native to your code. Then for strings, subclass String and add functionality to it to make it work like the one from Qt. C# is great because it borrows so heavily from C++ AND Java. If you code like this, then you can make a simple script that will convert your C# to either C++ or Java without any intervention and then porting is EASY!!!!
Intel will have an offering which provides equal performance for approximately the same price. I love AMD, thanks to them, it makes it possible for me to buy new Intel chips for less money all the time.
The plane would only be transparent so long as the vomit doesn't chew through it.
Let's not forget the "do all your shopping in your pajamas" effect. While I can buy most of my stuff online these days, and I can even save enough money that I can buy three different models and toss the two I like less because of it, I still get off my ass and go to the store... while I can still get off my fat ass that is.
If you need to go out and shop around to find what you're looking for, there is walking involved. There's bending over and picking things up. Sure, this doesn't sound like a lot of exercise, but 2 years ago, I went from 200lbs to 152lbs in 4 months by doing a little of this and a little of that.
I found simply being out and about made me eat less and eating is something a lot of people do when "I'm bored" sets in. The exercise from walking alone isn't just a matter of exercise, but it's a means of doing something other than eating.
Shopping in a mall probably wastes gas in the car and pumps carbon into the atmosphere. But shopping from your couch adds wrappers to the trash bin, not to mention all the other paper involved. As you're shopping from your couch for things like radios and ipods, eventually your ass gets big enough that you start shopping for pants and shirt because you're embarrassed to go into the Big and Fat shop. Of course, if you find yourself sitting next to a skinny little hottie on a bench while she's nipping on a yogurt with muesli and you're chopping down on a triple whopper with extra bacon, you should take the hint then and there.
The environmental impact is much much worse from shopping online than the article leads. People stop moving, they get fat, they buy bigger cars to haul their fat asses around and they buy bigger articles of clothing and guess what... they order more stuff online because moving their fat asses is even harder than before.
Now that I'm "in shape" I loaded up a bag 50lbs of weight (about the same as I lost, 23ish KG for the rest of the world) and I went for a 5km walk with it on. I damn near died from that. I have far more muscle mass now than I did then, but carrying that weight and walking a decent speed was bordering on impossible. If I had that weight back for real, I'd lay in bed and order everything online and say screw it. I'd have a house completely filled with blister packs and cardboard because I wouldn't even bother moving my ass to bring it to the trash can.
Should be ban shopping online because it's killing the planet? Hell no, I love shopping online and it's keeping the guy who has to move my 100lb boxes around from getting a fat ass. But we certainly should consider a way we can get fat assed people out of their couches as well.
I have purchased approximately a thousand DVDs over the past 14 years. I have a room in my house damn near consumed by them. They take tons of space. Before that, I bought tons of VHS (possibly literally, they were heavy)
:) Oh.. this is Norway, wouldn't matter.
What I hate is :
1) buying on VHS then having to pay again when I want DVD, then again when I want Bluray, then again when I want downloadable, etc... I have pirated films that I have already purchased on earlier formats. If the industry wants a proof of purchase for every film in my library and a serial number for it, cool... where do I send it. Then when they release it in a new format, let me download it.
2) If I can't buy a film downloadable in another country, I simply download it elsewhere. I'm perfectly happy to pay for films I watch. But I sure as shit won't be treated like a 3rd rate citizen because they decided where I live isn't convenient to them to sell in (Norway is too small to interest many companies).
3) If they are price gouging because of my location (very common, BluRay in U.S. $14, in Norway $45) and they won't let me download it at U.S. prices + Norwegian taxes, then I'll pirate it today and wait for it to land on the $15 rack in clearance down the line.
As I said, I've bought about 1000 DVD's at an average price of $22 (have a spreadsheet) a disc. That's $22,000 U.S., if they want to call me a pirate for trying to give them money but refusing to pay a 200% premium for my location, then send the cops my way
I actually work in the industry as well. Used to develop software at a post production shop that produced DVDs for the Scandinavian market for WB, BV, etc... Disney used to ask me to rip DVDs for them because they couldn't get their hands on their own films due to licensing limitations. So to make local trailers for films, they'd send my pre-release DVDs and ask me to rip them for them.
I often pirate video games, can't find them in the local stores and I won't pay $15 shipping for a $10 disc... I use cracks constantly, I buy games and use cracks so I don't need to leave the disc in.
Oh... I have all 1400 discs ripped to my server so I don't need to screw with finding discs. Also, I'm bloody tired of paying over and over for the same film if my kid scratches the disc. So, both my Wii's are cracked and play from hard disc, my playstation 2 too.
At the moment, I can't hook my DVD player's HDMI up to my projector since it's a cheap projector with only VGA on it. It has RCA also, but when I use it, the speakers in the room start buzzing as the chasis of the DVD player isn't properly grounded. The HDCP crack will solve this. HDFury is a half assed solution, this will allow real solutions instead.
2 needs a brief revision. It wasn't specifically about cheap embedded devices, it was entirely about small and simple.
The issue was that they wanted to encourage (at pain of death) the vendors to embed the HDCP encoding hardware directly into the video decoding chips. The HDCP encoder needed to be tiny enough that it would take only a minuscule portion of the FPGA/ASIC containing the video decoder. This made it so vendors would agree to do it. If they would have implemented a more complex system similar to ZRTP (one of the more complex, yet secure streaming encryption systems) first of all, there would be tons of compatibility problems. Vendors of video IP would have tons of problems implementing it and the differing interpretations of the spec would guarantee uselessness of the spec.
Additionally, it should be able to be implemented using pure logic, arithmetic is expensive. A full adder adds a huge layer of gate depth into a chip. Each bit adds at least one layer of depth. This adds latency etc... a logical algorithm can perform an entire step in the encoding/decoding chain in one level of depth. Also, logic operations are REALLY simple to route in ASICs and FPGAs. Non-power-of-2 multiplications and divisions would add loss, not only damaging the quality of the signal, but possibly the control codes, rendering the output undecodable.
By placing the HDCP stuff directly within the ASIC/FPGAs of the decoders (the primary goal as earlier noted), it made it so that while the algorithm itself was generally weak to begin with, the keys would be much harder to obtain as they would be coded directly inside the ASIC. Also, it makes it impossible to intercept the decoded signal between the decoder and the encrypter.
DVD CSS is an entirely separate interesting case. I was attempting to license the DVD specs to make a licensed DVD player back before DeCSS was released. I was reading through the requirements for licensing and find it amazing that almost all the interesting bits have disappeared from the face of the planet. Here's some things I vaguely remember from it, they may be imprecise :
1) The spec was made by Intel (a pretty smart group of guys normally)
2) The spec itself cost $5,000 + delivery via a secure physical method.
3) The spec was licensed to a company, but only after meeting certain security requirements.
4) Only "cleared individuals" would be permitted access to the spec.
5) The spec was to exist ONLY within a sealed room, no windows, heavily locked door.
6) The door would only be able to be opened by individuals cleared for access by the DVD consortium.
7) When the cleared individuals were not in the room and at any time the door was open, the spec would have to be stored in a fireproof safe which limited access only to individuals (the ones cleared to be in the room).
8) All source code for implementing the spec would exist ONLY on a removable medium (floppy, removable hard disk, zip disk, this was before thumb drives).
9) All binary implementations of the code would be heavily encrypted and could not be decrypted or deciphered without the use of a multi-million dollar laboratory environment.
Oops! That's where the failure was. So long as code can be run, it has to be in a decrypted form. So once the code is decrypted to be executed, then ANY debugger which can function on an application in memory could be used to read the code in assembler form. The complexity of the algorithm is irrelevant, all that matters is the amount of time a person has to sit and sort it out. The film industry wanted this clause in, but I'm entirely convinced that the guys who made the spec knew before hand it was entirely unrealistic.
You'll note that they focused a great deal of effort on the protection of the physical specification. I'd actually be amazed if the formal specification even exists in electronic form anymore. I'd imagine that it's strictly in print and is photocopied for each new licensee if there are actually any.
Seriously, I would be a huge fan of cutting back violent sport for anyone that has enough testosterone to believe that yelling at an inanimate device such as a TV set would have any impact positive or negative to the outcome of the event which they're watching.
How do you ban selling violent games to kids that have parents who will simply buy them for them? What defines a violent game? Super Smash Bros is EXTREMELY violent. Should that be banned? For tht fact, I remember "Nintendo Duck Hunt" having been quite violent as the only object of the game was to kill animals with a pistol. On the other hand, my 8-year son has friends who rush home each day to play Call of Duty. Their parents apparently didn't believe they were violent. Some of the 10 year olds in my neighborhood play Grand Theft Auto, their dads are sure they know that it's not real.
Why do we waste time on laws that are a waste of time outside of the theory. We're making laws that to try and do the things that parents should be doing instead. If the parent lets their kid go out and buy a violent game, and the kid almost definitely got the money for the game from mom and dad, how hard is it to get their mom and dad to simply buy the game for them? Of course, most of those kids can't get the money from mom and dad for the game, so they just pirate it instead. So what you're really saying is, to the few kids that are honest enough to actually buy a copy of the game, they should pirate it instead?
Tell you what... when you figure out a way to keep 11 year old kids from looking at porn, I'll support your bill.
Apple's model for the telephone market is almost definitely better for application developers. An application developer can buy 3 models of phones (and iPad if you care about that thing), test on each and be sure that everything works. The fact that iOS is such a closed platform is fantastic and makes it so that we developers can be more confident of what we ship to the public. It also means that we can optimize code to run well on all the phones which run that operating system.
Android on the other hand is more like the next step of Symbian... with slightly better design and control. Symbian was a heap of shit for developers. The API was a nightmare, content delivery worked only sometimes. Their package management system was a tinker toy. Additionally, their memory model was designed with a 25 year old PDA in mind, and their argument for it was that it needs to work with GCC 2.91. They implemented an ad-hoc exception model with a "clean-up stack" which was a lame excuse for auto-pointers as 2.91 didn't have good template support.
Android on the other hand has a relatively simple development model and it seems as if application development (so long as native code isn't important) is really quite easy. You can code in their Java like language (I do this to help with the law suit to differentiate and call it something else) and make an app and get it running quickly. Unfortunately, it runs on about a billion different processors (there are tons of ARMs out there) with a gazillion (quite cool that word is in the spell checker) graphics subsystems out there (nVidia, frame buffer, TI, etc...) and there are a multitude of different types of touch screens (single touch, multi touch, hi-resolution, low resolution, no-touch, just joypad, high latency, low latency). There are a pile of audio subsystems, I won't even begin to cover the massive number of those, it's mind boggling.
Writing simple cook book and business apps for Andoid is a charm. Takes far less time than on iOS, almost as little time as on Windows Phone 7 (which is WAY EASY) and can be tested more or less in an emulator without any problems. The only issue is the touch screen input which can be averted by making the buttons all a little bigger.
Anything requiring high response rates, fancy input methods, real-time audio, etc... is a nightmare on the platform. It's even worse than on Windows. There are just too many methods of input.
Android is a pretty neat touch screen platform that allows absolutely any manufacturer out there to make a full blown smart phone for almost nothing. Chinese vendors are already pumping these things out by the truckload and it's only a matter of time before it's possible to buy full smartphones for $50 or less.
You can buy an after market iPhone screen and touch panel from China for $20 (free shipping). And they are pretty good replacements. This means that they can get them for less than half that. Cheap system on a chip ARM processors can be bought for less the same. It's entirely possible that you can get ALL the parts required to make a full Android phone in China for probably $30. The specs will be pathetic, but will improve rapidly over time. The result, an Android phone containing the bare minimum memory required to run the phone, the bare minimum CPU required to run a telephone call, the bare minimum audio quality required to hear the other person, probably not even enough specs to download an application.
Of course, noone would buy these phones right? Well, probably not more than 100,000 of each model (which is the target Nokia sets for their mid-range smart phones). Remember there are a shit load of Asian people buy Chinese knockoffs of all these things. And what's best is, these aren't even knock offs. Thanks to the open source nature of Android, it's 100% legitimate to make these things. Of course, no westerners would buy these things. Umm... or would they. DealExtreme.com will sell tens of thousands of these. They'll be sold all over the Mediterranean and Caribbean islands to t
Seriously, who gives a shit about a telephone that can do more than an iPhone or whatever can now. 720p 30fps video is certainly good enough for a mobile phone. 1080p is only useful when we start using the phone to replace PCs. For that we need a PC operating system.
When will Mac OS X run ALL Mac software on the iPhone? Or when will <insert vendor's name here>'s phone start running Windows 7 64-bit. I want a phone with dual core, 1.0+ Ghz or better x86 or something with a dynamic translator that can run Mac OS X or Windows 7 64-bit. It needs wireless USB, wireless HDMI and my office desk chair should have an inductive charger in it so I can leave it in my pocket while I'm working.
Until we have that much, then who gives a crap about dual core processors on telephones? Last I checked my Mario DS runs just fine on a 67Mhz ARM9 CPU with a cheap crap GPU. Need for Speed runs great on a 412Mhz ARM 11 on my old iPhone 3G etc...
If they want to speed up the H.264 CODEC to handle 1080p, that's actually quite nice, then I can use it for high def movie playback, but unless they add enough storage space to make 20+ Mbps practical, I think I'll stick to 9Mbps 720p. Why more than double the pixels without at least doubling the bit rate to support it. Good quality scaled 720p is much better than crap quality 1080p any day. In fact, I'd rather have 1080i to get the higher field rate.
I think I'll wait patiently for Atom based phones before bothering with dual core in a hand held device. All this crap in-between is a hold me over for the real thing.
I don't actually have an X-Box 360, haven't figured out what I would need one for. It doesn't play Blu-Ray, almost every 360 game worth playing will also be released for Wii or PC, so why have yet another box.
On the other hand, I have 5, YES 5 360 wireless controllers. Two on my game machine hooked to the projector. Two on my 8 year old son's PC and one on my daughter's PC. The X-Box 360 controller is the best thing ever to have to gaming on the PC. Finally a standard PC joystick. It only took a billion years. Last time there was a standard game controller for the PC was the original two button, analog axis controller in the 80's that you needed a separate A/D card (there was no sound cards yet) to connect to the machine. It was the biggest piece of crap ever, but the 360 controller is AWESOME. A standard console grade controller for the PC.
Well that said, once I got these controllers, I bought all the Lego games (Star Wars, Indiana Jones, Batman and Harry Potter), I bought Tomb Raider and piles of other games. After all, the PC just became the most powerful console and the best part is, it plays games from nearly every old console and it also plays Blu-ray etc...
Well, my kids LOVE playing games on the PC with their controllers. When the game says "press the green button", they press the green button. In the room with the projector (a cheap little 100 lumen BenQ Joybee) you can see the colors, but not the letters on the buttons. Also, since the projector typically is downscaling 1280x1024 to 800x600, it is really quite hard to read letters on the screen (even when playing Wii). So, the colors are GREAT.
Here comes Microsoft to tell us that if any of our controllers break, we'll be replacing them with controllers of lower usability. Oh well... good thing couldn't last forever could it? Let's just hope that Logitech and others don't change.
Yep, was thinking that this will be a big boon to the GPS tracking community. Unfortunately, that cops will probably be able to use AGPS as well, which means you'd have to block your cell phone signal. If you're actually in need of a device like this, then there's a really good chance that you probably need to use your phone (probably a tosser) to arrange where you're trying to get away to, book plane tickets, etc... so, you're kinda screwed then. And as a thought, the tracker is probably GSM based anyway, so AGPS would be a much cheaper solution for cops tracking more than one "perp".
Besides, the link you pointed us to is kinda cool, but the prices are obscene. GPS is really easy to block, so it shouldn't take more than a $20 frequency generator with a shit load of bleed. An old CB radio almost certainly can be modified to do this and that would cost nearly nothing.
Sounds like the HTC shift which runs Windows CE in one mode and Windows XP in the other
libc assumes and operating system. An operating system requires system resources, many of the systems at hand actually don't have room for that. For example, the systems you would use in a DVR that run on 100 milliwatts to control the power of the main DVR. These often function as the real time clock for the system as well. I won't suggest you're using a real time clock dependent leap second precision in a DVR environment, however I will say that there is date and time with second level accuracy. These systems are either written in C WITHOUT libc or they're written in assembler.
I often find myself writing Atmel assembler for tiny little chips that I use for configuring 3rd party ASICs. They're really useful as you can use a 2 wire interface to control a device that typically requires 8-16 wires to configure. Atmel chips that are too small for C cost about 20 cents in bulk or over a dollar for the same functionality but C programmable. Since we at least one of these on each daughter board in our devices, it does add up (there are sometimes 5 boards). And since we're making hundreds of thousands of them, we DO save a ton of money that can't be accounted for as "the development time to code and maintain assembler would make up for the additional chip cost".
One of those chips... is a real time clock with network clock synchronization and 0.1us precision. Don't assume that just because this is the 21st century that everyone has the tools available that you consider to be "bare minimum requirements". Hell, if resources were readily available, I would port all our code to C# and write a TI DSP optimized tracing JIT back end for the Mono ILM. But frankly, that's just not the way the cookie crumbles.
Non-idiot programmers have these awesome computing devices with operating systems and tons of ram and hard drive. Sadly, many of us idiot programmers work on devices that have just enough code to run the initial system check and then if you want something you need to compile it and do it yourself.
Of course, we could use libraries, but the boss doesn't want to buy a library for each thing and the open source library has a GPL license or a "not for use in commercial software" or whatever license. Or there's a "just give me credit in the program somewhere" license... umm there's no UI... where can I do that? Engrave it on the cover?
Or we could just open source everything we do, then the first company in China that decides it would be easy to build now that the software is done can make it and sell it on dealextreme for 1/20th the price.
Nope, we write it over and over and over. We like feeding out families.
Let's face it, in we've found copies of the Epic of Gilgamesh dating to 4500B.C., the old testament (which is really more a monotheistic superset of gilgamesh) is from 1800B.C.ish. The real reason only "scholars" got the whole math thing is because they were the only ones who knew why the were counting down instead of up.
Nobo has a point... but it would make it so that the hardware engineers would suffer instead of the software ones. 1/86,4000 of a day = 1 second could be a fair solution. All we would need to do then would be to come up with a new atomic clock which allows for the alteration and then come up with computer crystals that are accurate to the new system (hey, let's get ones that are accurate to begin with, that would be great).
But, since respectable companies tend to run their own SNTP servers and they themselves adjust against national servers (we hope), it could simply be a good idea to ditch the leap second in favor of fixing all the clocks.
But, I think the real issue of the article is the occasions where "17:59:60" is a valid time. I think for presentation (and databases), it would in fact have been better to simply prolong 17:59:59 or progressively added a millisecond for the next 1000 seconds for example. Although it might through off scientific calculations during that period, the impact would be less critical.
The proper solution is to make programmers aware of leap seconds. There are 86400 seconds in a normal day, however there is an additional second added once or twice a year to adjust for solar time.
Wikipedia documents it quite well and programmers in modern times should be heading to wikipedia almost constantly anyway. The real problem occurs when the date/time is given in seconds since an "event" such as Jan 1, 1970. Most programmers don't know about leap seconds and I must admit, I don't generally bother calculating for them. But if it were necessary, it would be relatively trivial to do so.
We shouldn't remove fixes to the clock just because programmers are undereducated. I'm quite convinced that just posting this on Slashdot will raise awareness across a high percentage of the programming world.
I also always wondered why undergraduate studies for computer science didn't make time a relevant issue. It seems as if it's one of the more complex topics and yet, we don't pay any attention to it. Last formal education I had on time (not talking about physic related, but calendar) was in primary school. There are so many time systems out there that we should pay more attention to educating programmers on it.
Audio carried via the SPDIF spec either over copper or optical are obviously immune to classic interference style degradation of audio since it is digital media and you can tell if the audio has errors or not.
That said, if you're running cables in RF noisy environments, the chances of bit errors over copper are infinitely higher than over optical. There is truly a clear benefit to using optical over copper in this case. Same even goes for Ethernet cabling, when making a nice long run across a building where we couldn't avoid RF interference, we could not achieve full gigabit speeds over copper, so we switched to fiber and all was good.
A better comparison is using a $1.99 10 foot optical cable vs. a $100 10 foot monster cable optical cable. While the monster cable might be better manufactured and could theoretically survive a shallower turn radius, in reality, the quality difference between Monster and Radio Shack brown packs is zero.
The only time I spend extra money on cables anymore is for the ones I keep in my laptop bag. I prefer a more durable molded cable which bends easier (you know those nice soft ones that feel almost like microfiber on the insulation?) so that they don't get destroyed while I'm bouncing around my almost never well packed notebook bag.
sounds promising on the track expansion, was wondering though, can modern high speed trains operate over these tracks or would they require new ones. And if they make new ones, then would the new tracks be usable for cargo transportation? Still sounds terribly silly to me that from my understanding the railroads themselves are generally privately owned.
:)
Ans also getting back to the topic, what's involved with noise abatement? Would it not be satisfactory to recess the tracks, piling up the moved earth and stone to the sides of the rail ways... obviously, drainage will be an issue, but I know very little about modern drainage techniques. I'd have thought the greatest cost would be the massive amount of stone that would have moved and laid for the foundation of the railway.
Also, somehow 4-5 days for Washington to New York sounds painfully long to me. I took an amtrak from Spokane to DC one time and I remember it taking just under 3 days. Are cargo trains much slower?
Sorry for all the "what if"s and "what about"s and such, but you seem to have some good information, I wanted to exploit you
The #1 key to a strong economy is the ability to move from point A to point B. We currently use trucks FAR more than we should to move around freight that should be moved by efficient trains instead. Problem is, it's too expensive. Additionally, with budget airlines out there, taking a plane is cheaper than Greyhound most of the time. That's just STUPID.
The biggest shortcoming of trains is their dependence on rails. There's no way around this. After all, trains are... well trains. But, let's be realistic here, a single mega-train with hundreds of cars can move around substantially more than a fleet of airplanes... but they're slow and for shipping, it's more efficient to move things by plane. Trains are far smarter than planes many times because they often eliminate the need for hotel rooms for business travelers. You can take an overnight train to your business destination and take a shower in the morning and be ready for a meeting when you get to where you're going. Additionally, if trains sold their rooms in a similar fashion to hotels, so you could bring your whole family by train on a vacation and just rent a room on the train instead of paying per person, then using a 200Mph train, you could transport a family from NY to CA in 18 hours and because the trains don't have a million security checks before boarding, it would actually take only a little longer than using an airplane (it definitely works this way on shorter trips, tried in Germany). Additionally, the "lost money" from the ticket sales can be made up for on food, drink, etc... use RyanJet as a model for trains. Pretty sure they're going to start charging for using the bathroom pretty soon.
Problem is, you can't possibly put a train system like this together in the U.S. using normal government contractors for $8 billion since they'll just piss away all the money on the big-wigs. This is a great opportunity to invest in small business. Have one firm design the rail, it's not that complex (compared to things like making a space shuttle or airplane), then find small subcontractors that are used to performing services for fixed rates and eating the losses if they fall short. Give them enough seed money to expand and buy construction vehicles and get them each to build a portion of the track. The result would be creating lots of new businesses and business opportunities. Making companies across the country that are ready to rebuild the parts of America that are falling to pieces (Detroit anyone?) and the return on that investment alone will pick up a huge amount of the business.
If the larger government contractors start failing and having to sell off chunks of their business, these new businesses which are established upon the basis of getting jobs done for what they quoted in the first place will step in and pick up after them. The large government contractors are big, fat and sloppy and spill more money than they use productively into fat men's pockets. If they can't perform competitively in a modern market where the little guy can, then let them fall and the little guy will take over instead.
As for the trains, well don't reinvent it. Germans, Chinese, Japanese etc... all have them already. Just buy what you need from them or buy the designs. They've done all the work already. Just build the rails to meet the needs of their trains.
Unlike the current rail business which is a disaster from end to end, it's time the U.S. simply owned its rail system. After all, the U.S. owns the interstate highway system, how is this any different? If the rail companies can't make it work, then maybe they just need to fade away.
The end result is, the U.S. has a new rail infrastructure. Train owners can rent tracks from the government and therefore new companies can become involved without being stonewalled by greedy railway owners. A bunch of new little big construction companies are established and willing to do big work for reasonable rates without the bureaucracy, people have an alternative to flying and hotels, and the cargo busi