Reminds me of a bug some years back, before the Firefox/Thunderbird rename of Mozilla. I lost all my settings and emails, and discovered that the reason was that all settings "*.js" files were written out line-by-line in place, without even any buffering. Any crash or system problem during the shutdown of the browser would leave a partial and corrupt user profile. On the next startup, this would be detected, and the browser would wipe your entire profile and reset everything to defaults for your convenience. Including all of your email. Archives too.
At the time, most people had POP3 email, and would rarely have a backup copy anywhere to synchronize with. There were something like 3 or 4 dupes of that bug, all many years old, each with literally thousands of comments from terrified users begging for help to recover their lost data.
It took me all of an hour to figure out what was going on just by using SysInternals Process Monitor to watch the behavior during shutdown. The fix would have been trivial, and was suggested by no less than about twenty posts in the Bugzilla forums: just write configuration changes to new files, and then 'swap' the file names with the existing files at the end. That in combination with the use of a buffered C++ IO stream or something similar would have also resulted in the shutdown process speeding up by almost a factor of a hundred.
I checked back every few months. That bug was there for at least three or four years after that. Sometimes old copies of the bug where closed as "will not fix" or "can't reproduce", bugs were merged, split, and thousands of new comments appeared from more panicked users. I suspect that in total, some idiot at Mozilla spent ten times as much time maintaining and rearranging those bug tickets than it would have taken to fix the issue in the first place.
I was almost tempted to try and fix it myself, but apparently several other people have tried but were shot down for not following Mozilla's rigorous coding standards. I laughed and laughed, and then switched to Outlook and never looked back.
97M pages at a generous 100kB per page is just under 10 TB, which costs about $1000 to store. Let be generous again, and multiply the cost of raw disk capacity by a factor of 100 to account for redundancy, hosting, rack space, and bandwidth... Nope, still only $100,000!
So, yeah, $97M is a bit much. The only way I can think to account for such a ludicrously high cost would be if they used an archaic manual technology, like making crude pigment-based marks on dead trees! But that would be ludicrous, it would make justice impossible to afford for the common man! Such a system wouldn't be allowed in a modern society, right?
An increasingly common quote I've been hearing from the consultants of technology giants recently is that "product enhancements" are only considered if they can be demonstrated to be critical to closing a sale.
It's absolutely asinine.
For example, paraphrasing somewhat, I once found a missing function that made an entire API useless. It was designed for manipulating objects, and there were functions for adding, changing, and deleting objects of several types, except one that could not be deleted. It's a simple mistake that can be quickly rectified with a hotfix. Nope. Sorry. We don't have any sales that would be affected by this. Err...
Microsoft products are riddled with abandoned, half-complete, and archaic code that nobody will ever improve or fix, because either no customer wants it desperately enough, or no manager within Microsoft cares, so nobody will get any gold stars for fixing it. Code that does boring things -- no matter how important -- gets no love. This is also where all those security vulnerabilities come, from ancient code that hasn't been modernized or even just looked over in a decade.
Don't believe me? Install Windows 8 Release Preview, and create a new ODBC connection using the control panel. That dialog box hasn't changed in something like 15 years. It's like a museum piece. The "Add new performance counter" window is the same story. You still can't resize it, even though many of the counter names are longer than the available space and can't be read.
This short-sightedness leads to products that are just layers and layers of ancient cruft that no current employee understands or is willing to even touch any more. Eventually the entire product becomes unsellable and has to be scrapped. With something as enormous as Windows, this could very well lead to the end of Microsoft as we know it. Of course, none of this is relevant to sales this quarter, so it doesn't matter...
This kind of stupid shit happens because management is not a science!
Just like how before chemistry we had alchemy, and before medicine we had blood letting, you're seeing the pre-scientific version of management. It shouldn't be surprising that it's a joke.
I've read a bunch of management books and the whole time I was screaming "CITATION NEEDED" inside my head. They're full of unscientific but plausible sounding mumbo-jumbo, and ridden with weasel words. No numbers. No studies. No control experiments to demonstrate an improvement relative to the existing gold standard. Probably because there aren't any standards to begin with.
Take stack ranking for example: how would you even begin to measure the effectiveness of such as technique, relative to other employee review systems? What units would you use? How would you run this experiment? I bet you can't answer this. I bet the person who came up with stack ranking can't answer it either.
Instead of trading in real-time, they just have to do what the banks have always done: batch processing. Collect buy and sell orders for an hour, and then process them all together in a fair way at the end of the hour.
Do you want to buy a stock? Put a buy order in at 13:45. Does someone want to sell the stock? They put a sell order in at 13:56. At 14:00, you get your stock and they get your money. Or more accurately, the database system will start batch processing at 14:00, and give your stock at 14:03 or something. It doesn't actually matter all that much how fast it all happens.
If you wanted to invest in a company for the next three years, a one hour wait is nothing.
If you wanted to sell your shares in a company that you've held for three years, a one hour wait is nothing.
If you wanted to flip a stock as quickly as possible to make a fraction of a cent on a dollar, then a one hour wait is an eternity.
Buying and selling will still be possible, HFT trading will not.
No new taxes required. No new fees. LESS hardware. SIMPLER software. NO chance of runaway side-effects from software trading 10,000x as fast as a human being's reaction time.
This is a trivial problem to solve. The technology is not the problem, the politics is. A small number of powerful people are making a lot of money by stealing cents on the dollar from the common man. They will not give this up without a fight. They will lie, they will cheat, and I suspect even kill to hold on their income stream. They're certainly not above bribery and vote buying.
If the "liquidity" provided by high frequency traders is valuable when performed on the order of milliseconds, then logically the extension is also valid: trading on the order of microseconds or even nanoseconds should be more valuable still.
Does that makes sense? Fuck no. It makes no sense whatsoever for a stock or a commodity to be traded in nanosecond timeframes. It's asinine. There is no possible way for this to make any sense whatsoever.
Hence, the only possible conclusion is that the original premise -- faster trades always provide more value -- is not true.
So, the question is, how fast do trades have to be?
HFT traders would argue milliseconds. Many people here have proposed seconds, or even minutes. Here's a thing to think about though: the markets close every day for hours, and the economy doesn't suddenly collapse due to this suspension of "liquidity". Hence, the logical conclusion is that trades could be performed as infrequently as once an hour, and nothing would really change.
I've had some really strange moderation in the past.
For example, a post pointing out that "cloud services" are dangerous for corporations outside of the United States. The US has spied on foreign corporations in the past to benefit their domestic interests. There is no reason to think such behavior wouldn't continue.
That got moderated into oblivion. It's one of the few times I've had a post go negative.
At the time, I couldn't figure it out, but in light of your comment it makes a lot more sense.
The linked post is technically correct, but wrong on several levels.
First, home charging doesn't have to be fast!
Second, where charging does have to be fast, at the electric equivalent of a gas station, special high-current and high-voltage power lines could be installed without issues.
Third, charging via electricity is not the only option. For example, I've seen designs where the battery is mostly liquid, and that liquid can be exchanged as fast as a car can be filled with gasoline.
The network stack was re-written for Vista, but never "tuned", because Vista was effectively a beta release of Windows 7.
On the other hand, the network stack in Windows 7 works very well, and can easily reach 100% of wire speed on gigabit Ethernet. I've seen it reach 95% of 10 gigabit Ethernet too, which is basically not possible with Windows XP.
I found a noticeable improvement in network copy speed when I switched to Windows 7, especially if the other end is also Windows 7 or Windows Server 2008 R2. If you're copying to and from some "NAS" with slow embedded processor running an older version of Linux or BSD, then the speed improvement won't be anywhere near as great.
Also, I noticed a substantial improvement in system performance by changing to 64-bit. The increased file-cache size does wonders when working with large files. For example, I have 16 GB of memory in my computer, which is enough to cache an entire HD movie file! Even if you have 4 GB of memory in XP, it can't use it all -- the kernel has an internal limit of approximately 570 MB for file caching.
My recommendation is buy an SSD at the same time. The combination of 64-bit, Windows 7, and an SSD will blow your socks off. 8)
Except neither you, nor Microsoft have figured out what to do for situations where the keyboard is not fully unavailable.
For example, connecting to a server via RDP, where the session isn't full screen. In that case, the "start" key will launch the start menu of the client computer, not the server.
Or, for example, connecting to a server via a virtual machine console that isn't full screen.
Or a remote server management system, like the ones IBM, Dell, and HP provide as standard.
In all of those situations, it's quite possible that the only available method to launch the start menu is to hit the 2x2 pixel "corner" of the screen, which is no longer in the corner of the monitor. Over a high-latency link, this is an exercise in frustration.
It's obvious that neither you nor the "designers" at Microsoft have tried to do any real work with the Metro UI.
Have you even used Windows 8, at all? Because this was literally the first thing I noticed when I ran it in a virtual machine!
Inherently with any software, there are sometimes bugs which is of course always going to be a frustration for support staff.
In which case, it's their job to pass it up to people who can fix it, the developers. Blizzard support never does it. NEVER. They explained that to me in writing. If they don't have a workaround already, you are shit out of luck as a customer.
While your mileage may vary between the representative you speak to, most of the people I worked with were all passionate about the games and about giving the best support they could. I know this sounds cheesy, but it made my day when I managed to help someone out with a really obscure issue, or that I got a compliment on the service I gave.
That's nice, but why would I care as a customer if support fails to help me... with a smile?
I don't know enough about your particular issue to comment on the real cause, but as the launcher is working fine on my 64 bit win7 installation, it leads me to believe this is only affecting a minority of those users meaning it could be a very hard one for the developers to track down. However, support requests costs them money and I would imagine people are being appropriately pestered to get it fixed.
Something that breaks 1% of the time because of a bug, has a bug in it 100% of the time. Deadlock and timing issues are notorious for this. YOU CANNOT BLAME THE CUSTOMER FOR BEING UNLUCKY! Saying that "it works for me" doesn't help the customer, who then just becomes even more angry.
I can offer some very generic advice, almost always, it was background program's or antivirus interfering with the game. Do try a selective startup with nothing else running in the background and see if it helps:
Are you fucking kidding me? Next you'll be telling me to try turning it off and on again.
I spent days writing carefully worded messages to Blizzard support explaining in painstaking detail that no amount of generic cut & paste support bullshit is going to help if the Launcher design is fundamentally broken. There are thousands of posts complaining about the same issue: The Launcher used to work, it was modified for the D3 and MoP betas, and has never worked since. Either thousands of Blizzard's customers broke their PCs at the same time in the same way, or Blizzard fucked up the re-design of the Launcher. Which do you think is more likely?!?
Also, keep in mind that for every forum post about that issue, there are likely 10, 20 or maybe a hundred other users with no problem at all and thus haven't posted there. It is one of those unfortunate thing about support forums, you tend to only see the problems and never all the other users with none.
Why would I give a shit about other users? *I* have a problem! Again, you're shifting the blame, which is typical support tactics from a company that doesn't care about their products or their customers.
Here's a better statistic for you, which you really should memorize: For every customer that complains, there's between a hundred and a thousand that have the same problem but didn't bother to complain.
I read that in a scientific research paper, and I have personally collected hard statistics over years to back that up: An automated crash reporting system that I managed would collect several hundred crashes for every support phone call regarding the same crash.
It's not a question of QA, the code is badly designed to begin with. No amount of QA or support will ever fix a bad design.
The design of the Launcher is staggeringly bad. It uses loopback TCP to talk to something like 4 or 5 processes. I mean.. what the actual fuck? Why would an installer require five processes to run, with complex network protocols in between them!?! Normal people would just use either a single process, or something like windows event based message passing, but nooooo....
The Launcher is a Rube Goldberg machine that deadlocks if you look at it wrong. For example, people on the forums have reported that it can't handle a modem or VPN connection -- I don't mean connecting via a VPN or modem connection (which should work anyway), but that it will fail to talk to processes on the same machine if you just have a VPN connection configured and unused! I'm guessing some dumbass had written code to "enumerate interfaces with IP enabled", and uses the first one returned.
I've never seen a human-readable error message out of a Blizzard product. All of their software either deadlocks, or returns a numeric error code. They have no exceptions, stack traces, or error reporting back to their head office. If something crashes, it's your problem, they don't care.
On top of that, the scheduled outages are just the most ludicrous thing I've ever seen. What billion-dollar company turns their Internet presence off for six hours a week!?! Haven't these people heard of load-balancers and rolling updates? Seriously, this is not new technology, it's not rocket surgery -- everyone else does it! When was the last time Valve's Steam was down for maintenance? I can't remember! Meanwhile, Battle.net goes down like clockwork. It has scheduled failures!
If any Blizzard employees read this: seriously, contact me, I'd be happy to fly over there, and for a low price of merely $2000 per day for my consulting time, I'll explain to your architects how grown-ups design software.
As someone who until recently worked in Blizzard customer support...
That must have been a frustrating job.
Just recently, I submitted a problem with the new Launcher (also shared with the Mists of Pandaria beta), which basically doesn't work on Windows 7 64-bit. It has deadlocks in it that are timing sensitive. For some people, retrying over and over eventually works, but for me it never works. I can only install the game by installing it on my work laptop, and copying the files across to my PC. It doesn't even work on the laptop every time, I usually have to retry three or four times. That's idiotic.
Blizzard support basically told me to fuck off and complain somewhere else. Eventually, I got them to admit that there was literally nothing they could do to support the product. They couldn't even submit an internal bug report! Unless there was a known workaround already, they wouldn't help. What's even more stunning is that they weren't coming up with workarounds themselves, they were just trolling the forum posts where the users had figured out a workaround, and then just pasting those as responses to tech support requests. They literally knew nothing that users hadn't already figured out in the forums.
I gotta ask: was Blizzard support always this useless, or is this a new thing? Is there any way at all for a Blizzard customer to actually get a technical problem with the game resolved, or should I just ask for my money back now, and save myself the hassle?
So basically what I said: its purpose is military. Blowing things up is one hundred percent wasteful, this new space plane just wastes slightly less.
Meanwhile, there are real problems to solve that are struggling for funding. Not just here on Earth, but in space too. Too many space probes that could have done valuable science have been cancelled for lack of funding!
What makes you think Fusion research is not being funded or actively being researched around the world? Until the science (ie. math) is deemed feasible off the blackboard why should the government throw 150B at them?
The whole point of doing the research is to get to that point. Asking for certainty before funding research ends up with no research and no progress. At this stage, the problems are fairly well known, and solvable, given sufficient money, which they're not getting.
And as far the space program goes why does nobody ever mention the US X3-B vehicles and derivative projects built using information collected from the old space shuttle program. There has has been a stealth capable unmanned reusable space shuttle in use for 2+ years with a manned version already being prototyped. Why should the US waste any more time and money playing delivery man to the ISS? Chances are good the US has already used the ISS project to gather, refine, and test certain technologies and processes so they can build an orbital satellite of their own capable of providing everything from human living quarters, labs, docking services, to serve as the starting point for orbital industry construction?
All of that is pointless. Re-usable space-planes are only worthwhile if they're cheaper than non-reusable craft, and none of the previous ones ever were (Shuttle, Buran), and none of the current or proposed ones are either. When you're paying $1000 per kg or more, bringing back tons of stuff back down to Earth is pure, unadulterated waste. Most space-plane designs are boosted by a rocket, and aren't rockets themselves. They're nothing more than a heavy wrapper for the payload. I mean, seriously, look at it, it literally sits on top of a traditional no-reusable rocket like a warhead! I guarantee you that it was designed either to return satellites for purely military purposes, or purely as a pork program.
I have heard of realistic designs for re-usable spaceplanes that would bring launch costs down, but the X3-B isn't it. I do think that putting $150B towards re-usable space launch technology that has a chance of lowering costs might bring a good return on investment, particularly because a lot of the proposals could also do double duty as hypersonic or semi-orbital passenger aircraft. The Skylon spaceplane comes to mind. There are other designs out there too, but they're all woefully underfunded.
My point was that going to the Moon is nowhere near as useful, and not cost-efficient. We wouldn't get anywhere near as much technology out of it that's useable on Earth.
In-orbit manufacturing is the same. You'd need a worthwhile destination outside Earth's orbit first for the economics of orbital construction to make sense. There is nothing in our solar system even remotely worth sending squishy humans to. Every planet except Earth is either a frozen rock, an overheated rock, a lump of ice, or a ball of gas. There's no life out there, no aliens to meet, no technology to pick up, no unobtanium to mine. There are rocks here, at more convenient temperatures!
Except that some of us actually use our laptops for their intended purpose, and move them around with us, which makes a dock or other full-size peripherals impractical. Also, even if I had a keyboard, this places the laptop further away from me, making the screen harder to read. Not all of us have 20/20 vision.
Unlike people who buy a laptop to just to leave it sitting there in the same location, I have a desktop PC for that purpose... with a USB keyboard.
Nope, that's still crap. Take a look at this photo, which shows the keyboard clearly.
All the keys are jammed up against each other. The space between the number row and the function key row is missing, the arrow keys butt up against three other keys, etc... In other words, it has exactly the issues as every other laptop keyboard.
I've looked: to my knowledge, no laptop manufacturer on Earth has ever made a decent laptop keyboard, despite oodles of room on most 17" models. No such thing exists.
The spacing on the Apple keyboard is the same between all the keys.
Take your hand off your keyboard, close your eyes, and then quickly press the F7 key.
You can't, can you?
I can on mine, because it's in an easy to find position. There's an extra large gap between F8 and F9, which is easy to find by touch, and then I move just one key to the left, and tada... F7 is right there!
Some of use all the keys on the keyboard, not just A-Z, which is apparently all that is required to write snarky comments as an anonymous coward.
I think it's a combination of momentum, apathy, and ignorance.
It's definitely worth it, because I can't be the only human being on earth who wants to use one of the only two input devices on a laptop for its intended purpose! If I'm willing to pay more for a decent keyboard, so are other people.
The problem is more likely that until relatively recently (compared to PCs), all laptops were 15" or smaller. Monitors still had a 4:3 aspect ratio, so there just wasn't enough space for a full-width keyboard in any laptop model. Hence, compromises had to be made.
Since then, wide-screen 17" laptops have become commonplace, but nobody has realized that with the keyboard width constraint gone, there's no reason to persist with the design optimized for smaller devices. It just hasn't "clicked" with anybody that this is an important compromise to undo.
I guarantee you that the first decent 17" laptop with a proper keyboard will cost on the order of $1 more to make, and it will be a best-seller for the corporation with the "vision" to make it.
Don't want to be a dick, but my Macbook Pro 17" has the same key spacing as my Apple Keyboard.
Like this one, with the speakers on either side of a tiny, cramped keyboard?
Or did your mean the new new 2012 models, which clearly have exactly the same keyboard for both the larger and smaller models?
Or maybe your point is that that Apple has made their normal keyboards cramped too?
Compare those to real keyboard, made for people who use them for typing: It has a gap between the numbers and function keys, the ESC button is separate, the function keys are grouped in sets of four, the arrow keys have a space around them in all directions and are normal sized, and there are dedicated keys for insert, delete, home, end, pg-up, and pg-dn, in the standard position, with a space around them.
By the way, I just compared the width of my laptop to my normal, standard-sized desktop PC keyboard. Ignoring the numeric keypad, which I never use, the laptop is 4 cm wider than the desktop keyboard. There is absolutely no reason why it couldn't have been made to have the exact same layout!
The doubly stupid part of this whole thing is that laptop keyboards are replaceable. They're manufactured as this little metal tray thing that can be separated easily from the rest of the laptop. Why don't manufacturers make half a dozen different layouts, and let people chose? Some people may want a numeric keypad, some may want dedicated ins/del/etc... keys instead, some people may want media-control keys, others might prefer properly spaced function keys, etc...
Slashdot Mirror
Among the more recent comments for that bug:
Happy 10-year anniversary! :o)
Hahaha... classic! 8)
Reminds me of a bug some years back, before the Firefox/Thunderbird rename of Mozilla. I lost all my settings and emails, and discovered that the reason was that all settings "*.js" files were written out line-by-line in place, without even any buffering. Any crash or system problem during the shutdown of the browser would leave a partial and corrupt user profile. On the next startup, this would be detected, and the browser would wipe your entire profile and reset everything to defaults for your convenience. Including all of your email. Archives too.
At the time, most people had POP3 email, and would rarely have a backup copy anywhere to synchronize with. There were something like 3 or 4 dupes of that bug, all many years old, each with literally thousands of comments from terrified users begging for help to recover their lost data.
It took me all of an hour to figure out what was going on just by using SysInternals Process Monitor to watch the behavior during shutdown. The fix would have been trivial, and was suggested by no less than about twenty posts in the Bugzilla forums: just write configuration changes to new files, and then 'swap' the file names with the existing files at the end. That in combination with the use of a buffered C++ IO stream or something similar would have also resulted in the shutdown process speeding up by almost a factor of a hundred.
I checked back every few months. That bug was there for at least three or four years after that. Sometimes old copies of the bug where closed as "will not fix" or "can't reproduce", bugs were merged, split, and thousands of new comments appeared from more panicked users. I suspect that in total, some idiot at Mozilla spent ten times as much time maintaining and rearranging those bug tickets than it would have taken to fix the issue in the first place.
I was almost tempted to try and fix it myself, but apparently several other people have tried but were shot down for not following Mozilla's rigorous coding standards. I laughed and laughed, and then switched to Outlook and never looked back.
97M pages at a generous 100kB per page is just under 10 TB, which costs about $1000 to store. Let be generous again, and multiply the cost of raw disk capacity by a factor of 100 to account for redundancy, hosting, rack space, and bandwidth... Nope, still only $100,000!
So, yeah, $97M is a bit much. The only way I can think to account for such a ludicrously high cost would be if they used an archaic manual technology, like making crude pigment-based marks on dead trees! But that would be ludicrous, it would make justice impossible to afford for the common man! Such a system wouldn't be allowed in a modern society, right?
An increasingly common quote I've been hearing from the consultants of technology giants recently is that "product enhancements" are only considered if they can be demonstrated to be critical to closing a sale.
It's absolutely asinine.
For example, paraphrasing somewhat, I once found a missing function that made an entire API useless. It was designed for manipulating objects, and there were functions for adding, changing, and deleting objects of several types, except one that could not be deleted. It's a simple mistake that can be quickly rectified with a hotfix. Nope. Sorry. We don't have any sales that would be affected by this. Err...
Microsoft products are riddled with abandoned, half-complete, and archaic code that nobody will ever improve or fix, because either no customer wants it desperately enough, or no manager within Microsoft cares, so nobody will get any gold stars for fixing it. Code that does boring things -- no matter how important -- gets no love. This is also where all those security vulnerabilities come, from ancient code that hasn't been modernized or even just looked over in a decade.
Don't believe me? Install Windows 8 Release Preview, and create a new ODBC connection using the control panel. That dialog box hasn't changed in something like 15 years. It's like a museum piece. The "Add new performance counter" window is the same story. You still can't resize it, even though many of the counter names are longer than the available space and can't be read.
This short-sightedness leads to products that are just layers and layers of ancient cruft that no current employee understands or is willing to even touch any more. Eventually the entire product becomes unsellable and has to be scrapped. With something as enormous as Windows, this could very well lead to the end of Microsoft as we know it. Of course, none of this is relevant to sales this quarter, so it doesn't matter...
This kind of stupid shit happens because management is not a science!
Just like how before chemistry we had alchemy, and before medicine we had blood letting, you're seeing the pre-scientific version of management. It shouldn't be surprising that it's a joke.
I've read a bunch of management books and the whole time I was screaming "CITATION NEEDED" inside my head. They're full of unscientific but plausible sounding mumbo-jumbo, and ridden with weasel words. No numbers. No studies. No control experiments to demonstrate an improvement relative to the existing gold standard. Probably because there aren't any standards to begin with.
Take stack ranking for example: how would you even begin to measure the effectiveness of such as technique, relative to other employee review systems? What units would you use? How would you run this experiment? I bet you can't answer this. I bet the person who came up with stack ranking can't answer it either.
It's simpler than what is done now!
Instead of trading in real-time, they just have to do what the banks have always done: batch processing. Collect buy and sell orders for an hour, and then process them all together in a fair way at the end of the hour.
Do you want to buy a stock? Put a buy order in at 13:45. Does someone want to sell the stock? They put a sell order in at 13:56. At 14:00, you get your stock and they get your money. Or more accurately, the database system will start batch processing at 14:00, and give your stock at 14:03 or something. It doesn't actually matter all that much how fast it all happens.
If you wanted to invest in a company for the next three years, a one hour wait is nothing.
If you wanted to sell your shares in a company that you've held for three years, a one hour wait is nothing.
If you wanted to flip a stock as quickly as possible to make a fraction of a cent on a dollar, then a one hour wait is an eternity.
Buying and selling will still be possible, HFT trading will not.
No new taxes required. No new fees. LESS hardware. SIMPLER software. NO chance of runaway side-effects from software trading 10,000x as fast as a human being's reaction time.
This is a trivial problem to solve. The technology is not the problem, the politics is. A small number of powerful people are making a lot of money by stealing cents on the dollar from the common man. They will not give this up without a fight. They will lie, they will cheat, and I suspect even kill to hold on their income stream. They're certainly not above bribery and vote buying.
It goes further than this.
If the "liquidity" provided by high frequency traders is valuable when performed on the order of milliseconds, then logically the extension is also valid: trading on the order of microseconds or even nanoseconds should be more valuable still.
Does that makes sense? Fuck no. It makes no sense whatsoever for a stock or a commodity to be traded in nanosecond timeframes. It's asinine. There is no possible way for this to make any sense whatsoever.
Hence, the only possible conclusion is that the original premise -- faster trades always provide more value -- is not true.
So, the question is, how fast do trades have to be?
HFT traders would argue milliseconds. Many people here have proposed seconds, or even minutes. Here's a thing to think about though: the markets close every day for hours, and the economy doesn't suddenly collapse due to this suspension of "liquidity". Hence, the logical conclusion is that trades could be performed as infrequently as once an hour, and nothing would really change.
I've had some really strange moderation in the past.
For example, a post pointing out that "cloud services" are dangerous for corporations outside of the United States. The US has spied on foreign corporations in the past to benefit their domestic interests. There is no reason to think such behavior wouldn't continue.
That got moderated into oblivion. It's one of the few times I've had a post go negative.
At the time, I couldn't figure it out, but in light of your comment it makes a lot more sense.
The linked post is technically correct, but wrong on several levels.
First, home charging doesn't have to be fast!
Second, where charging does have to be fast, at the electric equivalent of a gas station, special high-current and high-voltage power lines could be installed without issues.
Third, charging via electricity is not the only option. For example, I've seen designs where the battery is mostly liquid, and that liquid can be exchanged as fast as a car can be filled with gasoline.
Next year? Do you have a citation for that?
From what I've heard, ReRAM is nowhere near the point that it could be used as a viable replacement for flash.
And not a dime of it is going to the people who made the thing you're looking for.
That's their own fault.
Content providers refuse to accept money for the service that the customers want, while The Pirate Bay provides a superior service for free.
The network stack was re-written for Vista, but never "tuned", because Vista was effectively a beta release of Windows 7.
On the other hand, the network stack in Windows 7 works very well, and can easily reach 100% of wire speed on gigabit Ethernet. I've seen it reach 95% of 10 gigabit Ethernet too, which is basically not possible with Windows XP.
I found a noticeable improvement in network copy speed when I switched to Windows 7, especially if the other end is also Windows 7 or Windows Server 2008 R2. If you're copying to and from some "NAS" with slow embedded processor running an older version of Linux or BSD, then the speed improvement won't be anywhere near as great.
Also, I noticed a substantial improvement in system performance by changing to 64-bit. The increased file-cache size does wonders when working with large files. For example, I have 16 GB of memory in my computer, which is enough to cache an entire HD movie file! Even if you have 4 GB of memory in XP, it can't use it all -- the kernel has an internal limit of approximately 570 MB for file caching.
My recommendation is buy an SSD at the same time. The combination of 64-bit, Windows 7, and an SSD will blow your socks off. 8)
Aren't we Mr Insightful?
Funny too!
Except neither you, nor Microsoft have figured out what to do for situations where the keyboard is not fully unavailable.
For example, connecting to a server via RDP, where the session isn't full screen. In that case, the "start" key will launch the start menu of the client computer, not the server.
Or, for example, connecting to a server via a virtual machine console that isn't full screen.
Or a remote server management system, like the ones IBM, Dell, and HP provide as standard.
In all of those situations, it's quite possible that the only available method to launch the start menu is to hit the 2x2 pixel "corner" of the screen, which is no longer in the corner of the monitor. Over a high-latency link, this is an exercise in frustration.
It's obvious that neither you nor the "designers" at Microsoft have tried to do any real work with the Metro UI.
Have you even used Windows 8, at all? Because this was literally the first thing I noticed when I ran it in a virtual machine!
Inherently with any software, there are sometimes bugs which is of course always going to be a frustration for support staff.
In which case, it's their job to pass it up to people who can fix it, the developers. Blizzard support never does it. NEVER. They explained that to me in writing. If they don't have a workaround already, you are shit out of luck as a customer.
While your mileage may vary between the representative you speak to, most of the people I worked with were all passionate about the games and about giving the best support they could. I know this sounds cheesy, but it made my day when I managed to help someone out with a really obscure issue, or that I got a compliment on the service I gave.
That's nice, but why would I care as a customer if support fails to help me... with a smile?
I don't know enough about your particular issue to comment on the real cause, but as the launcher is working fine on my 64 bit win7 installation, it leads me to believe this is only affecting a minority of those users meaning it could be a very hard one for the developers to track down. However, support requests costs them money and I would imagine people are being appropriately pestered to get it fixed.
Something that breaks 1% of the time because of a bug, has a bug in it 100% of the time. Deadlock and timing issues are notorious for this. YOU CANNOT BLAME THE CUSTOMER FOR BEING UNLUCKY! Saying that "it works for me" doesn't help the customer, who then just becomes even more angry.
I can offer some very generic advice, almost always, it was background program's or antivirus interfering with the game. Do try a selective startup with nothing else running in the background and see if it helps:
Are you fucking kidding me? Next you'll be telling me to try turning it off and on again.
I spent days writing carefully worded messages to Blizzard support explaining in painstaking detail that no amount of generic cut & paste support bullshit is going to help if the Launcher design is fundamentally broken. There are thousands of posts complaining about the same issue: The Launcher used to work, it was modified for the D3 and MoP betas, and has never worked since. Either thousands of Blizzard's customers broke their PCs at the same time in the same way, or Blizzard fucked up the re-design of the Launcher. Which do you think is more likely?!?
Also, keep in mind that for every forum post about that issue, there are likely 10, 20 or maybe a hundred other users with no problem at all and thus haven't posted there. It is one of those unfortunate thing about support forums, you tend to only see the problems and never all the other users with none.
Why would I give a shit about other users? *I* have a problem! Again, you're shifting the blame, which is typical support tactics from a company that doesn't care about their products or their customers.
Here's a better statistic for you, which you really should memorize: For every customer that complains, there's between a hundred and a thousand that have the same problem but didn't bother to complain.
I read that in a scientific research paper, and I have personally collected hard statistics over years to back that up: An automated crash reporting system that I managed would collect several hundred crashes for every support phone call regarding the same crash.
It's not a question of QA, the code is badly designed to begin with. No amount of QA or support will ever fix a bad design.
The design of the Launcher is staggeringly bad. It uses loopback TCP to talk to something like 4 or 5 processes. I mean.. what the actual fuck? Why would an installer require five processes to run, with complex network protocols in between them!?! Normal people would just use either a single process, or something like windows event based message passing, but nooooo....
The Launcher is a Rube Goldberg machine that deadlocks if you look at it wrong. For example, people on the forums have reported that it can't handle a modem or VPN connection -- I don't mean connecting via a VPN or modem connection (which should work anyway), but that it will fail to talk to processes on the same machine if you just have a VPN connection configured and unused! I'm guessing some dumbass had written code to "enumerate interfaces with IP enabled", and uses the first one returned.
I've never seen a human-readable error message out of a Blizzard product. All of their software either deadlocks, or returns a numeric error code. They have no exceptions, stack traces, or error reporting back to their head office. If something crashes, it's your problem, they don't care.
On top of that, the scheduled outages are just the most ludicrous thing I've ever seen. What billion-dollar company turns their Internet presence off for six hours a week!?! Haven't these people heard of load-balancers and rolling updates? Seriously, this is not new technology, it's not rocket surgery -- everyone else does it! When was the last time Valve's Steam was down for maintenance? I can't remember! Meanwhile, Battle.net goes down like clockwork. It has scheduled failures!
If any Blizzard employees read this: seriously, contact me, I'd be happy to fly over there, and for a low price of merely $2000 per day for my consulting time, I'll explain to your architects how grown-ups design software.
As someone who until recently worked in Blizzard customer support...
That must have been a frustrating job.
Just recently, I submitted a problem with the new Launcher (also shared with the Mists of Pandaria beta), which basically doesn't work on Windows 7 64-bit. It has deadlocks in it that are timing sensitive. For some people, retrying over and over eventually works, but for me it never works. I can only install the game by installing it on my work laptop, and copying the files across to my PC. It doesn't even work on the laptop every time, I usually have to retry three or four times. That's idiotic.
Blizzard support basically told me to fuck off and complain somewhere else. Eventually, I got them to admit that there was literally nothing they could do to support the product. They couldn't even submit an internal bug report! Unless there was a known workaround already, they wouldn't help. What's even more stunning is that they weren't coming up with workarounds themselves, they were just trolling the forum posts where the users had figured out a workaround, and then just pasting those as responses to tech support requests. They literally knew nothing that users hadn't already figured out in the forums.
I gotta ask: was Blizzard support always this useless, or is this a new thing? Is there any way at all for a Blizzard customer to actually get a technical problem with the game resolved, or should I just ask for my money back now, and save myself the hassle?
So basically what I said: its purpose is military. Blowing things up is one hundred percent wasteful, this new space plane just wastes slightly less.
Meanwhile, there are real problems to solve that are struggling for funding. Not just here on Earth, but in space too. Too many space probes that could have done valuable science have been cancelled for lack of funding!
Wow.
Suddenly, the decision to have an explicit 'bool' type in newer languages makes a lot more sense!
What makes you think Fusion research is not being funded or actively being researched around the world? Until the science (ie. math) is deemed feasible off the blackboard why should the government throw 150B at them?
The whole point of doing the research is to get to that point. Asking for certainty before funding research ends up with no research and no progress. At this stage, the problems are fairly well known, and solvable, given sufficient money, which they're not getting.
And as far the space program goes why does nobody ever mention the US X3-B vehicles and derivative projects built using information collected from the old space shuttle program. There has has been a stealth capable unmanned reusable space shuttle in use for 2+ years with a manned version already being prototyped. Why should the US waste any more time and money playing delivery man to the ISS? Chances are good the US has already used the ISS project to gather, refine, and test certain technologies and processes so they can build an orbital satellite of their own capable of providing everything from human living quarters, labs, docking services, to serve as the starting point for orbital industry construction?
All of that is pointless. Re-usable space-planes are only worthwhile if they're cheaper than non-reusable craft, and none of the previous ones ever were (Shuttle, Buran), and none of the current or proposed ones are either. When you're paying $1000 per kg or more, bringing back tons of stuff back down to Earth is pure, unadulterated waste. Most space-plane designs are boosted by a rocket, and aren't rockets themselves. They're nothing more than a heavy wrapper for the payload. I mean, seriously, look at it, it literally sits on top of a traditional no-reusable rocket like a warhead! I guarantee you that it was designed either to return satellites for purely military purposes, or purely as a pork program.
I have heard of realistic designs for re-usable spaceplanes that would bring launch costs down, but the X3-B isn't it. I do think that putting $150B towards re-usable space launch technology that has a chance of lowering costs might bring a good return on investment, particularly because a lot of the proposals could also do double duty as hypersonic or semi-orbital passenger aircraft. The Skylon spaceplane comes to mind. There are other designs out there too, but they're all woefully underfunded.
My point was that going to the Moon is nowhere near as useful, and not cost-efficient. We wouldn't get anywhere near as much technology out of it that's useable on Earth.
In-orbit manufacturing is the same. You'd need a worthwhile destination outside Earth's orbit first for the economics of orbital construction to make sense. There is nothing in our solar system even remotely worth sending squishy humans to. Every planet except Earth is either a frozen rock, an overheated rock, a lump of ice, or a ball of gas. There's no life out there, no aliens to meet, no technology to pick up, no unobtanium to mine. There are rocks here, at more convenient temperatures!
Can you imagine a Manhattan project style funding of new science?
Imagine throwing $150B at, say, fusion power research.
Instead of a one-off trip to some rock, we'd be well on the way to eliminating global warming and boosting industry through lowered energy costs.
Once developed, the work required to construct thousands of fusion plants would keep far more people employed than the space program ever did.
Close, it has visibly larger gaps where it should, but obviously it's not possible to buy it any more.
It still doesn't have the arrow keys or the ins/del/etc keys in the right locations.
Except that some of us actually use our laptops for their intended purpose, and move them around with us, which makes a dock or other full-size peripherals impractical. Also, even if I had a keyboard, this places the laptop further away from me, making the screen harder to read. Not all of us have 20/20 vision.
Unlike people who buy a laptop to just to leave it sitting there in the same location, I have a desktop PC for that purpose... with a USB keyboard.
Nope, that's still crap. Take a look at this photo, which shows the keyboard clearly.
All the keys are jammed up against each other. The space between the number row and the function key row is missing, the arrow keys butt up against three other keys, etc... In other words, it has exactly the issues as every other laptop keyboard.
I've looked: to my knowledge, no laptop manufacturer on Earth has ever made a decent laptop keyboard, despite oodles of room on most 17" models. No such thing exists.
The spacing on the Apple keyboard is the same between all the keys.
Take your hand off your keyboard, close your eyes, and then quickly press the F7 key.
You can't, can you?
I can on mine, because it's in an easy to find position. There's an extra large gap between F8 and F9, which is easy to find by touch, and then I move just one key to the left, and tada... F7 is right there!
Some of use all the keys on the keyboard, not just A-Z, which is apparently all that is required to write snarky comments as an anonymous coward.
I think it's a combination of momentum, apathy, and ignorance.
It's definitely worth it, because I can't be the only human being on earth who wants to use one of the only two input devices on a laptop for its intended purpose! If I'm willing to pay more for a decent keyboard, so are other people.
The problem is more likely that until relatively recently (compared to PCs), all laptops were 15" or smaller. Monitors still had a 4:3 aspect ratio, so there just wasn't enough space for a full-width keyboard in any laptop model. Hence, compromises had to be made.
Since then, wide-screen 17" laptops have become commonplace, but nobody has realized that with the keyboard width constraint gone, there's no reason to persist with the design optimized for smaller devices. It just hasn't "clicked" with anybody that this is an important compromise to undo.
I guarantee you that the first decent 17" laptop with a proper keyboard will cost on the order of $1 more to make, and it will be a best-seller for the corporation with the "vision" to make it.
Don't want to be a dick, but my Macbook Pro 17" has the same key spacing as my Apple Keyboard.
Like this one, with the speakers on either side of a tiny, cramped keyboard?
Or did your mean the new new 2012 models, which clearly have exactly the same keyboard for both the larger and smaller models?
Or maybe your point is that that Apple has made their normal keyboards cramped too?
Compare those to real keyboard, made for people who use them for typing: It has a gap between the numbers and function keys, the ESC button is separate, the function keys are grouped in sets of four, the arrow keys have a space around them in all directions and are normal sized, and there are dedicated keys for insert, delete, home, end, pg-up, and pg-dn, in the standard position, with a space around them.
By the way, I just compared the width of my laptop to my normal, standard-sized desktop PC keyboard. Ignoring the numeric keypad, which I never use, the laptop is 4 cm wider than the desktop keyboard. There is absolutely no reason why it couldn't have been made to have the exact same layout!
The doubly stupid part of this whole thing is that laptop keyboards are replaceable. They're manufactured as this little metal tray thing that can be separated easily from the rest of the laptop. Why don't manufacturers make half a dozen different layouts, and let people chose? Some people may want a numeric keypad, some may want dedicated ins/del/etc... keys instead, some people may want media-control keys, others might prefer properly spaced function keys, etc...