GWT is pretty good (as a Java solution). Yeah, once your project (more than a single person for an extended development period) gets big then static compile time checking can save you from a lot of trouble.
Rails is pretty good if all you do is web->DB form stuff. Once you start integrating with devices and complex webservices then Rails is not as good a fit as something like Java.
That's a fairly broad statement, "GWT doesn't hold up well". I'm curious to see why you think this is the case - my (extensive) experience has been the opposite - GWT scales well and you can do stuff with it that is painful with other technologies (eg. works across just about every browser - even with braindead IE6's limitations; you can also development very modular and re-usable components with GWT). So yeah, I'm interested to hear what general-purpose solution you think is better than GWT and why.
GWT is great if you want interactivity. Also fits in well since you are a Java programmer. I've also recently discovered vaadin (http://vaadin.com) which extends GWT and provides much nicer themes than the default. Best of all, all the tools you need for GWT/vaadin development are free (both zero cost and source is available); Java, Tomcat, Eclipse or Netbeans, GWT + GWT plugin (for debugging in your browser), vaadin and the all-important community documentation/forums (that is, you don't have to pay for a subscription like MSDN). Best of all vaadin/GWT handles almost all of the browser variation for you (although there are differences in CSS for advanced styling).
Since you are already a Java programmer I would say you'd be mad to go past GWT (which you also develop for, and can debug in Java). Other solutions are ok for static page-oriented viewing but for a complex dynamic site it is hard for a competent Java programmer to be more productive than with GWT (except for the smaller sites with not a lot going on). How do I know, well I use GWT daily - including building complex sites for managing medical information for our national-level health ministry.
Well, one of the benefits of languages like Java and Visual Basic are that people who don't have a passion for development can still be productive.
The other thing is that Java is somewhat like Linux - you just don't see it much on the desktop - but it actually drives a very significant amount of enterprise heavy lifting and the Internet (the old saying goes that you can't send an email to anyone outside your company without at least one Linux mail server touching it).
Since the enterprises aren't trying to sell you the software that gives them an edge in their business then you don't hear much about what systems they actually use. This is why many users only consider software other than shrink-wrapped software for the desktop. This is only a small part of development out there (although significant in terms of sales revenue to a few companies). But if you look at something like the Tiobe programming language index you will see that a huge amount of development is done in Java (around 5x what is done with C#.NET for example), although through the recession this has taken a hit as large corporate development is deferred (although it will almost certainly pick up again).
The real beauty of Java is its stability and simplicity - which doesn't make for exciting press, but the JVMs these days are very very fast, and good developers prize simplicity above nearly everything else (which also doesn't get much press, certainly not from developers with just a few years under their belts who pride themselves on obscure languages and constructs, instead of just getting the damn work done in a strategically sensible way [that is portable, since software that will be critical infrastructure for 20 years will have to be run on the platforms of the future]).
Thanks for giving me a chance to elaborate. I hope that my observations help to explain your observations a little.
It gets even worse. If this is indeed to be proven bullshit the loons around the world will still repeat it as if it was truth. The number of web pages repeating the fallacy completely dominate the page with the retraction. Hence it becomes a 'truth' that 'everyone knows' and modifies people's behaviour (eg. the loony fringe).
Here is a case study to prove my point. After years of sustained rocket attacks from Gaza on Southern Israel the Israelis launched Operation Cast Lead to stop the menace. The Israeli Army went in and stopped the attacks (for a while). The Palestinians shrieked that the Israelis had hit a school with artillery (ignoring the fact that their own fighters had targeted civilian infrastructure for *years* with the deliberate aim of killing Israeli civilians; and the Palestinians placed military installations in civilians areas so if hit they could claim propaganda victories). The UN repeated this claim. Then this claim was placed in Richard Goldstone's report claiming that the Israelis had committed war crimes. However, it turned out that this was all bunk (see http://www.washingtonpost.com/blogs/right-turn/post/richard-goldstone-continues-his-recantation/2011/11/01/gIQApEKecM_blog.html and references therein). There were three real instances where Israeli soldiers broke their military law and these were investigated, prosecuted and dealt with (stealing an ATM bank card; using a boy to check for bombs; and I can't recall the third instance). Little mention was made of the Palestinian war crimes (since the Gaza locals went, "Huh?" when pressed for information) - that is: aiming rockets at Israeli civilians; locating military installations amongst civilian housing; storing rockets in schools etc.
My point here is the following: Richard Goldstone's report contained lies (he was played by the Hamas propaganda organs) that Goldstone himself has subsequently recanted as false. Reasonable people who only know about the lies form opinions that the Israelis are bad and war criminals based on this information. The reality is that the Palestinians were conducting the war crimes and the Israelis took great pains to fight as ethically as they could (investigating and prosecuting their own soldiers), but this information is not known my most ordinary folks. If ordinary and reasonable people knew this then they wouldn't condemn Israel on this count (settlements, I think we can all agree, are bad - but that is another matter).
This is what makes putting out sensationalist lies so dangerous and why this guy 'reporting' hearsay to promote his book is so bad.
Talofa! Actually the 'look' of an app is not nearly as important as the 'feel' of it. Unfortunately most Swing programmers are like most programmers in general, not very good and with a poor understanding of their tools. They would block the 'Event Dispatch Thread' and cause their program to run slowly. Users would whinge about this. If you know what you are doing then you don't do this and your program runs as smooth as silk. Also Swing is so powerful compared to most other toolkits that you have enormous flexibility to make the interface work in a really intelligent way, if you are prepared to put in some effort. With regard to the 'look' I also suggest you look up not just 'rich clients' but 'filthy rich clients', which is a bunch of techniques that make Swing apps look incredibly good. Even with no effort just using the Nimbus Look & Feel produces a very nice looking application. Most modern Windows application don't look like the default controls anyway, so this is a bit of a red herring argument, and with Swing's Nimbus theme my clients often comment how good looking my apps are (even though I didn't have to do anything extra to make them look nice).
With regard to Java's performance. We'll Java is faster than C++ in some microbenchmarks and slower in others. The reality is that modern JVMs (past 1.6.0_u10) are very fast, using hardware (Direct X or OpenGL, depending on the platform) for all Java2D rendering. The only thing you need to do to ensure your Java is fast these days is run the excellent profiler built into the SDK (JVisualVM) and check you haven't done things such as used slow Strings when a StringBuffer or StringBuilder would be more appropriate (I've had 100x speedups when replacing XML String concatenation in loops with StringBuffer instead). With regard to gaming, well I'm currently working on a modern jet combat simulator and I find Java not only fast enough but the ability to easily manage resources in many threads means that the overall code run is faster than C++ (where multi-threaded resource management is prone to deadlocks and leaks - it takes an enormous effort to sort it out, whereas Java has explicit language and library support for these, eg. java.util.concurrent is awesome). Furthermore, for games the real speedup is moving code from the CPU to the GPU. At the moment I find my Java program is rendering excellently using Vertex Buffer Objects (VBOs) controlled through JoGL. Then I use libraries like JInput (joystick input) and JOAL (over OpenAL, for sound). The resulting code works on my MacBook Pro (so I can code on a nice leather lazyboy), on my Windows gaming rig, and with the server-side component on my Linux server (which means I can deploy to cheap virtual machines around the world, so everyone gets low latency for 'gunzo' close in cannon fights). In short, WORA works for me (thanks in part to intelligent design of JoGL, JInput, JOAL etc) - hence I say WORA is not a myth. With Java you can get excellent performance, fast development time and deliver your product to the platform that your *users* choose to run on (not twisting their arm to run a platform they don't wanna). How good is that? [hint: a lot better and lot less effort than using crufty old C++ or platform-inhibited C#]
Actually "Write Once, Run Anywhere" works - I know, back in the day I would write Java on a 16-bit Windows and run it on a 64-bit Irix machine. Stuff like that has continued for me throughout the years (these days I write on Mac OS X 64-bit and deploy to Linux, Windows 7 32 and 64), and I've never had a problem so far as long as I've only used the proper APIs and not implementation-specific class.
Hence, my own experience over 17 years with Java has been that WORA actually works. I'd be interested in hearing your experience where you used a standard Java library and it didn't work. Otherwise, you are repeating incorrect hearsay.
> Maybe so, but how could a valid utility patent on a.NET library routine not extend to a functionally equivalent routine on Java or any other platform?
Because the JDK license includes a patent grant for a compatible implementation (or at least, it used to). Note that Dalvik is not compatible with the JDK, but then Google do not claim it is - which is why Oracle felt they could do Google for copyright infringement.
From Miguel: "These days we no longer believe that Silverlight is a suitable platform for write-once-run-anywhere technology, there are just too many limitations for it to be useful"
Who was he trying to kid (besides himself) ?.NET was never designed to be "write once, run anywhere" and never will be (since Mono will never implement all the libraries of the Microsoft implementation). If you want "write once, run anywhere" then just stick to Java, which is designed for this purpose - and yes, it means it can't use every feature of every platform, eg. Windows, but at least your software will run. everywhere (which gives you a better return-on-investment since you can sell to Windows *and* Linux *and* Mac, etc).
There is no patent cloud over Java. This was proven less than two weeks ago. Plus, you are explicitly allowed to create compatible versions of the JVM (but cannot use the trademark Java unless you pay for the Test Compatibility Kit).
Note that the patent cloud over.NET is not for the C# language itself - that is indeed open. The.NET libraries are proprietary and protected by patents.
In short, if you are worried about creating a Java or.NET implementation then Java has a *much* safer legal foundation for doing so (just say you are compatible with OpenJDK:) ). For users this is all irrelevant, users are safe usig Java or C# (but I still prefer Java, it will be around long after Microsoft have abandoned.NET).
Well, it wasn't like the community didn't try and warn the Mono dudes. Microsoft does this sooner or later with all their technology (lose focus on something to try and sell 'teh new shiny' instead).
Israel would not nuke Iran unless Israel's existence was threatened. Even if Iran used a chemical WMD Israel would likely respond with (a lot of) conventional munitions. Really the Israelis have nukes so that they won't have to use them (crazy, eh? but it is the same with most of the other nuke armed powers).
The Americans have met their objectives though: destroyed terrorist training camps; destroyed Al Qaeda (which is just a meme now); destroyed Bin Laden; installed a Western-friendly government; destroyed drugs in most of the government-controlled areas (in the Taliban areas they are still grown, as the drug money is used to support the Taliban while sticking it to the West). Only thing not done is wipe out the Taliban, which will never happen as long as Pakistan supports them (hell, the Taliban are a Pakistani creation - the Taliban are not the same as the jihadis that fought the Soviets, like the Northern Alliance).
In theory the Afghan government could now beat the Afghan Taliban if NATO withdrew and Pakistan kept its mitts out (it won't). With current objectives completed it is a good time to go before new objectives get invented.
You do know that often the 'civilians' buried in drone strikes are the associates of terrorists in the compounds, or just terrorists who aren't carrying guns at the time, yeah? There have also been cases where donkeys got killed in a drone strike in Afghanistan and the villagers buried it and claimed the compensation for a 'civilian' killed. The US drones don't just fire willy nilly on villages for the fsck of it. They actually have Special Forces to observe and target terrorists. These guys do hold back on strikes to try prevent civilian casualties but sometimes the target is so time-sensitive (hard to find you have to take the shot) or surround themselves with civilians that civilian casualties result. The Rules of Engagement of the US and NATO prevent deliberately targeting civilians.For a balanced view you need to contrast this with the policy of both the Afghan and Pakistani Taliban to not only kill civilians (so called 'involuntary martyrdom') but to surround themselves with civilians to exploit the NATO Rules of Engagement (that stipulate civilian casualties are to be avoided). Who is the bad guy?
Yes, war is hell - particularly if you are an innocent in the region. However if you are going to start a war you better fscking well finish it with victory. Fortunately the US has actually achieved its aims in Afghanistan. Terrorist camps destroyed: check; Al Qaeda crippled: check; Western friendly (although corrupt) government installed: check; Bin Laden got: check; poppy fields destroyed in Afghan government controlled areas: check. What the West now relies on is the Afghan government to manage their own country (and everyone whinges how the locals should have sovereignty rather than NATO, so a tick here too), but everyone also knows that this is a shaky proposition since the Taliban (funded by the drug lords in Helmand who need instability to continue their trade, and the Pakistani ISI) is a jobs program for those who could not get work any other way.
Pakistan is a great source of trouble you are right. It is better to have some influence over them than none. Cutting off military aid is daft. Just because Pakistan is not perfect doesn't mean they are not sometimes useful (at least as much as they are sometimes troublesome). Fortunately the US government understands this better than its poorly informed populace (poorly informed by choice, they care more about the Kardasians than what is going on elsewhere in the World - even if it could bite them in the ass).
> Like RMS, I'm getting old, and travel a lot to do talks. If I fall ill or get hit by a car, I hope you turkeys never find out.
Lol. Very funny. I hope that when you do get ill (and it will happen to all of us at one time or another) it is a long time from now. Keep up the great work Bruce, and don't let the gobbling of loco turkeys ever discourage you! I must be getting old too, while it is a very slippery slope with paedophilia what consitutes a child depends on where on the globe you are (eg. IIRC, places in the US where a consenting 17 year old and 18 year have the legal threat of statutory rape is madness), so Stallman makes valid points IMHO - people should think about what he's saying rather than always think he means the worse case (acts with children where there is more than a few years age difference).
> Are new enterprise apps still being written as Java applets,
Not usually. There is a mostly a move to webapps instead (for many of our clients we do stuff in Google Web Toolkit, which you program in Java). We do have some existing applets that are still worked on - mostly because we have a niche that displays road video surveys and engineering features and existing video players simply don't cut it (the players and even video formats are designed for playing movies in a forward direction, not the complexities of single stepping of video forward and backwards).
> or is the lifespan for enterprise apps just longer than consumer apps?
Yes. The enterprise life cycle is usually much longer than the consumer cycle. Consumers think they have a big home network if they have more than a couple of machines on them, and updating and patching them manually is possible and tolerable. Enterprises usually have a lot of customized business stuff in house that makes economic sense to wring the maximum amount of use of (given the high capital investment in developing them). In big enterprises a lot of this stuff is Java (applets or applications) or Java-based webapps or webservices - but consumers never hear or see this - mostly because the enterprises want to keep their 'secret sauce' secret (because it gives them efficiencies and competitive advantage over rivals) or simply have no need to publicize their custom software (publicizing requires effort and costs time/money, and if they aren't selling it to you then there is no point telling you about it). This is why Java is still massively used according to the Tiobe Index (http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html) yet it is invisible to most ordinary folks who see the entire scope of computing as their little Windows desktop. The fortunes of Java have dipped in recent years thanks to enterprises trying to save money by deferring development projects. As the economy improves I would expect the enterprise use of Java to increase again.
I'm using JInput in one of my own projects (a modern jet air combat simulator, works great and very fast with Java/JOGL/JOAL/JInput [yes, Java is more than fast enough for games if you know how to utilize the GPU!]). Turns out that JInput *is portable* among Windows, Linux and Mac (because that is where I'm using it). It isn't truly portable (you'd have to do porting work for Solaris, BSD etc but who uses those anyway;) ). More importantly (and more relevant to this discussion) is that JInput doesn't expose you to dependencies that break with minor revisions of the JVM, which is what the IT admin guys fear - at least to no greater extent than patches to the O/S breaks the USB access layer (which does happen with big O/S revisions, but then all programs break, not just the JInput-using ones).
In this case the problem is that not all Java developers know how to build portable applications, or they do, but they sometimes feel they need to use non-portable libraries/imports when under pressure by management to "just get it done". Then I've also seen code where the developers were clueless with regard to file-names, assuming that every system the software would be run on was case-insenstive like Windows (when I fixed this, the software in question worked perfectly on Linux).
My point here is that there are bad Java developers just as much as there are bad developers in any other language. The software these people make can break with any update (although if they knew what they were doing it wouldn't). Unfortunately the poor IT department gets hammered after an update and they blame Java - when in fact it is the shitty developers who do not avoid non-portable Java libraries (eg. internal classes that are even documented that they should not be used for portable applications; where portability means between different JVM implementations and versions, as well as different operating-systems).
It simply isn't Java's fault that some developers are incompetent - although not everyone knows enough to make this call. Consider if a product wasn't written in Java, say if it was written in C++ or C#.NET then portability is a non-issue, there is no real attempt at true portability across operating systems, platform implementations [MS.NET vs Mono libraries]; or library versions [try replacing your libc or msvcrt.ddl with a different major version and see how much software still works]. That product would very often break if your environment (C library, O/S version) changed, yet people understand this. Most often Java applications don't break when you make major changes to their environment (if they are written properly), they can break 'for no good reason' when you update the JVM and have been badly written. Expect better from your Java developers/Java application suppliers!
Some of this is the typical 'use cases' that differ between a home user and an enterprise user.
A home user mostly passively gets stuff over the web. When they do interact with the web it is in very, very simple ways that are easily implemented with simple page-oriented forms or even some AJAX scattered around for a more dynamic experience (eg. using the excellent Google Web Toolkit, which you program in *gasp*, Java!).
Some enterprise users need a lot more complicated interfaces than that of a simple (or even AJAX) web page. This is because many of these users are actually *creating* things. Mostly a web interface is good enough but for specialized applications the web is woefully short (yes, even HTML5). Examples where the web doesn't cut it in the enterprise space: UML modelling (VisualParadigm is a an excellent Java product); Computer Aided Design/Drafting; anything graphically intensive; anything with high bandwidth requirements for feedback.
So, Java makes perfect sense on the desktop, especially in enterprises that have varied operating systems (eg. design department uses Macs; logistics with SAP on Unix; banks on UNIX; plebs and management on Windows etc) but need to have common applications. Java makes sense at home for products that must support Macs and Windows. Java rarely makes sense for business that are targeting the Windows-only market, which are the budget home users (except of course where Java is running reliably and where it is hidden away from ordinary folks; your DVD/BluRay player; inside your car; your Android phone; running your microwave; etc).
Slashdot Mirror
GWT is pretty good (as a Java solution). Yeah, once your project (more than a single person for an extended development period) gets big then static compile time checking can save you from a lot of trouble.
Rails is pretty good if all you do is web->DB form stuff. Once you start integrating with devices and complex webservices then Rails is not as good a fit as something like Java.
JQuery, no thanks. GWT does a much better job of insulating you from browser craziness IMHO.
That's a fairly broad statement, "GWT doesn't hold up well". I'm curious to see why you think this is the case - my (extensive) experience has been the opposite - GWT scales well and you can do stuff with it that is painful with other technologies (eg. works across just about every browser - even with braindead IE6's limitations; you can also development very modular and re-usable components with GWT). So yeah, I'm interested to hear what general-purpose solution you think is better than GWT and why.
GWT is great if you want interactivity. Also fits in well since you are a Java programmer. I've also recently discovered vaadin (http://vaadin.com) which extends GWT and provides much nicer themes than the default. Best of all, all the tools you need for GWT/vaadin development are free (both zero cost and source is available); Java, Tomcat, Eclipse or Netbeans, GWT + GWT plugin (for debugging in your browser), vaadin and the all-important community documentation/forums (that is, you don't have to pay for a subscription like MSDN). Best of all vaadin/GWT handles almost all of the browser variation for you (although there are differences in CSS for advanced styling).
Since you are already a Java programmer I would say you'd be mad to go past GWT (which you also develop for, and can debug in Java). Other solutions are ok for static page-oriented viewing but for a complex dynamic site it is hard for a competent Java programmer to be more productive than with GWT (except for the smaller sites with not a lot going on). How do I know, well I use GWT daily - including building complex sites for managing medical information for our national-level health ministry.
Well, one of the benefits of languages like Java and Visual Basic are that people who don't have a passion for development can still be productive.
The other thing is that Java is somewhat like Linux - you just don't see it much on the desktop - but it actually drives a very significant amount of enterprise heavy lifting and the Internet (the old saying goes that you can't send an email to anyone outside your company without at least one Linux mail server touching it).
Since the enterprises aren't trying to sell you the software that gives them an edge in their business then you don't hear much about what systems they actually use. This is why many users only consider software other than shrink-wrapped software for the desktop. This is only a small part of development out there (although significant in terms of sales revenue to a few companies). But if you look at something like the Tiobe programming language index you will see that a huge amount of development is done in Java (around 5x what is done with C#.NET for example), although through the recession this has taken a hit as large corporate development is deferred (although it will almost certainly pick up again).
The real beauty of Java is its stability and simplicity - which doesn't make for exciting press, but the JVMs these days are very very fast, and good developers prize simplicity above nearly everything else (which also doesn't get much press, certainly not from developers with just a few years under their belts who pride themselves on obscure languages and constructs, instead of just getting the damn work done in a strategically sensible way [that is portable, since software that will be critical infrastructure for 20 years will have to be run on the platforms of the future]).
Thanks for giving me a chance to elaborate. I hope that my observations help to explain your observations a little.
Actually, Hugo is having trouble shaking his cancer. I'd put my money on that getting him first.
It gets even worse. If this is indeed to be proven bullshit the loons around the world will still repeat it as if it was truth. The number of web pages repeating the fallacy completely dominate the page with the retraction. Hence it becomes a 'truth' that 'everyone knows' and modifies people's behaviour (eg. the loony fringe).
Here is a case study to prove my point. After years of sustained rocket attacks from Gaza on Southern Israel the Israelis launched Operation Cast Lead to stop the menace. The Israeli Army went in and stopped the attacks (for a while). The Palestinians shrieked that the Israelis had hit a school with artillery (ignoring the fact that their own fighters had targeted civilian infrastructure for *years* with the deliberate aim of killing Israeli civilians; and the Palestinians placed military installations in civilians areas so if hit they could claim propaganda victories). The UN repeated this claim. Then this claim was placed in Richard Goldstone's report claiming that the Israelis had committed war crimes. However, it turned out that this was all bunk (see http://www.washingtonpost.com/blogs/right-turn/post/richard-goldstone-continues-his-recantation/2011/11/01/gIQApEKecM_blog.html and references therein). There were three real instances where Israeli soldiers broke their military law and these were investigated, prosecuted and dealt with (stealing an ATM bank card; using a boy to check for bombs; and I can't recall the third instance). Little mention was made of the Palestinian war crimes (since the Gaza locals went, "Huh?" when pressed for information) - that is: aiming rockets at Israeli civilians; locating military installations amongst civilian housing; storing rockets in schools etc.
My point here is the following: Richard Goldstone's report contained lies (he was played by the Hamas propaganda organs) that Goldstone himself has subsequently recanted as false. Reasonable people who only know about the lies form opinions that the Israelis are bad and war criminals based on this information. The reality is that the Palestinians were conducting the war crimes and the Israelis took great pains to fight as ethically as they could (investigating and prosecuting their own soldiers), but this information is not known my most ordinary folks. If ordinary and reasonable people knew this then they wouldn't condemn Israel on this count (settlements, I think we can all agree, are bad - but that is another matter).
This is what makes putting out sensationalist lies so dangerous and why this guy 'reporting' hearsay to promote his book is so bad.
Talofa! Actually the 'look' of an app is not nearly as important as the 'feel' of it. Unfortunately most Swing programmers are like most programmers in general, not very good and with a poor understanding of their tools. They would block the 'Event Dispatch Thread' and cause their program to run slowly. Users would whinge about this. If you know what you are doing then you don't do this and your program runs as smooth as silk. Also Swing is so powerful compared to most other toolkits that you have enormous flexibility to make the interface work in a really intelligent way, if you are prepared to put in some effort. With regard to the 'look' I also suggest you look up not just 'rich clients' but 'filthy rich clients', which is a bunch of techniques that make Swing apps look incredibly good. Even with no effort just using the Nimbus Look & Feel produces a very nice looking application. Most modern Windows application don't look like the default controls anyway, so this is a bit of a red herring argument, and with Swing's Nimbus theme my clients often comment how good looking my apps are (even though I didn't have to do anything extra to make them look nice).
With regard to Java's performance. We'll Java is faster than C++ in some microbenchmarks and slower in others. The reality is that modern JVMs (past 1.6.0_u10) are very fast, using hardware (Direct X or OpenGL, depending on the platform) for all Java2D rendering. The only thing you need to do to ensure your Java is fast these days is run the excellent profiler built into the SDK (JVisualVM) and check you haven't done things such as used slow Strings when a StringBuffer or StringBuilder would be more appropriate (I've had 100x speedups when replacing XML String concatenation in loops with StringBuffer instead). With regard to gaming, well I'm currently working on a modern jet combat simulator and I find Java not only fast enough but the ability to easily manage resources in many threads means that the overall code run is faster than C++ (where multi-threaded resource management is prone to deadlocks and leaks - it takes an enormous effort to sort it out, whereas Java has explicit language and library support for these, eg. java.util.concurrent is awesome). Furthermore, for games the real speedup is moving code from the CPU to the GPU. At the moment I find my Java program is rendering excellently using Vertex Buffer Objects (VBOs) controlled through JoGL. Then I use libraries like JInput (joystick input) and JOAL (over OpenAL, for sound). The resulting code works on my MacBook Pro (so I can code on a nice leather lazyboy), on my Windows gaming rig, and with the server-side component on my Linux server (which means I can deploy to cheap virtual machines around the world, so everyone gets low latency for 'gunzo' close in cannon fights). In short, WORA works for me (thanks in part to intelligent design of JoGL, JInput, JOAL etc) - hence I say WORA is not a myth. With Java you can get excellent performance, fast development time and deliver your product to the platform that your *users* choose to run on (not twisting their arm to run a platform they don't wanna). How good is that? [hint: a lot better and lot less effort than using crufty old C++ or platform-inhibited C#]
You do realise that Stallman has subsequently said Java is ok since it is now GPL-ed, yeah?
Actually "Write Once, Run Anywhere" works - I know, back in the day I would write Java on a 16-bit Windows and run it on a 64-bit Irix machine. Stuff like that has continued for me throughout the years (these days I write on Mac OS X 64-bit and deploy to Linux, Windows 7 32 and 64), and I've never had a problem so far as long as I've only used the proper APIs and not implementation-specific class.
Hence, my own experience over 17 years with Java has been that WORA actually works. I'd be interested in hearing your experience where you used a standard Java library and it didn't work. Otherwise, you are repeating incorrect hearsay.
> Maybe so, but how could a valid utility patent on a .NET library routine not extend to a functionally equivalent routine on Java or any other platform?
Because the JDK license includes a patent grant for a compatible implementation (or at least, it used to). Note that Dalvik is not compatible with the JDK, but then Google do not claim it is - which is why Oracle felt they could do Google for copyright infringement.
From Miguel: "These days we no longer believe that Silverlight is a suitable platform for write-once-run-anywhere technology, there are just too many limitations for it to be useful"
Who was he trying to kid (besides himself) ? .NET was never designed to be "write once, run anywhere" and never will be (since Mono will never implement all the libraries of the Microsoft implementation). If you want "write once, run anywhere" then just stick to Java, which is designed for this purpose - and yes, it means it can't use every feature of every platform, eg. Windows, but at least your software will run. everywhere (which gives you a better return-on-investment since you can sell to Windows *and* Linux *and* Mac, etc).
There is no patent cloud over Java. This was proven less than two weeks ago. Plus, you are explicitly allowed to create compatible versions of the JVM (but cannot use the trademark Java unless you pay for the Test Compatibility Kit).
Note that the patent cloud over .NET is not for the C# language itself - that is indeed open. The .NET libraries are proprietary and protected by patents.
In short, if you are worried about creating a Java or .NET implementation then Java has a *much* safer legal foundation for doing so (just say you are compatible with OpenJDK :) ). For users this is all irrelevant, users are safe usig Java or C# (but I still prefer Java, it will be around long after Microsoft have abandoned .NET).
Well, it wasn't like the community didn't try and warn the Mono dudes. Microsoft does this sooner or later with all their technology (lose focus on something to try and sell 'teh new shiny' instead).
Israel would not nuke Iran unless Israel's existence was threatened. Even if Iran used a chemical WMD Israel would likely respond with (a lot of) conventional munitions. Really the Israelis have nukes so that they won't have to use them (crazy, eh? but it is the same with most of the other nuke armed powers).
Isn't sufficiently advanced technology indistinguishable from magic?
The Americans have met their objectives though: destroyed terrorist training camps; destroyed Al Qaeda (which is just a meme now); destroyed Bin Laden; installed a Western-friendly government; destroyed drugs in most of the government-controlled areas (in the Taliban areas they are still grown, as the drug money is used to support the Taliban while sticking it to the West). Only thing not done is wipe out the Taliban, which will never happen as long as Pakistan supports them (hell, the Taliban are a Pakistani creation - the Taliban are not the same as the jihadis that fought the Soviets, like the Northern Alliance).
In theory the Afghan government could now beat the Afghan Taliban if NATO withdrew and Pakistan kept its mitts out (it won't). With current objectives completed it is a good time to go before new objectives get invented.
You do know that often the 'civilians' buried in drone strikes are the associates of terrorists in the compounds, or just terrorists who aren't carrying guns at the time, yeah? There have also been cases where donkeys got killed in a drone strike in Afghanistan and the villagers buried it and claimed the compensation for a 'civilian' killed. The US drones don't just fire willy nilly on villages for the fsck of it. They actually have Special Forces to observe and target terrorists. These guys do hold back on strikes to try prevent civilian casualties but sometimes the target is so time-sensitive (hard to find you have to take the shot) or surround themselves with civilians that civilian casualties result. The Rules of Engagement of the US and NATO prevent deliberately targeting civilians.For a balanced view you need to contrast this with the policy of both the Afghan and Pakistani Taliban to not only kill civilians (so called 'involuntary martyrdom') but to surround themselves with civilians to exploit the NATO Rules of Engagement (that stipulate civilian casualties are to be avoided). Who is the bad guy?
Yes, war is hell - particularly if you are an innocent in the region. However if you are going to start a war you better fscking well finish it with victory. Fortunately the US has actually achieved its aims in Afghanistan. Terrorist camps destroyed: check; Al Qaeda crippled: check; Western friendly (although corrupt) government installed: check; Bin Laden got: check; poppy fields destroyed in Afghan government controlled areas: check. What the West now relies on is the Afghan government to manage their own country (and everyone whinges how the locals should have sovereignty rather than NATO, so a tick here too), but everyone also knows that this is a shaky proposition since the Taliban (funded by the drug lords in Helmand who need instability to continue their trade, and the Pakistani ISI) is a jobs program for those who could not get work any other way.
Pakistan is a great source of trouble you are right. It is better to have some influence over them than none. Cutting off military aid is daft. Just because Pakistan is not perfect doesn't mean they are not sometimes useful (at least as much as they are sometimes troublesome). Fortunately the US government understands this better than its poorly informed populace (poorly informed by choice, they care more about the Kardasians than what is going on elsewhere in the World - even if it could bite them in the ass).
> Like RMS, I'm getting old, and travel a lot to do talks. If I fall ill or get hit by a car, I hope you turkeys never find out.
Lol. Very funny. I hope that when you do get ill (and it will happen to all of us at one time or another) it is a long time from now. Keep up the great work Bruce, and don't let the gobbling of loco turkeys ever discourage you! I must be getting old too, while it is a very slippery slope with paedophilia what consitutes a child depends on where on the globe you are (eg. IIRC, places in the US where a consenting 17 year old and 18 year have the legal threat of statutory rape is madness), so Stallman makes valid points IMHO - people should think about what he's saying rather than always think he means the worse case (acts with children where there is more than a few years age difference).
> Are new enterprise apps still being written as Java applets,
Not usually. There is a mostly a move to webapps instead (for many of our clients we do stuff in Google Web Toolkit, which you program in Java). We do have some existing applets that are still worked on - mostly because we have a niche that displays road video surveys and engineering features and existing video players simply don't cut it (the players and even video formats are designed for playing movies in a forward direction, not the complexities of single stepping of video forward and backwards).
> or is the lifespan for enterprise apps just longer than consumer apps?
Yes. The enterprise life cycle is usually much longer than the consumer cycle. Consumers think they have a big home network if they have more than a couple of machines on them, and updating and patching them manually is possible and tolerable. Enterprises usually have a lot of customized business stuff in house that makes economic sense to wring the maximum amount of use of (given the high capital investment in developing them). In big enterprises a lot of this stuff is Java (applets or applications) or Java-based webapps or webservices - but consumers never hear or see this - mostly because the enterprises want to keep their 'secret sauce' secret (because it gives them efficiencies and competitive advantage over rivals) or simply have no need to publicize their custom software (publicizing requires effort and costs time/money, and if they aren't selling it to you then there is no point telling you about it). This is why Java is still massively used according to the Tiobe Index (http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html) yet it is invisible to most ordinary folks who see the entire scope of computing as their little Windows desktop. The fortunes of Java have dipped in recent years thanks to enterprises trying to save money by deferring development projects. As the economy improves I would expect the enterprise use of Java to increase again.
... for home users. Yes, you are correct in this with the "home user" proviso.
I'm using JInput in one of my own projects (a modern jet air combat simulator, works great and very fast with Java/JOGL/JOAL/JInput [yes, Java is more than fast enough for games if you know how to utilize the GPU!]). Turns out that JInput *is portable* among Windows, Linux and Mac (because that is where I'm using it). It isn't truly portable (you'd have to do porting work for Solaris, BSD etc but who uses those anyway ;) ). More importantly (and more relevant to this discussion) is that JInput doesn't expose you to dependencies that break with minor revisions of the JVM, which is what the IT admin guys fear - at least to no greater extent than patches to the O/S breaks the USB access layer (which does happen with big O/S revisions, but then all programs break, not just the JInput-using ones).
In this case the problem is that not all Java developers know how to build portable applications, or they do, but they sometimes feel they need to use non-portable libraries/imports when under pressure by management to "just get it done". Then I've also seen code where the developers were clueless with regard to file-names, assuming that every system the software would be run on was case-insenstive like Windows (when I fixed this, the software in question worked perfectly on Linux).
My point here is that there are bad Java developers just as much as there are bad developers in any other language. The software these people make can break with any update (although if they knew what they were doing it wouldn't). Unfortunately the poor IT department gets hammered after an update and they blame Java - when in fact it is the shitty developers who do not avoid non-portable Java libraries (eg. internal classes that are even documented that they should not be used for portable applications; where portability means between different JVM implementations and versions, as well as different operating-systems).
It simply isn't Java's fault that some developers are incompetent - although not everyone knows enough to make this call. Consider if a product wasn't written in Java, say if it was written in C++ or C#.NET then portability is a non-issue, there is no real attempt at true portability across operating systems, platform implementations [MS.NET vs Mono libraries]; or library versions [try replacing your libc or msvcrt.ddl with a different major version and see how much software still works]. That product would very often break if your environment (C library, O/S version) changed, yet people understand this. Most often Java applications don't break when you make major changes to their environment (if they are written properly), they can break 'for no good reason' when you update the JVM and have been badly written. Expect better from your Java developers/Java application suppliers!
Some of this is the typical 'use cases' that differ between a home user and an enterprise user.
A home user mostly passively gets stuff over the web. When they do interact with the web it is in very, very simple ways that are easily implemented with simple page-oriented forms or even some AJAX scattered around for a more dynamic experience (eg. using the excellent Google Web Toolkit, which you program in *gasp*, Java!).
Some enterprise users need a lot more complicated interfaces than that of a simple (or even AJAX) web page. This is because many of these users are actually *creating* things. Mostly a web interface is good enough but for specialized applications the web is woefully short (yes, even HTML5). Examples where the web doesn't cut it in the enterprise space: UML modelling (VisualParadigm is a an excellent Java product); Computer Aided Design/Drafting; anything graphically intensive; anything with high bandwidth requirements for feedback.
So, Java makes perfect sense on the desktop, especially in enterprises that have varied operating systems (eg. design department uses Macs; logistics with SAP on Unix; banks on UNIX; plebs and management on Windows etc) but need to have common applications. Java makes sense at home for products that must support Macs and Windows. Java rarely makes sense for business that are targeting the Windows-only market, which are the budget home users (except of course where Java is running reliably and where it is hidden away from ordinary folks; your DVD/BluRay player; inside your car; your Android phone; running your microwave; etc).