The obvious question that immediately came to my mind was "what is the new 'commonly-accepted heaviest stable element'"? And that question doesn't seem to be answered by the posting or the article. Anybody know the answer? Thanks.
Well, piracy still happens. People go out on their yachts and never come back, and the yacht is later discovered on the black market. The cannonballs are a lot smaller (.44, not 4) and few hoist the Jolly Roger or say "Arr, Matey!", but there is still a reason to patrol the seas.
No. Win 3.1 was cooperative (mostly -- DOS apps got pre-empted). Mac was cooperative (fully, as far as I know).
Win95 was cooperative only for Win16 apps. Otherwise it was pre-emptive.
Where it wasn't "industrial strength" was in resource compartmentalization. Many resources, including certain parts of memory, were shared between processes, allowing one process to take down the whole system. This was done for backwards compatibility as well as performance (more resource sharing meant less resource needed). This was similar to the Mac before OS X.
Win31 - Not much memory protection - Cooperative multitasking (pre-emptive for DOS) MacOS - Partial memory protection - Cooperative multitasking WinNT - Full memory protection - Pre-emptive multitasking Win95 - Partial memory protection - Pre-emptive multitasking MacOSX - Full memory protection - Pre-emptive multitasking
When it comes down to it, I don't think there is much to complain about on either side.
Microsoft has had NT for a long time and has wanted it to be the core OS for consumer and business as soon as the market was ready (since it required a beefier computer and caused some compatibility breaks). I'd say MS has been way head of Apple as far as underlying OS architecture goes for a long time (Apple finally imported Mach and BSD to make up for this).
On the other hand, for simplicity, ease of use, and user interface, Apple has been and still is ahead of Microsoft.
It therefore should surprise nobody that there is rampant "borrowing" going on in both camps. Kind of like the "innovations" in next year's Toyota Camry are going to be similar to the "new ideas" in next year's Nissan Altima. The industry advances together. Real "innovation" doesn't actually get to market for 5-20 years after it is discovered, and by that time, it is kind of hard to say who really innovated. And when it comes down to it, it doesn't really matter. What matters is who got the innovation into a product first, and who did the better job of turning the innovation into a useful feature.
Transmeta is now "edging out" (or perhaps more accurately "edging in on") Intel's StrongARM processor, not Intel's x86. Transmeta is putting the x86-compatible Crusoe in places where previously only embedded chips like ARM and MIPs were thriving. (Though the NS Geode has already made some inroads...)
So Windows NT 3.1, from 1994 (or thereabouts) and even Windows 95 (to a degree), was getting its wisdom on how to provide those features by embracing the black turtle neck?
Yeah, and what about Red Dwarf? And definitely the Millenium Falcon. And a bus full of penguins, gotta have the penguins. And how big is a Beowulf cluster of those? And yes, I'd like to Super-Size that.
It was stated that the system can handle up to 4 GB RAM. This is a bit ambiguous, but it could easily be interpreted as saying that the OS can only handle a 4 GB address space. If a device mapped any address space, the corresponding RAM wouldn't be needed. It could theoretically be mapped to addresses above 4 GB, but by our interpretation of the statement above, the OS doesn't access memory above 4 GB. So the memory just goes to waste.
I would have to look at the details on the specs for the OS first. But I wouldn't be surprised either way.
I'm not sure what you're asking, but if you are trying to come up with a team activity that would help them think hard, my favorite is "Puzzle Day". I've done it a few times, and I come off of it with the biggest high and a new confidence in my abilities to solve tough problems, as well as a respect for the abilities of my friends & co-workers.
My teams were 10 to 15 people, but two teams of 5 or 6 would probably work. Come up with some fun but HARD problems. Hard means that 5 people working for 12 hours might solve 5 out of 10 problems. Try to have a theme.
Puzzles come in all shapes and sizes. Cryptogram word-search puzzles are one example: take a word search puzzle, then replace all A's with G's, B's with T's, etc.
The best puzzles are layered: After solving the word search, the uncircled letters make up the phrase "fear of the great mole rat", and the final answer to the puzzle is "Zemmiphobia". Then all of the answers to all of the puzzles come together to solve one final puzzle...
Maybe this isn't your cup o' tea, but I sure found it fun...
You can't transparently cache, but you can set up an SUS server and point your clients at it. Software Update Services FAQ. I don't think it costs anything (beyond the cost of a Windows 2000 Server or Windows 2003 Server), and I don't see anywhere that it says you can only use it in a business... Wouldn't that work?
There was more to it than that. It appears that you followed the case via Slashdot, not by reading up on it yourself. There was more to the case than what appears on Slashdot, contrary to popular Slashdot herd mentality.
If platform-independence is Java's only redeeming feature, then it sucks, since 1) even without MS in the picture, it never delivered on this promise, and 2) that isn't enough to make a programming environment usable. Java has additional useful features, and Microsoft wanted to make them even better for Windows developers by adding a few extensions. That way, developers who were willing to give up platform independence could trade it in for access to extra features. That kind of action was and is normal (for Microsoft as well as everyone else -- take something, add extra value, and provide that to others). The only difference was that there was a license involved. Microsoft read the license in a way that they thought they had the right to do it. The courts decided otherwise. (Note that the case wasn't cut-and-dried, and that it required a court to decide how to interpret the contract, since Microsoft DID have license to do some very similar things. Microsoft wasn't in flagrant violation of the contract, just had interpreted it differently than Sun did, and the courts decided to favor Sun's interpretation. You ought to know how easy it is to misread legalese, and sometimes even lawyers disagree on how to interpret various statements. And even more likely, the MS developers were trying to come up with useful features, and failed to check with the legal department to see if what they were doing was ok by the license, since nobody likes to deal with Legal any more than is necessary!)
As a result of this court decision, the contract was now clarified. The court had decided on what the contract meant, and Microsoft was in violation of this meaning. Therefore, Microsoft was subject to penalties, and was required to abide by the new meaning of the contract. Sun, Microsoft, and the courts negotiated the penalties. One of the penalties was that Microsoft was no longer licensed to implement future versions of the Java spec. Another was that they had to stop shipping Java after a certain date. These were penalties that Sun requested, and I think they backfired.
So Sun has effectively told Microsoft to drop Java. For reasons of customer support, MS continues to fix the bugs in its JVM, but otherwise drops Java. After a while, it decides to get out of the Java market completely. MS doesn't ship the MS JVM with Windows XP, thus complying with the court order 2 years early. Sun has a hissy fit, since now Windows doesn't ship with Java. Microsoft wonders what the problem is: isn't that what you wanted in the first place?
Sun explicitly told MS to get out of the Java business. MS stayed in just to support their customers. When MS finally complies and gets out, Sun gets mad. Why? Because Sun's plan backfired: by getting Microsoft out of the Java business, Java is no longer installed by default on the most popular desktop platform. Sun asked for it, and Sun got it. They just didn't really want what they had asked for.
The new court order about Java comes because of Microsoft's monopoly situation. (Note: it is legal to be a monopoly, but when you are a monopoly, you have to abide by restrictions that other businesses do not have to abide by.)
I think a lot of people are barking up the wrong tree on this one. Here is a brief summary from the developer's perspective (as opposed to the marketer's or the lawyer's perspective):
1. If you are developing for Windows CE and you need source code to figure out what the OS is doing (for debugging, or to make up for anything left out of the documentation), you now have almost ALL of the source code, instead of SOME or NONE (like it was before).
2. If you need to add your own features by writing your own code (new DLLs, new programs, etc.) you do so. Microsoft doesn't care.
3. If you want to change any of Microsoft's source code that came with the Windows CE Platform Builder kit, you can change it, recompile it, and sell it, but you need to send your changes to Microsoft.
4. You still pay the same for copies of Windows CE that you distribute, whether or not you made any changes in the code.
#4 has been a fact of life for a long time -- you want to sell our OS, you pay us a royalty so that the nice guys who stayed up late at night writing our OS can feed their families.
#3 seems remarkably similar to the GPL -- if you make a change and sell it, you make your change available for everybody else. Of course, the code is now licensed to Microsoft. But if your code is generally applicable, you have 6 months to profit from it before anyone else steals your ideas (and they would have stolen them anyway, just now they get source). If your code is specific to your hardware (a very common situation), your code is now in the core OS distribution, so optimizations that make Windows CE work better with your hardware are now everywhere, making your hardware more attractive to potential customers.
The embedded world is a lot different from the software-only world. It is often much more like the hardware world. If you sell network cards, you WANT Microsoft to include support for your card in the OS. You WANT Microsoft to have the code for getting the best performance from your card. You would LOVE for Microsoft to let you suggest architectural changes to the OS that would make your product better.
Remember, this is all about give and take. Microsoft is now giving customers the ability to change the OS as needed. It is taking the changes and incorporating them into future editions of the product to make a better product. It basically picks up some of the benefits of open source, while still allowing all involved to pay the bills.
I think it is pretty cool. Everybody wins. And if a company decides that it isn't in their best interest to participate, they don't have to.
I don't know the particulars about the software you are using, but there is one of two things going on here: either the software simply sucks, and thus breaks when moving from one VM to another, or it uses lots of MS's extensions to Java (or both).
In the first case, this is definitely not the fault of Microsoft. Microsoft's JVM was written in the early days of Java when not everything in the spec was completely nailed down. (As far as that goes, it still isn't nailed down.) Microsoft's JVM implemented the spec just as well as Sun's, with the exception of Microsoft's extensions (which didn't break any code that was written according to the spec). There were some differences in interpretation between Sun's implementation and MS's. I know -- I've run into some in my own programs. But they are pretty reasonable differences in ambiguous situations, and once you know what the differences are, the problems are easy to fix. So if this is your problem, the fault lies with the company selling you lousy incompatible software and refusing to bring it up to standard when problem are found. They are depending on undocumented assumptions, and while this isn't necessarily a bad thing when there is no good documentation, it is a bad thing when they won't fix the problems when they are uncovered.
If it is the second case, then the argument is a bit different. The company decided to use Microsoft's extensions to gain additional features not available with Sun's JVM, with the cost of locking themselves to Microsoft's JVM. If they hadn't done this and needed the additional features, they wouldn't have been able to write their program in Java. Perhaps that would have been better. But Microsoft made the technology available, and never hid the fact that using the extensions would make the program incompatible with the Sun JVM.
I've used Microsoft's "controversial" J++ tool to make some applets that run just fine on Netscape's JVM, Microsoft's JVM, and Sun's JVM. There are a few places where I have to keep track of which JVM I'm running on so that I can use different workarounds for bugs or differences in implementation present in the JVMs (each has a few, especially when interacting with JavaScript or trying to access "privileged" operations). And as versions have been upgraded, I've had to track down new problems exposed by changes to the JVMs so that my clients could safely upgrade. Never did J++ twist my arm or force me to use incompatible extensions. In fact, if I accidentally used any, it would pop up compiler warnings so I could make sure my code was compatible.
They created a language with extra features. That was the name of the game back then, and still is. Pascal sucked, but Turbo Pascal, with its nifty extensions, was actually useful. Original HTML was pretty limited, but Mosaic added features allowing multimedia, and Netscape added JavaScript, and so on. Java was very limited since it couldn't access any native OS functionality, and Sun was stonewalling on Microsoft's proposals for Native APIs, so Microsoft added some extensions to fill in the gaps. The only difference between any of these scenarios was that Sun was maintaining a stranglehold on Java, and pounced on Microsoft for making the changes.
Sun committed Java-suicide when it told Microsoft that it couldn't mess with Java anymore. Microsoft stopped messing with Java (sounds reasonable enough, no?). Without active work from Microsoft, Java starts going down the toilet on the desktop. Microsoft finally decides to ditch it earlier than Sun's agreement would have forced them to do so. So Sun gets mad and cries to mommy about how mean Bill Gates won't ship Java for them. What did they expect -- Sun had already given them a court order to stop shipping Java, so why are they now mad when MS complies early?
Very good question. I suppose it is all about comittment. If you are married, it is understood that you have made a certain comittment and have agreed to certain obligations. Getting out of the comittment is possible, but definitely a hassle. And some of the oblications are permenant.
Though this is definitely a limitation on freedom and flexibility, that is not entirely a bad thing. Why do some computer languages enforce strict type checking? Not all limits on freedom are bad. They allow assumptions to be made. -- I can generally assume that I can drive through an intersection when the light is green and that cars will not drive in front of me.
A legal marriage provides at least a small degree of assurance that the partnership will not dissolve at the slightest whim. In addition, it provides some idea of what will happen if the partnership is dissolved. Without that, some risks might not be safe (such as merging assets, having children, or incurring other obligations such as a mortgage).
Finally, I think it is a good thing for some relationships that there is some barrier to separation. I mean, relationships go through rocky times, and sometimes they recover. From what I've seen, the strongest relationships are those that have survived challenges. A comittment such as marriage can be an obstacle to an overly hasty separation.
Yes, divorce sucks. But marriage is a sign of a real comittment. And real comittment means a loss of some freedom, such as the freedom to walk out at any time. I think I would rather have the comittment. (YMMV.)
Microsoft decided not to fix one vulnerability found in NT4. Why? This particular vulnerability should already not be an issue for most people (the port is normally filtered by most firewalls, and if you're running without a firewall, you're in trouble anyway). It has a workaround that works for most people. The fix would require significant code rework, and it is likely to break some existing applications. Microsoft doesn't want to introduce a major change this late in NT4's life cycle.
Why does the life cycle have anything to do with it? Basically, those who are still using NT4 for critical systems are likely to be those who don't want anything to change. They might have a setup that "just works" and they don't want to mess with it. Adding a patch that might change behavior, even if just a little bit, is risky. With 2000 and XP, this isn't such an issue: first, the code (supposedly) doesn't have to be changed quite as much; second, if it does change something, people are more likely to be actively developing for 2000 and XP, so they can adapt to the change.
In other words, both Microsoft and end users are less willing to make changes to NT4 systems than they are to newer OS'es. Stuff that is running on NT4 systems is likely to have less active support, and therefore is often best left alone or worked around if a problem arises.
Obviously, it would be nice if Microsoft were to make the patch available and then let users decide. But that would cost them developer/tester time, cost them support ("this patch screwed up my server!!"), and likely not provide any benefit that wasn't available just as easily with a supported workaround. Perhaps they are right to spend that same developer time on other issues that might actually be of benefit to somebody.
This has nothing to do with file sharing. That is a completely separate issue. (There isn't much money in it, so how could it be funding terrorism?)
Microsoft is talking about counterfeiting. There is a lot of money to be made by selling rip-off copies at discount prices. Think: it costs $.50 to burn a CD. If you can sell it for $10 to $20 bucks by ignoring copyright laws, you can make a lot of money. It is a profitable business. And it is very possible that some of the money made from this is used to support terrorism.
Basically, if there is a black market for something, some of the people making money from it are probably sending some of the money to "terrorist" causes. Software/Music/Movie counterfeiting is probably no different from any other black market industry. No worse, no better. Microsoft is just using this fact to try to get the government to crack down on it.
In 1999-2000, I had this computer that I used as a web server in college for a school group. It was a 486 with 16MB RAM running FreeBSD, Apache, PHP, etc. I got a friend to let me leave it in the corner of a building on campus. One cord to the power outlet, and one cord to the ethernet jack.
Well, one day I get an email that the server is dead. Web pages don't show up, but it responds to pings. I telnet in, but any command locks up the telnet session. So I run reboot, and it never comes back. Final diagnosis: hard drive failure.
Replaced the hard drive, and restored the web site. All is well until I get another email that the server is dead. No pings this time. Turns out that the water main in the floor above it had broken, and it had been thrown into a pile of computers that were behind a makeshift "dam". Once students were allowed back into the area, I searched around, found my computer, plugged it in, and found that it was once again working as expected.
Besides those two events, this old Gateway 486/66 never had to be rebooted or repaired. Ran without a hitch until I unplugged it on the last day of finals.
1. Telemarketers never call my cell phone (seems to be a common phenomenon, and I hope it stays that way). Yay! So now I just have a cell phone and no direct line. I suppose this doesn't work for everyone, and you have to find some other way to get Internet (like cable), but it worked for me.
2. Before cable was available, we had a phone line. And since once in a while it came in handy, we didn't turn off the ringer. Thus I learned that there is a moment of silence between when the autodialer detects that you answered the phone and when your call gets picked up by a telemarketroid. This is your golden opportunity to hang up before you get caught up in a conversation you don't want to have. Pick up the phone, keep a finger on the hook, say hello, and if you hear dead silence, hang up immediately. (If you made a mistake, they'll call back!)
As far as I know: ME and XP actually mean what they are advertised to mean (i.e. have no additional history beyond their invention by some guy in marketing) -- Millenium Edition and eXPerience.
NT, we now know (assuming our source is accurate), comes from N-Ten. Some marketroid came up with "New Technology" and a lot of perceptive geeks noticed that VMS --> WNT, but neither is the real origin of the acronym.
It took me almost 1 year actually WORKING on the Windows CE team before I found out what CE officially stands for (a lot of people on the team knew, but I just always asked the wrong people, and didn't really worry about the question very much...). CE means (or originally meant) "Consumer Electronics." However, for a while there was a whiteboard in the hall with a growing list of new interpretations for "CE"...
In a complex build environment, the BUILD ENVIRONMENT ITSELF changes from day to day. The various compilers, linkers, and "makefiles" (actually "sources" files in the Windows build system) change, and things like that don't get picked up from just an incremental build. For a complicated system, there is more to the build system than just "build the out of date files."
Slashdot Mirror
The obvious question that immediately came to my mind was "what is the new 'commonly-accepted heaviest stable element'"? And that question doesn't seem to be answered by the posting or the article. Anybody know the answer? Thanks.
Well, piracy still happens. People go out on their yachts and never come back, and the yacht is later discovered on the black market. The cannonballs are a lot smaller (.44, not 4) and few hoist the Jolly Roger or say "Arr, Matey!", but there is still a reason to patrol the seas.
Requires a 40GB hard drive for chat functionality? What the heck!?!
(I'm sure there is a good reason, but it just struck me as strange.)
No. Win 3.1 was cooperative (mostly -- DOS apps got pre-empted). Mac was cooperative (fully, as far as I know).
Win95 was cooperative only for Win16 apps. Otherwise it was pre-emptive.
Where it wasn't "industrial strength" was in resource compartmentalization. Many resources, including certain parts of memory, were shared between processes, allowing one process to take down the whole system. This was done for backwards compatibility as well as performance (more resource sharing meant less resource needed). This was similar to the Mac before OS X.
Win31 - Not much memory protection - Cooperative multitasking (pre-emptive for DOS)
MacOS - Partial memory protection - Cooperative multitasking
WinNT - Full memory protection - Pre-emptive multitasking
Win95 - Partial memory protection - Pre-emptive multitasking
MacOSX - Full memory protection - Pre-emptive multitasking
When it comes down to it, I don't think there is much to complain about on either side.
Microsoft has had NT for a long time and has wanted it to be the core OS for consumer and business as soon as the market was ready (since it required a beefier computer and caused some compatibility breaks). I'd say MS has been way head of Apple as far as underlying OS architecture goes for a long time (Apple finally imported Mach and BSD to make up for this).
On the other hand, for simplicity, ease of use, and user interface, Apple has been and still is ahead of Microsoft.
It therefore should surprise nobody that there is rampant "borrowing" going on in both camps. Kind of like the "innovations" in next year's Toyota Camry are going to be similar to the "new ideas" in next year's Nissan Altima. The industry advances together. Real "innovation" doesn't actually get to market for 5-20 years after it is discovered, and by that time, it is kind of hard to say who really innovated. And when it comes down to it, it doesn't really matter. What matters is who got the innovation into a product first, and who did the better job of turning the innovation into a useful feature.
Actually, this is still true.
:P
Transmeta is now "edging out" (or perhaps more accurately "edging in on") Intel's StrongARM processor, not Intel's x86. Transmeta is putting the x86-compatible Crusoe in places where previously only embedded chips like ARM and MIPs were thriving. (Though the NS Geode has already made some inroads...)
To boldly go where no x86 has gone before...
And how is that Macintosh 1997?
So Windows NT 3.1, from 1994 (or thereabouts) and even Windows 95 (to a degree), was getting its wisdom on how to provide those features by embracing the black turtle neck?
Huh?
Yeah, and what about Red Dwarf? And definitely the Millenium Falcon. And a bus full of penguins, gotta have the penguins. And how big is a Beowulf cluster of those? And yes, I'd like to Super-Size that.
Remember, there was a stock split in there somewhere... :P
I think what he's trying to get at is as follows:
It was stated that the system can handle up to 4 GB RAM. This is a bit ambiguous, but it could easily be interpreted as saying that the OS can only handle a 4 GB address space. If a device mapped any address space, the corresponding RAM wouldn't be needed. It could theoretically be mapped to addresses above 4 GB, but by our interpretation of the statement above, the OS doesn't access memory above 4 GB. So the memory just goes to waste.
I would have to look at the details on the specs for the OS first. But I wouldn't be surprised either way.
If going out and buying one Windows machine could save many gigabytes of bandwidth, would it be worth it?
I'm not sure what you're asking, but if you are trying to come up with a team activity that would help them think hard, my favorite is "Puzzle Day". I've done it a few times, and I come off of it with the biggest high and a new confidence in my abilities to solve tough problems, as well as a respect for the abilities of my friends & co-workers.
My teams were 10 to 15 people, but two teams of 5 or 6 would probably work. Come up with some fun but HARD problems. Hard means that 5 people working for 12 hours might solve 5 out of 10 problems. Try to have a theme.
Puzzles come in all shapes and sizes. Cryptogram word-search puzzles are one example: take a word search puzzle, then replace all A's with G's, B's with T's, etc.
The best puzzles are layered: After solving the word search, the uncircled letters make up the phrase "fear of the great mole rat", and the final answer to the puzzle is "Zemmiphobia". Then all of the answers to all of the puzzles come together to solve one final puzzle...
Maybe this isn't your cup o' tea, but I sure found it fun...
You can't transparently cache, but you can set up an SUS server and point your clients at it. Software Update Services FAQ. I don't think it costs anything (beyond the cost of a Windows 2000 Server or Windows 2003 Server), and I don't see anywhere that it says you can only use it in a business... Wouldn't that work?
Mod this up!!! This is the exact answer to the question.
There was more to it than that. It appears that you followed the case via Slashdot, not by reading up on it yourself. There was more to the case than what appears on Slashdot, contrary to popular Slashdot herd mentality.
If platform-independence is Java's only redeeming feature, then it sucks, since 1) even without MS in the picture, it never delivered on this promise, and 2) that isn't enough to make a programming environment usable. Java has additional useful features, and Microsoft wanted to make them even better for Windows developers by adding a few extensions. That way, developers who were willing to give up platform independence could trade it in for access to extra features. That kind of action was and is normal (for Microsoft as well as everyone else -- take something, add extra value, and provide that to others). The only difference was that there was a license involved. Microsoft read the license in a way that they thought they had the right to do it. The courts decided otherwise. (Note that the case wasn't cut-and-dried, and that it required a court to decide how to interpret the contract, since Microsoft DID have license to do some very similar things. Microsoft wasn't in flagrant violation of the contract, just had interpreted it differently than Sun did, and the courts decided to favor Sun's interpretation. You ought to know how easy it is to misread legalese, and sometimes even lawyers disagree on how to interpret various statements. And even more likely, the MS developers were trying to come up with useful features, and failed to check with the legal department to see if what they were doing was ok by the license, since nobody likes to deal with Legal any more than is necessary!)
As a result of this court decision, the contract was now clarified. The court had decided on what the contract meant, and Microsoft was in violation of this meaning. Therefore, Microsoft was subject to penalties, and was required to abide by the new meaning of the contract. Sun, Microsoft, and the courts negotiated the penalties. One of the penalties was that Microsoft was no longer licensed to implement future versions of the Java spec. Another was that they had to stop shipping Java after a certain date. These were penalties that Sun requested, and I think they backfired.
So Sun has effectively told Microsoft to drop Java. For reasons of customer support, MS continues to fix the bugs in its JVM, but otherwise drops Java. After a while, it decides to get out of the Java market completely. MS doesn't ship the MS JVM with Windows XP, thus complying with the court order 2 years early. Sun has a hissy fit, since now Windows doesn't ship with Java. Microsoft wonders what the problem is: isn't that what you wanted in the first place?
Sun explicitly told MS to get out of the Java business. MS stayed in just to support their customers. When MS finally complies and gets out, Sun gets mad. Why? Because Sun's plan backfired: by getting Microsoft out of the Java business, Java is no longer installed by default on the most popular desktop platform. Sun asked for it, and Sun got it. They just didn't really want what they had asked for.
The new court order about Java comes because of Microsoft's monopoly situation. (Note: it is legal to be a monopoly, but when you are a monopoly, you have to abide by restrictions that other businesses do not have to abide by.)
I think a lot of people are barking up the wrong tree on this one. Here is a brief summary from the developer's perspective (as opposed to the marketer's or the lawyer's perspective):
1. If you are developing for Windows CE and you need source code to figure out what the OS is doing (for debugging, or to make up for anything left out of the documentation), you now have almost ALL of the source code, instead of SOME or NONE (like it was before).
2. If you need to add your own features by writing your own code (new DLLs, new programs, etc.) you do so. Microsoft doesn't care.
3. If you want to change any of Microsoft's source code that came with the Windows CE Platform Builder kit, you can change it, recompile it, and sell it, but you need to send your changes to Microsoft.
4. You still pay the same for copies of Windows CE that you distribute, whether or not you made any changes in the code.
#4 has been a fact of life for a long time -- you want to sell our OS, you pay us a royalty so that the nice guys who stayed up late at night writing our OS can feed their families.
#3 seems remarkably similar to the GPL -- if you make a change and sell it, you make your change available for everybody else. Of course, the code is now licensed to Microsoft. But if your code is generally applicable, you have 6 months to profit from it before anyone else steals your ideas (and they would have stolen them anyway, just now they get source). If your code is specific to your hardware (a very common situation), your code is now in the core OS distribution, so optimizations that make Windows CE work better with your hardware are now everywhere, making your hardware more attractive to potential customers.
The embedded world is a lot different from the software-only world. It is often much more like the hardware world. If you sell network cards, you WANT Microsoft to include support for your card in the OS. You WANT Microsoft to have the code for getting the best performance from your card. You would LOVE for Microsoft to let you suggest architectural changes to the OS that would make your product better.
Remember, this is all about give and take. Microsoft is now giving customers the ability to change the OS as needed. It is taking the changes and incorporating them into future editions of the product to make a better product. It basically picks up some of the benefits of open source, while still allowing all involved to pay the bills.
I think it is pretty cool. Everybody wins. And if a company decides that it isn't in their best interest to participate, they don't have to.
I don't know the particulars about the software you are using, but there is one of two things going on here: either the software simply sucks, and thus breaks when moving from one VM to another, or it uses lots of MS's extensions to Java (or both).
In the first case, this is definitely not the fault of Microsoft. Microsoft's JVM was written in the early days of Java when not everything in the spec was completely nailed down. (As far as that goes, it still isn't nailed down.) Microsoft's JVM implemented the spec just as well as Sun's, with the exception of Microsoft's extensions (which didn't break any code that was written according to the spec). There were some differences in interpretation between Sun's implementation and MS's. I know -- I've run into some in my own programs. But they are pretty reasonable differences in ambiguous situations, and once you know what the differences are, the problems are easy to fix. So if this is your problem, the fault lies with the company selling you lousy incompatible software and refusing to bring it up to standard when problem are found. They are depending on undocumented assumptions, and while this isn't necessarily a bad thing when there is no good documentation, it is a bad thing when they won't fix the problems when they are uncovered.
If it is the second case, then the argument is a bit different. The company decided to use Microsoft's extensions to gain additional features not available with Sun's JVM, with the cost of locking themselves to Microsoft's JVM. If they hadn't done this and needed the additional features, they wouldn't have been able to write their program in Java. Perhaps that would have been better. But Microsoft made the technology available, and never hid the fact that using the extensions would make the program incompatible with the Sun JVM.
I've used Microsoft's "controversial" J++ tool to make some applets that run just fine on Netscape's JVM, Microsoft's JVM, and Sun's JVM. There are a few places where I have to keep track of which JVM I'm running on so that I can use different workarounds for bugs or differences in implementation present in the JVMs (each has a few, especially when interacting with JavaScript or trying to access "privileged" operations). And as versions have been upgraded, I've had to track down new problems exposed by changes to the JVMs so that my clients could safely upgrade. Never did J++ twist my arm or force me to use incompatible extensions. In fact, if I accidentally used any, it would pop up compiler warnings so I could make sure my code was compatible.
They created a language with extra features. That was the name of the game back then, and still is. Pascal sucked, but Turbo Pascal, with its nifty extensions, was actually useful. Original HTML was pretty limited, but Mosaic added features allowing multimedia, and Netscape added JavaScript, and so on. Java was very limited since it couldn't access any native OS functionality, and Sun was stonewalling on Microsoft's proposals for Native APIs, so Microsoft added some extensions to fill in the gaps. The only difference between any of these scenarios was that Sun was maintaining a stranglehold on Java, and pounced on Microsoft for making the changes.
Sun committed Java-suicide when it told Microsoft that it couldn't mess with Java anymore. Microsoft stopped messing with Java (sounds reasonable enough, no?). Without active work from Microsoft, Java starts going down the toilet on the desktop. Microsoft finally decides to ditch it earlier than Sun's agreement would have forced them to do so. So Sun gets mad and cries to mommy about how mean Bill Gates won't ship Java for them. What did they expect -- Sun had already given them a court order to stop shipping Java, so why are they now mad when MS complies early?
And I can't spell. It's "commitment" not "comittment." Whoops.
Very good question. I suppose it is all about comittment. If you are married, it is understood that you have made a certain comittment and have agreed to certain obligations. Getting out of the comittment is possible, but definitely a hassle. And some of the oblications are permenant.
Though this is definitely a limitation on freedom and flexibility, that is not entirely a bad thing. Why do some computer languages enforce strict type checking? Not all limits on freedom are bad. They allow assumptions to be made. -- I can generally assume that I can drive through an intersection when the light is green and that cars will not drive in front of me.
A legal marriage provides at least a small degree of assurance that the partnership will not dissolve at the slightest whim. In addition, it provides some idea of what will happen if the partnership is dissolved. Without that, some risks might not be safe (such as merging assets, having children, or incurring other obligations such as a mortgage).
Finally, I think it is a good thing for some relationships that there is some barrier to separation. I mean, relationships go through rocky times, and sometimes they recover. From what I've seen, the strongest relationships are those that have survived challenges. A comittment such as marriage can be an obstacle to an overly hasty separation.
Yes, divorce sucks. But marriage is a sign of a real comittment. And real comittment means a loss of some freedom, such as the freedom to walk out at any time. I think I would rather have the comittment. (YMMV.)
They decided that one particular vulnerability wouldn't be fixed. That is not the same as refuse-to-support.
(And there are some decent reasons, even from the consumer's point of view, for why this one won't be fixed. So please just let it go.)
Sheesh! Enough already.
Microsoft decided not to fix one vulnerability found in NT4. Why? This particular vulnerability should already not be an issue for most people (the port is normally filtered by most firewalls, and if you're running without a firewall, you're in trouble anyway). It has a workaround that works for most people. The fix would require significant code rework, and it is likely to break some existing applications. Microsoft doesn't want to introduce a major change this late in NT4's life cycle.
Why does the life cycle have anything to do with it? Basically, those who are still using NT4 for critical systems are likely to be those who don't want anything to change. They might have a setup that "just works" and they don't want to mess with it. Adding a patch that might change behavior, even if just a little bit, is risky. With 2000 and XP, this isn't such an issue: first, the code (supposedly) doesn't have to be changed quite as much; second, if it does change something, people are more likely to be actively developing for 2000 and XP, so they can adapt to the change.
In other words, both Microsoft and end users are less willing to make changes to NT4 systems than they are to newer OS'es. Stuff that is running on NT4 systems is likely to have less active support, and therefore is often best left alone or worked around if a problem arises.
Obviously, it would be nice if Microsoft were to make the patch available and then let users decide. But that would cost them developer/tester time, cost them support ("this patch screwed up my server!!"), and likely not provide any benefit that wasn't available just as easily with a supported workaround. Perhaps they are right to spend that same developer time on other issues that might actually be of benefit to somebody.
HELLO, PEOPLE!
This has nothing to do with file sharing. That is a completely separate issue. (There isn't much money in it, so how could it be funding terrorism?)
Microsoft is talking about counterfeiting. There is a lot of money to be made by selling rip-off copies at discount prices. Think: it costs $.50 to burn a CD. If you can sell it for $10 to $20 bucks by ignoring copyright laws, you can make a lot of money. It is a profitable business. And it is very possible that some of the money made from this is used to support terrorism.
Basically, if there is a black market for something, some of the people making money from it are probably sending some of the money to "terrorist" causes. Software/Music/Movie counterfeiting is probably no different from any other black market industry. No worse, no better. Microsoft is just using this fact to try to get the government to crack down on it.
In 1999-2000, I had this computer that I used as a web server in college for a school group. It was a 486 with 16MB RAM running FreeBSD, Apache, PHP, etc. I got a friend to let me leave it in the corner of a building on campus. One cord to the power outlet, and one cord to the ethernet jack.
Well, one day I get an email that the server is dead. Web pages don't show up, but it responds to pings. I telnet in, but any command locks up the telnet session. So I run reboot, and it never comes back. Final diagnosis: hard drive failure.
Replaced the hard drive, and restored the web site. All is well until I get another email that the server is dead. No pings this time. Turns out that the water main in the floor above it had broken, and it had been thrown into a pile of computers that were behind a makeshift "dam". Once students were allowed back into the area, I searched around, found my computer, plugged it in, and found that it was once again working as expected.
Besides those two events, this old Gateway 486/66 never had to be rebooted or repaired. Ran without a hitch until I unplugged it on the last day of finals.
Just goes to show that BSD will never die...
Well, it worked for me, anyway.
1. Telemarketers never call my cell phone (seems to be a common phenomenon, and I hope it stays that way). Yay! So now I just have a cell phone and no direct line. I suppose this doesn't work for everyone, and you have to find some other way to get Internet (like cable), but it worked for me.
2. Before cable was available, we had a phone line. And since once in a while it came in handy, we didn't turn off the ringer. Thus I learned that there is a moment of silence between when the autodialer detects that you answered the phone and when your call gets picked up by a telemarketroid. This is your golden opportunity to hang up before you get caught up in a conversation you don't want to have. Pick up the phone, keep a finger on the hook, say hello, and if you hear dead silence, hang up immediately. (If you made a mistake, they'll call back!)
As far as I know: ME and XP actually mean what they are advertised to mean (i.e. have no additional history beyond their invention by some guy in marketing) -- Millenium Edition and eXPerience.
NT, we now know (assuming our source is accurate), comes from N-Ten. Some marketroid came up with "New Technology" and a lot of perceptive geeks noticed that VMS --> WNT, but neither is the real origin of the acronym.
It took me almost 1 year actually WORKING on the Windows CE team before I found out what CE officially stands for (a lot of people on the team knew, but I just always asked the wrong people, and didn't really worry about the question very much...). CE means (or originally meant) "Consumer Electronics." However, for a while there was a whiteboard in the hall with a growing list of new interpretations for "CE"...
In a complex build environment, the BUILD ENVIRONMENT ITSELF changes from day to day. The various compilers, linkers, and "makefiles" (actually "sources" files in the Windows build system) change, and things like that don't get picked up from just an incremental build. For a complicated system, there is more to the build system than just "build the out of date files."