Call me flamebait, but why should Sun endorse PHP?!
There are so many better technologies around. Why should Sun endorse something that can't properly diferentiate between a zero, a null and an empty string and that (AFAIK until the last version) insists in doing shallow copies of objects when you assign them to a new variable.
I know... I know... Many people like PHP. But, if you really like it, I would suggest widening your horizons a little bit and doing some programming (not the web kind) with languages like Ruby or Python just to see what they have to offer.
1) The runtime and libraries are huge because it is a very complete library set. They encompass many things (some of them I wonder if are ever used) that other frameworks (or even OSs) don't offer. It reimplements many things because it does rely on the guest OS for almost nothing, as different OSs tend to do things differently.
2) VM bloat has nothing to do with core library size. The memory footprint you are complaining about is probably due to the optimizing VM itself plus the code that had to be loaded for the Hello Wold application. I suspect most of it is for the VM and very little for the Hello application. The same VM can also run NetBeans with the same VM memory footprint (of course, the rest of NetBeans takes a whole lot of memory)
3) The VM start-up time is only a problem for desktop applications. For server-side stuff (which os where Java excels) the VM is started about, say, once every year. I don't run many Hello World applications when at work, so, if the VM takes 5 seconds to load and 50 to load the application, I don't complain.
Of course, there are other tools (I just love Python and Zope and I am exploring Twisted and TurboGears) that may have a faster start-up times or smaller memory footprints, but Java is a more than good enough "OS within the OS" coupled with a good enough language.
However, this is definitely one of the weakest points in Sun's lifetime and it may scare away potential enterprise level decision makers into going with Java and Solaris
People can always buy a Fujitsu PRIMEPOWER instead. Remember, SPARC is an open architecture and, if OpenSolaris fail to deliver a decent OS, Linux runs on just about anything.
Do we really need an actual example before we can think it is possible?
It is possible. Has it been done? I think not. Will it be done someday? I bet it will.
Remember the time you could fry (sparks included) a CGA monitor by sending it the wrong frequencies just by reprogramming the 6845 CRT controller? Well... I do.
Besides the points already considered in this thread, the malware program could also be made to sabotage all thermal management on your processor, motherboard and hard-disk, effectively frying your computer while you attempt to temove it.
After all, the cost of the license for the server version of Windows, plus the client access licenses is quite respectable and many times surpasses the cost of the hardware itself. That is not what happens with desktop PCs - the license that comes with them is, quite often, neglectable (as in ordering the same PC from Dell without an OS is frequently more expensive). Most big businesses disregard the OS that comes installed and roll out their own Select license. I don't know if they demand refund for the unused licenses explicitly or simply bundle it in their large-scale licenses.
But I still think that being able to run Win32 apps is irrelevant to a given OSs success or failure. What is needed is a killer capability - something that makes that OS so desirable that it outweights the cost of purchasing it and moving away from your current OS.
The PC, even with 327x emulation, provided a very desirable capability - the possibility to buy and run or to quickly build programs for it. Were this capability not there (say, if PC-DOS were not bundled with MS BASIC, or Visicalc didn't exist), the PC could have ended up in a forgotten corner of the IT world.
What a computing platform really needs to survive is a competitive edge - if it is an edge in a niche, it will survive, but if it is an edge in the general market, it may even prevail. Be it more stability, a new capability that is very needed, very low cost computers to run it on or a kind of eye-candy that makes all other computers before it look like CP/M boxes. I don't know how it will begin - I don't know what kind of development is required - but I do know how it will end: it will end in the said platform having its own software patterned more or less to the software of the previous generation, with the added twist of whatever made it so "great" in the first place.
I am sure that being able to mimic that other dominant platform is not the best way to spend development resources. It was not enough to save OS/2. And it was a very nice OS, by the way.;-)
Of course, OS/2, OPENSTEP, BeOS, SCO OpenDesktop failed for various reasons.
For instance, I gave up on OS/2 when, while ordering a computer (an Aptiva 486) from IBM, I learned that having it with OS/2 instead of Windows was more expensive.
Yet, I insist, one of the reasons was not having a decent software library that Windows couldn't run better.
You won't migrate away from Windows to run Office slightly worse than you can do with the Windows that came pre installed (more or less for free) with your computer. The only reason you will migrate OSs is if you need or want badly an application that is only available for the other OS and has no equivalent on your current OS.
I did migrate away from Windows (I run Ubuntu now) when I found a better environment for my Java/Python/Zope development. This way, my computer was a closer reproduction of the servers my applications would run on and was able to run stuff Windows wouldn't. Having a faithful copy of your target platform (or, at least, not a totally different one) is a huge timesaver. Were you something more than a mindless parrot, you would know.
Recently I started using a Windows box again because I had to do.NET development. I would not settle for running Visual Studio over Wine (even if it were possible) because I could not trust the environment to be a faithful reproduction of the target environment. Strange things would break in strange ways and I am not paid to debug my environment. Besides that, Visual Studio is a lot better than MonoDevelop.
Having software that can't run (or adequately run) on other systems can give an OS a competitive edge (PageMaker and PhotoShop on MacOS, Excel on Windows vs. Lotus 123/G on OS/2, CAD/CAE/CAM on various Unixes come to mind). Being able to poorly mimic another OS hardly gives you an edge over it.
And, if OS B can run adequately applications written for OS A, I would not bother develop anything for OS B simply because by developing for OS A (and just testing and making sure it runs on OS B) gives me access to both platforms. That is one (if only one of many) of the reasons OS/2 never got anywhere.
Mostly every second election. That seems to be a good bet for a president to be re-elected. Since they cannot serve three consecutive terms, they may want to limit beligerance on their second term to what is needed for the party to be able to elect the next president.
Interestingly the people at MS research are expecting just that - they are writing Singularity in what is essentially C# with extensions (extensions mostly in the form of formal specification semantics to allow more complete static checking). The upside to doing this is that, when combined with a better ground up approach to security as is being used in Singularity, you get a remarkably robust and secure kernel for an operating system.
Of course, I couldn't resist to say that Singularity sucks.
Better yet. Donate sperm to a sperm-bank. This way you will have a bunch of children at zero cost and tie up resources of someone else who would (from your selfish gene point of view) pollute the gene pool. Your offspring will also have a couple advantages:
- The more descendants you leave (the more your impact on the gene pool), the more likely for them to find compatible blood, bone marrow or organ donors.
- If you and your descendants share a common genetic trait that may cause a disease or other medical condition, making it more common will increase the odds of someone developing a treatment.
Have you ever approached an investor with NDA in hand?
Probably not.
They won't listen to you because the guy next in line behind you (or the guy that came in yesterday) could have had the very same brilliant idea. That guy may seem more viable than you and giving you the ammo you need to come back at them later is just plain stupid. They listen to lots of people with lots of ideas and, quite often, to many people with the same idea.
But, if what you have can be patented and you got a patent on that, they may dump the more promising guy and go for you.
Fortunately, we don't need firearms for this. We can stop using and recomending DRM capable hardware and we can halt software development for it. We must be very vocal in our opposition to this. We may may be few, but I am sure this audience is more influential than the average.
The main difference between BSD and GPL is that GPL obliges anyone using it to play nice with others, to give back what they do (if they choose to pass it along). This is what made GPL the most widely used free software license as it did not allow anyone to grab GPL software and turn it into something proprietary (as opposed to BSD-ish stuff) - it fostered an ecosystem where software could evolve and thrive while BSD-ish sofwtare is mostly a starting point.
The main problem with this approach is with companies like NTP, who never had a product and never, ever, intended to have one. They can't possibly infringe on somebody's patents because they don't really work on anything.
Columbia's crew died because small pieces of foam falling off tanks got to be routine, and eventually after 100 missions a big one fell off and hit probably the single worst place on the whole Orbiter.
It always struck me as slightly amazing that nobody ever carried out even a visual inspection of the orbiter for damage done on lift-off. It took about a hundred missions and a catastrophic loss of crew and vehicle.
After all, the hardest part (getting the thing to orbit) was already done. It would not even require an EVA - a station flyby could suffice (even if some guys that were at Mir could get anxious with someone suggesting close flybys).
Slashdot Mirror
Call me flamebait, but why should Sun endorse PHP?!
There are so many better technologies around. Why should Sun endorse something that can't properly diferentiate between a zero, a null and an empty string and that (AFAIK until the last version) insists in doing shallow copies of objects when you assign them to a new variable.
I know... I know... Many people like PHP. But, if you really like it, I would suggest widening your horizons a little bit and doing some programming (not the web kind) with languages like Ruby or Python just to see what they have to offer.
1) The runtime and libraries are huge because it is a very complete library set. They encompass many things (some of them I wonder if are ever used) that other frameworks (or even OSs) don't offer. It reimplements many things because it does rely on the guest OS for almost nothing, as different OSs tend to do things differently.
2) VM bloat has nothing to do with core library size. The memory footprint you are complaining about is probably due to the optimizing VM itself plus the code that had to be loaded for the Hello Wold application. I suspect most of it is for the VM and very little for the Hello application. The same VM can also run NetBeans with the same VM memory footprint (of course, the rest of NetBeans takes a whole lot of memory)
3) The VM start-up time is only a problem for desktop applications. For server-side stuff (which os where Java excels) the VM is started about, say, once every year. I don't run many Hello World applications when at work, so, if the VM takes 5 seconds to load and 50 to load the application, I don't complain.
Of course, there are other tools (I just love Python and Zope and I am exploring Twisted and TurboGears) that may have a faster start-up times or smaller memory footprints, but Java is a more than good enough "OS within the OS" coupled with a good enough language.
Even if it's not fit for writing poetry.
And John Sculley came from Pepsi.
Of course, I am being ironic.
People can always buy a Fujitsu PRIMEPOWER instead. Remember, SPARC is an open architecture and, if OpenSolaris fail to deliver a decent OS, Linux runs on just about anything.
Do we really need an actual example before we can think it is possible?
It is possible. Has it been done? I think not. Will it be done someday? I bet it will.
Remember the time you could fry (sparks included) a CGA monitor by sending it the wrong frequencies just by reprogramming the 6845 CRT controller? Well... I do.
Besides the points already considered in this thread, the malware program could also be made to sabotage all thermal management on your processor, motherboard and hard-disk, effectively frying your computer while you attempt to temove it.
I am sure that, as bad as Windows can be, it does not qualify as a crime against humanity. Bill is nowhere near as relevant.
Samba is a very good idea.
;-)
After all, the cost of the license for the server version of Windows, plus the client access licenses is quite respectable and many times surpasses the cost of the hardware itself. That is not what happens with desktop PCs - the license that comes with them is, quite often, neglectable (as in ordering the same PC from Dell without an OS is frequently more expensive). Most big businesses disregard the OS that comes installed and roll out their own Select license. I don't know if they demand refund for the unused licenses explicitly or simply bundle it in their large-scale licenses.
But I still think that being able to run Win32 apps is irrelevant to a given OSs success or failure. What is needed is a killer capability - something that makes that OS so desirable that it outweights the cost of purchasing it and moving away from your current OS.
The PC, even with 327x emulation, provided a very desirable capability - the possibility to buy and run or to quickly build programs for it. Were this capability not there (say, if PC-DOS were not bundled with MS BASIC, or Visicalc didn't exist), the PC could have ended up in a forgotten corner of the IT world.
What a computing platform really needs to survive is a competitive edge - if it is an edge in a niche, it will survive, but if it is an edge in the general market, it may even prevail. Be it more stability, a new capability that is very needed, very low cost computers to run it on or a kind of eye-candy that makes all other computers before it look like CP/M boxes. I don't know how it will begin - I don't know what kind of development is required - but I do know how it will end: it will end in the said platform having its own software patterned more or less to the software of the previous generation, with the added twist of whatever made it so "great" in the first place.
I am sure that being able to mimic that other dominant platform is not the best way to spend development resources. It was not enough to save OS/2. And it was a very nice OS, by the way.
Of course, OS/2, OPENSTEP, BeOS, SCO OpenDesktop failed for various reasons.
.NET development. I would not settle for running Visual Studio over Wine (even if it were possible) because I could not trust the environment to be a faithful reproduction of the target environment. Strange things would break in strange ways and I am not paid to debug my environment. Besides that, Visual Studio is a lot better than MonoDevelop.
For instance, I gave up on OS/2 when, while ordering a computer (an Aptiva 486) from IBM, I learned that having it with OS/2 instead of Windows was more expensive.
Yet, I insist, one of the reasons was not having a decent software library that Windows couldn't run better.
You won't migrate away from Windows to run Office slightly worse than you can do with the Windows that came pre installed (more or less for free) with your computer. The only reason you will migrate OSs is if you need or want badly an application that is only available for the other OS and has no equivalent on your current OS.
I did migrate away from Windows (I run Ubuntu now) when I found a better environment for my Java/Python/Zope development. This way, my computer was a closer reproduction of the servers my applications would run on and was able to run stuff Windows wouldn't. Having a faithful copy of your target platform (or, at least, not a totally different one) is a huge timesaver. Were you something more than a mindless parrot, you would know.
Recently I started using a Windows box again because I had to do
Having software that can't run (or adequately run) on other systems can give an OS a competitive edge (PageMaker and PhotoShop on MacOS, Excel on Windows vs. Lotus 123/G on OS/2, CAD/CAE/CAM on various Unixes come to mind). Being able to poorly mimic another OS hardly gives you an edge over it.
And, if OS B can run adequately applications written for OS A, I would not bother develop anything for OS B simply because by developing for OS A (and just testing and making sure it runs on OS B) gives me access to both platforms. That is one (if only one of many) of the reasons OS/2 never got anywhere.
The failure of OS/2 made clear an operating system capable of running software of the dominant OS tends not to develop a native software catalog.
For a long time, even within IBM, OS/2 was being used as a Windows clone, running Windows binaries.
Commenting your sig, it's all right. The guys who took the liberties as payment will never deliver security.
Lets just not forget people are running a three to five year-old office suite atop a five year old operating system.
This stuff was made for sub-gigahertz CPUs with less than half a gigabyte of RAM.
When I have to rank Microsoft's products, I usually put SQL Server on third, right behind the natural keyboards and optical mice.
Mostly every second election. That seems to be a good bet for a president to be re-elected. Since they cannot serve three consecutive terms, they may want to limit beligerance on their second term to what is needed for the party to be able to elect the next president.
Most probably, the same ethics committee that cleared the use of torture in Guantanamo base.
Sadly, I don't expect much ethics from these guys.
And yes, this is sickening.
Of course, I couldn't resist to say that Singularity sucks.
Better yet. Donate sperm to a sperm-bank. This way you will have a bunch of children at zero cost and tie up resources of someone else who would (from your selfish gene point of view) pollute the gene pool. Your offspring will also have a couple advantages:
- The more descendants you leave (the more your impact on the gene pool), the more likely for them to find compatible blood, bone marrow or organ donors.
- If you and your descendants share a common genetic trait that may cause a disease or other medical condition, making it more common will increase the odds of someone developing a treatment.
Plus a couple more.
As long as the client spends less money, that's fine.
Have you ever approached an investor with NDA in hand?
Probably not.
They won't listen to you because the guy next in line behind you (or the guy that came in yesterday) could have had the very same brilliant idea. That guy may seem more viable than you and giving you the ammo you need to come back at them later is just plain stupid. They listen to lots of people with lots of ideas and, quite often, to many people with the same idea.
But, if what you have can be patented and you got a patent on that, they may dump the more promising guy and go for you.
Not to mention that, since the demise of the 5.25" floppy, drive letters were A:, C:, D: and so on, skipping B:.
With new computers, they just start from C:
Try to explain why to a newbie.
We should not run. We should fight back.
Fortunately, we don't need firearms for this. We can stop using and recomending DRM capable hardware and we can halt software development for it. We must be very vocal in our opposition to this. We may may be few, but I am sure this audience is more influential than the average.
Does it mean we will have even-more-universal binaries for 64-bit x86 MacOS X?
The main difference between BSD and GPL is that GPL obliges anyone using it to play nice with others, to give back what they do (if they choose to pass it along). This is what made GPL the most widely used free software license as it did not allow anyone to grab GPL software and turn it into something proprietary (as opposed to BSD-ish stuff) - it fostered an ecosystem where software could evolve and thrive while BSD-ish sofwtare is mostly a starting point.
The main problem with this approach is with companies like NTP, who never had a product and never, ever, intended to have one. They can't possibly infringe on somebody's patents because they don't really work on anything.
It always struck me as slightly amazing that nobody ever carried out even a visual inspection of the orbiter for damage done on lift-off. It took about a hundred missions and a catastrophic loss of crew and vehicle.
After all, the hardest part (getting the thing to orbit) was already done. It would not even require an EVA - a station flyby could suffice (even if some guys that were at Mir could get anxious with someone suggesting close flybys).