1) How many of the people who reflexivly project hate at Java have solid, defensible reasons for doing so? Do you think people form opinions based on evidence and thoughtful consideration?
2) Examples, please.
3) Well C is not a long-lived programming language, then. COBOL, FORTRAN and LISP have been going strong since the 1950's!
3.1) Plenty. Most of the applications on my phone runs Java (Android). My cable box (OCAP). My Blu-ray player (BD-J). Several of the applications I administer or use professionally (IntelliJ, Confluence, Jira, Zimbra, JMeter).
Just because you don't see it doesn't mean you don't use it. And even if you don't use doesn't mean other people don't use it. Java is an incredible pervasive language in the embedded and server space.
3.2) And?
3.3) With a twelve year history, multiple implementations and no hint of a successor, this strikes me as needless fear-mongering. Microsoft is proprietary, not fickle. Most of their standards have excellent longevity and they have far longer support cycles for their products than most of the Unix world.
I suspect it is about establishing precedent and combating the idea that EV owners are entitled to "free" power, not about recovering costs in this specific incident.
That's because Democrats wanted people to believe default was a credible threat. It wasn't. No one who knew what they were talking about thought it was, hence the stability of interest rates.
Really, it's the sort of thing that you don't realize you want until you have it, then you are annoyed if you don't have it.
Even under the best conditions, half a second is a bit of an exaggeration, and you're often not under the best of conditions. The cable slipped behind the desk, or you realize that you forgot to plug your phone in after you turned off the light or took off your glasses, or you're driving and don't want to be the asshole who causes an accident because you were fiddling with a cablet instead of keeping your eyes on the road, or your cable is getting dodgy and you keep forgetting to order a new one.
And then there is the annoyance of having your phone go dead because you were too tired to bother flipping the light back on, or never did hit a red light long enough to plug your phone in, or you jostled your phone and it stopped charging.
Is it a need? Certainly not. But it is damned handy.
Big businesses get discounts because they sort their mail, prepare them for automated processing and sometimes transport them to postal facilities. In other words, they pay less because they cost less to service. You could pay the same price if you printed your mail with IMpb barcodes, pre-sorted it and sent it in batches of five hundred or more.
Commecial first class and bulk mail services provide the volume necessary to keep postal rates relatively low and routes open. Increase their rates, volume will drop, and everybody's rates go up.
The raw number isn't very meaningful. There are over 250 million passenger vehicles registered in the United States. That comes to one fire per 1641 vehicles. There are something like 12 thousand Teslas, but also only three fires, which comes to one fire per 4000 vehicles.
So with those numbers, they still seem better off than combustion driven vehicles. However, fleet size isn't the only variable. The majority of Teslas are less than a year old, while overall only 13.5% of passenger vehicles are under two years old, and 60.6% are at least seven years old. Teslas are also a luxury vehicle with a base price of $69,900 (before tax credits) and allegedly stringent manufacturing standards, compared to the overall market with an average new sale price of $24,764 and varying quality. And finally, maintenance needs to be taken into consideration. Not only are Teslas newer and almost certainly under warranty, the average owner is relatively wealthy and presumably can afford proper maintenance.
But even that is not necessarily the right set of stats to look at. The three Tesla fires occurred after an accident, so it may be more informative to look at the liklihood of accidents to result in fire. According to the NFPA, only 4% of automobile fires were the result of a collision or roll-over; most were caused by a mechanical or electrical failure. There were about 5.4 million accidents in 2010, so at 4% of 152300 accident it comes out to approximately one fire per 900 accidents. There are no hard statistics on precisely how many accidents there are for Tesla specifically, but it would have to be a significantly high number to have a lower fire to accident ratio. And that isn't even taking into account the relative severity of accidents into account.
Of course the real problem is that there simply isn't enough data available. Three times in as many months could be a fluke. Or it could be a pattern.
Apple managed to roll out iOS 7 - which was in development for only a year - successfully to over 140 million devices in a single day. Rapidly address bug aside, it is nowhere near the scale of mess that the HCX rollouts have been.
The launch of the online component of GTA V has certainly been rough, but we are again talking about a drastically different scale of problem. Online gaming is an extended, high bandwidth, latency sensitive distributed interaction and relies heavily on third party infrastructure (namely the Playstation Network and XBox Live). Going to a website to create a login or submit an application should be short, consume little bandwidth, be tolerant of far longer latencies, and involve only two-party transactions.
We are also talking a drastically different scale of budget. The budget for GTA V is mind-boggling for video games (~$265 million for development and marketing) but it pales in comparison to the mountain of money being poured into the health care exchanges. Kaiser reported that $3.6 BILLION had been awarded in grants to the states to set up the exchanges, and that number doesn't include state funding. California alone spend $313 million on their exchange. It is difficult to know just how much has been spent on the federal exchange, since not all spending is called out explicitly, but the HHS has awarded well over $600 million to GCI alone (the $50-something million number was just one component).
And in the end, the exchanges did not experience a particularly high level of traffic. Though not all states have released numbers, the ones we have are fairly modest. California, the most populous state in the union, reported a scant 514,000 unique visitors on Tuesday. That is NOTHING.
Horse pucky. There is no reason to assume there would be one (and only one) peak load. If anything, one would expect annual spikes during the enrollment periods. And the enrollment projections have all seen mounting demand year over year. So this is an invented dichotomy to begin with.
But even if one accepted that as a design precept, there is no need to simply fall down under peak. You make your UX components asynchronous and dump requests into a work queue. Return control to the user, give them a link to check on their status. Not only are these techniques well understood, they are neither difficult to implement nor particularly expensive. You can easily push a million messages PER MINUTE on a single box using free software like ActiveMQ.
And the idea that performance is the determining factor between SQL and no-SQL betrays fatuous. There have been high performance SQL deployments far longer than the current crop of "no-SQL" databases. For transactional workloads with stable data models, you're likely to take a performance hit trying to shoe-horn into the flavor of the week data store.
You also have to consider the repercussion of a malfunctioning system. Not only are you going to generate MORE traffic by being improperly sized for peak load (due to people reloading or making multiple visits), they also have manned phone banks to handle technical difficulties. Servers are cheap. People are not.
Long story short, this has been an epic fail of the sort that would be inexcusable in the private sector. And unfortunately for the tax payer, it isn't a matter of being underfunded. The Department of Health and Human Services reportedly paid $55.7 million to Canadian (!) tech firm CGI Federal for this steaming pile, and may pay them upwards of $40 million more during the tenure of the contract.
Pupil to teacher ratios have been declining for decades. Public schools are already down to a student/teacher ratio of 15.4, with a median class size of 20.0 for public elementary schools (according to the National Center for Educational Statistics).
Pay for teachers has more than outpaced inflation since the 1980's, rising from an inflation adjusted ~ $44k to ~ $54k. Once you factor in benefits, extended summer vacations (or additional income earned teaching summer sessions), pensions and the potential for tenure, the overall compensation picture is hardly unfair or unattractive. And the UNESCO statistics show that starting salaries are actually relatively competitive, internationally speaking; behind Switzerland, Germany,Demark and the Netherlands, but ahead of Australia, Spain, Norway, Ireland, Austria, Iceland, France, Greece, Italy, Portugal and Sweeden.
Money is good. Some schools are legitimately underfunded. However, it is not a panacea, and there is little evidence that too little money is pumped into public education. Consider this; the best funded school district in my area, on a per pupil basis, currently spends over $27,000 per student and achieves the absolute worst results, as measured by performance on standardized tests, graduation rates and college attendance. Other districts excel with half the funding. Parochial schools outperform with less than a QUARTER the funding. And nationally, home schooled children consistently out-perform their peers, in spite of per-pupil spending that is often measured in the hundreds, rather than the thousands.
You claim that the judge eliminated self-defense, not because it wasn't germane, but because of an alleged - but not established - sequence of events. That is simply fatuous, and demonstratably false. The jury was clearly instructed that George Zimmerman's claim of self-defense was the central consideration of the case:
"An issue in this case is whether George Zimmerman acted in self-defense. It is a defense to the crime of Second Degree Murder, and the lesser included offense of Manslaughter, if the death of Trayvon Martin resulted from the justifiable use of deadly force.
If in your consideration of the issue of self-defense you have a reasonable doubt on the question of whether George Zimmerman was justified in the use of deadly force, you should find George Zimmerman not guilty.
However, if from the evidence you are convinced beyond a reasonable doubt that George Zimmerman was not justified in the use of deadly force, you should find him guilty if all the elements of the charge have been proved."
That is not the standard of law in Florida, nor is it that pertinent. He wasn't escaping from someone walking down the street, unaware of him; he was escaping someone who punched him in the face, bashed his head into concrete and was straddling him.
(The pertinent section of the jury instructions: "If George Zimmerman was not engaged in an unlawful activity and was attacked in any place where he had a right to be, he had no duty to retreat and had the right to stand his ground and meet force with force, including deadly force if he reasonably believed that it was necessary to do so to prevent death or great bodily harm to himself or another or to prevent the commission of a forcible felony.")
This is roughly as stupid as when people claimed Florida banned sex a couple years ago. Read the bill. It clearly refers to devices that perform gaming function on activation; not general purpose devices.
Common sense, people. What is more likely? A hundred and forty legislators voting to ban computers, or a bogus lawsuit with a silly premise?
While systemd takes a drastically different approach to initialization than upstart, the similarities between mir and wayland far outweigh their differences, and most of the stated reasons for not adopting wayland seem based on misunderstanding rather than significant issues with the architecture.
Perhaps mir will be a winning move for Canonical. Perhaps it will even be adopted outside Ubuntu. I wouldn't bet on it.
Return works fine; trying to return from something that isn't a method doesn't. The most striking thing about this question is that the submitter never explains what they are doing or why they insist on doing it in a non-idiomatic way. "I'd really prefer to be able to use 'return' in this scenario" is not a compelling argument. What reason did he have for trying to work around a reasonable, well established language paradigm?
As far as people abandoning Ruby, I'm not sure it is more or less common than other languages. There have been a couple high profile "abandonments", like Twitter, which seem to get paraded out as proof that the language does not scale. There are a number of countervailing points, however; specifically:
1) Twitter, even back when they first started their migration of critical components from Ruby to Scala, was at a scale unfathomable to most projects. If anything, that a message queue written in pure Ruby backed a site that size at all is an achievement.
2) Twitter made their switch back back in the days of Ruby 1.8; i.e. back when Ruby was executed by directly traversing the AST, without native threading, and with only a primitive GC implementation. Ruby 1.9 and 2.0 are far more sophisticated and far better performing.
3) Not only has the original Matz implementation of Ruby improved dramatically, there are several other first class implementations of the language that have different performance and operational characteristics. Using JRuby, for example, buys you a highly advanced VM, a variety of GC options, and the ability to seamlessly integrate Java classes for performance (or integration).
4) There are quite a few major sites run on Ruby. For example, GitHub.
Now is it a perfect language? Hardly. Is it suitable for every project? Of course not. But it doesn't deserve to be dismissed out of hand.
How is $storage in perl any less ambiguous? It might be a simple scalar, a scalar reference, an array reference, a hash reference, or an object of indeterminate type. At least in Ruby you know - for certain - that it is a class instance variable.
Java may have made minimal impact on the desktop, but how many Android phones are out there? How many "dumb phones" running JME*? How many set top boxes running OCAP? How many Blu-ray players running BD-J? How many devices (ATMs, printers, cash registers, kiosks) running Java? How many web applications backed by server-side Java?
Java runs on billions of devices, all the way from embedded devices too massive distributed applications. It's fairly safe to say that everyone in the developed world unwittingly relies on Java applications every day of their lives.
That isn't to say there aren't disadvantages to the platform, or that Oracle has been an exemplary steward of the technology, but to argue that it has not been a dramatic success and a transformative force in the industry is just ignorant.
Slashdot Mirror
1) How many of the people who reflexivly project hate at Java have solid, defensible reasons for doing so? Do you think people form opinions based on evidence and thoughtful consideration?
2) Examples, please.
3) Well C is not a long-lived programming language, then. COBOL, FORTRAN and LISP have been going strong since the 1950's!
3.1) Plenty. Most of the applications on my phone runs Java (Android). My cable box (OCAP). My Blu-ray player (BD-J). Several of the applications I administer or use professionally (IntelliJ, Confluence, Jira, Zimbra, JMeter).
Just because you don't see it doesn't mean you don't use it. And even if you don't use doesn't mean other people don't use it. Java is an incredible pervasive language in the embedded and server space.
3.2) And?
3.3) With a twelve year history, multiple implementations and no hint of a successor, this strikes me as needless fear-mongering. Microsoft is proprietary, not fickle. Most of their standards have excellent longevity and they have far longer support cycles for their products than most of the Unix world.
As for performance, you're full of it. :)
TRON.
I, for one, welcome our new MCP overlord.
I suspect it is about establishing precedent and combating the idea that EV owners are entitled to "free" power, not about recovering costs in this specific incident.
That's because Democrats wanted people to believe default was a credible threat. It wasn't. No one who knew what they were talking about thought it was, hence the stability of interest rates.
Really, it's the sort of thing that you don't realize you want until you have it, then you are annoyed if you don't have it.
Even under the best conditions, half a second is a bit of an exaggeration, and you're often not under the best of conditions. The cable slipped behind the desk, or you realize that you forgot to plug your phone in after you turned off the light or took off your glasses, or you're driving and don't want to be the asshole who causes an accident because you were fiddling with a cablet instead of keeping your eyes on the road, or your cable is getting dodgy and you keep forgetting to order a new one.
And then there is the annoyance of having your phone go dead because you were too tired to bother flipping the light back on, or never did hit a red light long enough to plug your phone in, or you jostled your phone and it stopped charging.
Is it a need? Certainly not. But it is damned handy.
Big businesses get discounts because they sort their mail, prepare them for automated processing and sometimes transport them to postal facilities. In other words, they pay less because they cost less to service. You could pay the same price if you printed your mail with IMpb barcodes, pre-sorted it and sent it in batches of five hundred or more.
Commecial first class and bulk mail services provide the volume necessary to keep postal rates relatively low and routes open. Increase their rates, volume will drop, and everybody's rates go up.
The raw number isn't very meaningful. There are over 250 million passenger vehicles registered in the United States. That comes to one fire per 1641 vehicles. There are something like 12 thousand Teslas, but also only three fires, which comes to one fire per 4000 vehicles.
So with those numbers, they still seem better off than combustion driven vehicles. However, fleet size isn't the only variable. The majority of Teslas are less than a year old, while overall only 13.5% of passenger vehicles are under two years old, and 60.6% are at least seven years old. Teslas are also a luxury vehicle with a base price of $69,900 (before tax credits) and allegedly stringent manufacturing standards, compared to the overall market with an average new sale price of $24,764 and varying quality. And finally, maintenance needs to be taken into consideration. Not only are Teslas newer and almost certainly under warranty, the average owner is relatively wealthy and presumably can afford proper maintenance.
But even that is not necessarily the right set of stats to look at. The three Tesla fires occurred after an accident, so it may be more informative to look at the liklihood of accidents to result in fire. According to the NFPA, only 4% of automobile fires were the result of a collision or roll-over; most were caused by a mechanical or electrical failure. There were about 5.4 million accidents in 2010, so at 4% of 152300 accident it comes out to approximately one fire per 900 accidents. There are no hard statistics on precisely how many accidents there are for Tesla specifically, but it would have to be a significantly high number to have a lower fire to accident ratio. And that isn't even taking into account the relative severity of accidents into account.
Of course the real problem is that there simply isn't enough data available. Three times in as many months could be a fluke. Or it could be a pattern.
There is a reason why other distributions - even ones that had switched to Upstart - adopted systemd.
There is a reason why other distributions - and toolkit developers - opted against supporting Mir.
And it has nothing to do with the tea party.
"we received a ton of (mostly worthless to us as a business but very expensive) consulting help"
Your tax dollars hard at work, ladies and gentlemen.
Apple managed to roll out iOS 7 - which was in development for only a year - successfully to over 140 million devices in a single day. Rapidly address bug aside, it is nowhere near the scale of mess that the HCX rollouts have been.
The launch of the online component of GTA V has certainly been rough, but we are again talking about a drastically different scale of problem. Online gaming is an extended, high bandwidth, latency sensitive distributed interaction and relies heavily on third party infrastructure (namely the Playstation Network and XBox Live). Going to a website to create a login or submit an application should be short, consume little bandwidth, be tolerant of far longer latencies, and involve only two-party transactions.
We are also talking a drastically different scale of budget. The budget for GTA V is mind-boggling for video games (~$265 million for development and marketing) but it pales in comparison to the mountain of money being poured into the health care exchanges. Kaiser reported that $3.6 BILLION had been awarded in grants to the states to set up the exchanges, and that number doesn't include state funding. California alone spend $313 million on their exchange. It is difficult to know just how much has been spent on the federal exchange, since not all spending is called out explicitly, but the HHS has awarded well over $600 million to GCI alone (the $50-something million number was just one component).
And in the end, the exchanges did not experience a particularly high level of traffic. Though not all states have released numbers, the ones we have are fairly modest. California, the most populous state in the union, reported a scant 514,000 unique visitors on Tuesday. That is NOTHING.
I
Horse pucky. There is no reason to assume there would be one (and only one) peak load. If anything, one would expect annual spikes during the enrollment periods. And the enrollment projections have all seen mounting demand year over year. So this is an invented dichotomy to begin with.
But even if one accepted that as a design precept, there is no need to simply fall down under peak. You make your UX components asynchronous and dump requests into a work queue. Return control to the user, give them a link to check on their status. Not only are these techniques well understood, they are neither difficult to implement nor particularly expensive. You can easily push a million messages PER MINUTE on a single box using free software like ActiveMQ.
And the idea that performance is the determining factor between SQL and no-SQL betrays fatuous. There have been high performance SQL deployments far longer than the current crop of "no-SQL" databases. For transactional workloads with stable data models, you're likely to take a performance hit trying to shoe-horn into the flavor of the week data store.
You also have to consider the repercussion of a malfunctioning system. Not only are you going to generate MORE traffic by being improperly sized for peak load (due to people reloading or making multiple visits), they also have manned phone banks to handle technical difficulties. Servers are cheap. People are not.
Long story short, this has been an epic fail of the sort that would be inexcusable in the private sector. And unfortunately for the tax payer, it isn't a matter of being underfunded. The Department of Health and Human Services reportedly paid $55.7 million to Canadian (!) tech firm CGI Federal for this steaming pile, and may pay them upwards of $40 million more during the tenure of the contract.
Unity, first commit:
Committer: Neil Jagdish Patel
Date: 2009-10-15 10:40:35 UTC
Revision ID: neil.patel@canonical.com-20091015104035-ijthyaoq3rwqu8r7
[build] Initial commit
Gnome-shell, first commit:
tag name 2.27.0 (37b3bb8ab0012a3ba39e775d78772c652eacf804)
tag date 2009-08-10 22:37:47 (GMT)
tagged by Owen W. Taylor
And early development was done in SVN rather than git, so the true start date is much earlier. The first mock-ups appeared in April of that year:
https://wiki.gnome.org/action/info/GnomeShell/Design/Iterations/AppBrowsingAlternative?action=info
The first public demo was at the Gran Canaria Desktop Summit:
http://blogs.gnome.org/marina/2009/07/05/gcds-and-the-gnome-shell-sneak-peak/
So relatively close, but Gnome-Shell was definitely first.
With Mir and Wayland, it isn't even close. The first commit to Mir was in Feburary of this year. Wayland hit 1.0 in October 2012.
Pupil to teacher ratios have been declining for decades. Public schools are already down to a student/teacher ratio of 15.4, with a median class size of 20.0 for public elementary schools (according to the National Center for Educational Statistics).
Pay for teachers has more than outpaced inflation since the 1980's, rising from an inflation adjusted ~ $44k to ~ $54k. Once you factor in benefits, extended summer vacations (or additional income earned teaching summer sessions), pensions and the potential for tenure, the overall compensation picture is hardly unfair or unattractive. And the UNESCO statistics show that starting salaries are actually relatively competitive, internationally speaking; behind Switzerland, Germany,Demark and the Netherlands, but ahead of Australia, Spain, Norway, Ireland, Austria, Iceland, France, Greece, Italy, Portugal and Sweeden.
Money is good. Some schools are legitimately underfunded. However, it is not a panacea, and there is little evidence that too little money is pumped into public education. Consider this; the best funded school district in my area, on a per pupil basis, currently spends over $27,000 per student and achieves the absolute worst results, as measured by performance on standardized tests, graduation rates and college attendance. Other districts excel with half the funding. Parochial schools outperform with less than a QUARTER the funding. And nationally, home schooled children consistently out-perform their peers, in spite of per-pupil spending that is often measured in the hundreds, rather than the thousands.
You claim that the judge eliminated self-defense, not because it wasn't germane, but because of an alleged - but not established - sequence of events. That is simply fatuous, and demonstratably false. The jury was clearly instructed that George Zimmerman's claim of self-defense was the central consideration of the case:
"An issue in this case is whether George Zimmerman acted in self-defense. It is a defense to the crime of Second Degree Murder, and the lesser included offense of Manslaughter, if the death of Trayvon Martin resulted from the justifiable use of deadly force.
If in your consideration of the issue of self-defense you have a reasonable doubt on the question of whether George Zimmerman was justified in the use of deadly force, you should find George Zimmerman not guilty.
However, if from the evidence you are convinced beyond a reasonable doubt that George Zimmerman was not justified in the use of deadly force, you should find him guilty if all the elements of the charge have been proved."
http://www.scribd.com/doc/153354467/George-Zimmerman-Trial-Final-Jury-Instructions
That is not the standard of law in Florida, nor is it that pertinent. He wasn't escaping from someone walking down the street, unaware of him; he was escaping someone who punched him in the face, bashed his head into concrete and was straddling him.
(The pertinent section of the jury instructions: "If George Zimmerman was not engaged in an unlawful activity and was attacked in any place where he had a right to be, he had no duty to retreat and had the right to stand his ground and meet force with force, including deadly force if he reasonably believed that it was necessary to do so to prevent death or great bodily harm to himself or another or to prevent the commission of a forcible felony.")
Thankfully the jury had a better grasp of the standard of self-defense than you do.
Um, no. The judge doesn't get to make that determination. The jury does. After reviewing the evidence and hearing testimony.
[mberg@mberg ~]$ ldd /usr/bin/gnome-shell | wc -l /usr/bin/gnucash | wc -l /usr/bin/konsole | wc -l /opt/google/chrome/chrome | wc -l
100
[mberg@mberg bin]$ ldd
123
[mberg@mberg bin]$ ldd
92
[mberg@mberg ~]$ ldd
70
This is roughly as stupid as when people claimed Florida banned sex a couple years ago. Read the bill. It clearly refers to devices that perform gaming function on activation; not general purpose devices.
Common sense, people. What is more likely? A hundred and forty legislators voting to ban computers, or a bogus lawsuit with a silly premise?
While systemd takes a drastically different approach to initialization than upstart, the similarities between mir and wayland far outweigh their differences, and most of the stated reasons for not adopting wayland seem based on misunderstanding rather than significant issues with the architecture.
Perhaps mir will be a winning move for Canonical. Perhaps it will even be adopted outside Ubuntu. I wouldn't bet on it.
Return works fine; trying to return from something that isn't a method doesn't. The most striking thing about this question is that the submitter never explains what they are doing or why they insist on doing it in a non-idiomatic way. "I'd really prefer to be able to use 'return' in this scenario" is not a compelling argument. What reason did he have for trying to work around a reasonable, well established language paradigm?
As far as people abandoning Ruby, I'm not sure it is more or less common than other languages. There have been a couple high profile "abandonments", like Twitter, which seem to get paraded out as proof that the language does not scale. There are a number of countervailing points, however; specifically:
1) Twitter, even back when they first started their migration of critical components from Ruby to Scala, was at a scale unfathomable to most projects. If anything, that a message queue written in pure Ruby backed a site that size at all is an achievement.
2) Twitter made their switch back back in the days of Ruby 1.8; i.e. back when Ruby was executed by directly traversing the AST, without native threading, and with only a primitive GC implementation. Ruby 1.9 and 2.0 are far more sophisticated and far better performing.
3) Not only has the original Matz implementation of Ruby improved dramatically, there are several other first class implementations of the language that have different performance and operational characteristics. Using JRuby, for example, buys you a highly advanced VM, a variety of GC options, and the ability to seamlessly integrate Java classes for performance (or integration).
4) There are quite a few major sites run on Ruby. For example, GitHub.
Now is it a perfect language? Hardly. Is it suitable for every project? Of course not. But it doesn't deserve to be dismissed out of hand.
How is $storage in perl any less ambiguous? It might be a simple scalar, a scalar reference, an array reference, a hash reference, or an object of indeterminate type. At least in Ruby you know - for certain - that it is a class instance variable.
I suddenly want to re-purpose a classic Easy Bake oven as a computer case.
Java may have made minimal impact on the desktop, but how many Android phones are out there? How many "dumb phones" running JME*? How many set top boxes running OCAP? How many Blu-ray players running BD-J? How many devices (ATMs, printers, cash registers, kiosks) running Java? How many web applications backed by server-side Java?
Java runs on billions of devices, all the way from embedded devices too massive distributed applications. It's fairly safe to say that everyone in the developed world unwittingly relies on Java applications every day of their lives.
That isn't to say there aren't disadvantages to the platform, or that Oracle has been an exemplary steward of the technology, but to argue that it has not been a dramatic success and a transformative force in the industry is just ignorant.
* Quite a few: http://www.netmarketshare.com/operating-system-market-share.aspx?qprid=9&qpcustomb=1
Never heard of autism? Down syndrome? Ceberal Palsey? Rett syndrome? Martin Bell syndrome?