If you do sign an NDA as part of a job contract, it expires as soon as your job assignment ends. You are free to use whatever knowledge you have inside your head. (Taking along actual source code with you is another matter. In that case, you are actually employed, and the NDA is in effect.)
If you really want to know, ask your doctor. He will be able to tell you what the effects of HGH are, whether or not you can legally obtain it. In most countries, it would be classified as medication and thus subject to several regulations. Ordering this product from a spammer could be illegal.
If your doctor decides it would have beneficial effects on you and gives you a prescription, use that to obtain it, and not buy from the spammer. After all, you don't know what they will sell you, and when you have already paid, it is too late.
Whatever you do, do not buy anything from a spammer. (You don't want them to use you as an example to show that "spam works", do you?)
Personally, I wouldn't believe them. After all, if it really works, would they have to resort to spam to sell it?
Microsoft's strategy has always been to bind everything to the Windows environment: develop on Windows, run on Windows. That is one of the reasons they went after Java: Java would allow people to easily develop programs that don't run on Windows.
Occasionally, they have made software that ran on other platforms: Office for Mac (mainly to avoid anti-trust issues),.Net for *BSD.
From MS' point of view, it doesn't make sense to create Linux software. They'd lose their main advantage: expertise of the Windows platform. Not because of "hidden APIs", but just because their application developers are Windows developers (rather obvious). It would take a lot of time before they know Linux well enough to develop for it.
On the other hand, with Microsoft, nothing is impossible. It wouldn't be the first time they'd change their policy. Just don't bet on it.
On the gripping hand, even if they would produce Linux software, it won't be open source (as noted by other posters). Shared Source seems likely, BSD licence for a few parts. GPL is out of the question. (As noted above, nothing is impossible, but Microsoft releasing GPL'd code is extremely unlikely.)
if I'm silent while driving, I'm not sulking (I'm concentrating on driving)
Sounds familiar. My girlfriend insists on having conversations in the car while I'm driving. That is terribly distracting.
Even worse are the "alarm signals". Ever had someone sitting next to you yell "STOP!" or "WATCH OUT!"? It gets to you. (Most of the time, there's no danger at all. And when there is something really dangerous, I saw it and I don't need someone to scream about it.)
I'm looking forward to the day she has her driver's licence, then I can let her drive and feel for herself how difficult it is. (My mother has the same kind of reactions, and her husband always lets her drive.)
This reminds me of a story I heard a few years ago:
My friend was a network admin at the time. One morning, he gets a message from a remote site: a control unit was down. He checks, and sees everything is OK. Next morning, another report about the same control unit. Looking more closely, my friend notices that the machine has shut down at 7:45, and came back 15 minutes later.
So he activates all logging on the control unit. Next day: logging stops at 7:45 for no apparent reason. Machine reboots at 8 o'clock.
My friend decides to go to the machine and monitor it in person. Next morning, he goes to the remote site, locates the control unit and waits.
7:45. Nothing happens. 8:00. Still everything is all right. Next day (when my friend is back on his regular place), machine goes down at 7:45.
My friend goes back to the scene of the crime and now he hides in the closet. 7:45. A man walks up to the control unit, pulls the power cord and plugs in an electrical razor.
My friend was so astonished that he just stayed in the closet watching this man shave himself.
... "if anything's wrong with my computer, it's a virus" phenomena
True story: a while ago, my girlfriend's PC was acting strange, and she thought it was a virus infection. To me, the symptoms didn't seem like a virus. After looking around in the system settings, I figured out that the proxy server was fubar, so I disabled it, hereby solving the problem.
Yes, his name is Dave Cutler. IIRC he quit working for M$ in discust about the time NT 3.5.1 was released (please correct me here).
IIRC, MS hired him back to work on W2k.
The Win cmd shell is nothing like VMS.
True. The kernel underneath is similar - some of the internal structures have the same names (that's Cutler's heritage), but the user interface is different. Even the kernel has a different design: WinNT is a microkernel, while OpenVMS is monolithic.
Correct, at one point in time, Digital's marketing division decided to rename it to OpenVMS. The change was made retroactively for all versions.
Digital vision on the word "Open" was "support all industry standards". Technically, a lot of things were changed as well. (TCP/IP, POSIX support, DECNET V that ran on top of TCP/IP instead of it's own transport...)
This name change happened at the same time as the introduction of the Alpha AXP, causing many people to believe this caused the change. The new name applies to all VMS versions, even the ones that only run on VAX. (IIRC, Alpha support was added in version 6.)
When MS removed their JVM from Windows, it opened up new possibilities for Java:
The "Windows only" extensions would die, so that everyone could concentrate on writing "real" Java programs.
Removing Java from the browser will lead to a more wide-spread adoption of SUN's Java plugin. This eliminates the need for browser-specific applets, and even more important, it allows Swing applets. (Almost nobody uses applets nowadays, but this could change.)
So singling out a single pkg format is just ridiculous.
I don't think the poster specifically meant to target RPMs, but included them simply as an example.
All distributions I know supply the MD5 sums for their packages (whether.rpm,.tgz,.deb or other). It's up to the user to check them (and as the poster pointed out, they probably don't).
Signing/checking with public key signing is much much better.
Debian packages are signed with the maintainer's key. Unfortunately, the standard configuration file disables the checking of this signature.
noexec disables the execution of binaries FROM that drive. Programs can still affect it
In most cases, users do not have write access outside of their home directory. If they are stupid enough to save the program to disk, they won't be able to execute it. (Of course, you would need the noexec on/tmp and/var/tmp as well to make this effective.)
This is obviously not an option for a development system: programmers would not be able to test their work. (Supplying a directory where they can run programs would nullify the advantage of the noexec environment.)
Most code tends to include a lot of error handling stuff. In C, you can end up with more than half of your code devoted to error handling.
(C++ and Java are better, because of exceptions.)
If you read such a program, you'll have a hard time figuring out what it does in the general (non-error) case.
Cobalt used to sell Linux rack servers before they were bought by SUN. I believe SUN plans to migrate those products to Solaris, but I'm not sure whether this has been done already.
At my current assignment, we are implementing a web site generation system, backed up by the SilverStream e-portal CMS.
The system will be used to create a web site, based on the information in the CMS database. The whole system is implemented using standard JSP components (including struts), replacing the e-portal GUI.
There are plans to do an Interwoven roll-out at this company (mainly for political reasons). But even then, our code will be used because Interwoven doesn't have the capacities for on-line content delivery (at least not in the way we need it). In this scenario, content would be transfered between Interwoven and Silverstream for the online delivery.
At my own company, they use Silverstream as well for a CMS, but there they use Silverstream's interface.
Not true. This is what the FHS says about/usr/local:
The place for locally installed software and other files. Distributions may not install anything in here. It is reserved solely for the use of the local administrator. This way he can be absolutely certain that no updates or upgrades to his distribution will overwrite any extra software he has installed locally.
/opt is not mentioned as far as I can see. I remember reading that it was deprecated.
/usr/local is not obsolete, and won't be. The only rule is that a package manager (dpkg, rpm,...) should never touch that directory (beyond creating it on a new install).
Before Windows 98 or 2000, setting up a TCP/IP network was not so easy.
Either you had a DHCP server (small chance in a home network), or you'd had to assign an IP address to each PC.
IPX on the other hand, "just works" out of the box for small networks (like a typical LAN party). Simply connect a couple PCs, and they'll figure out the network themselves. (This has always been one of the strongest selling points for Novell.)
IMHO, that is the real reason for IPX games.
Windows 98 and later have a feature to autoconfigure a TCP/IP network, in much the same way as IPX. Novell has switched to TCP/IP for Netware servers as well, making IPX largely obsolete.
I think you're mixing two concepts here: garbage collection and heap optimization.
A GC finds all unused objects and marks them as free. Heap optimization shuffles memory around to reduce fragmentation. Both systems can be implemented independently. (E.g., older versions of MacOS would reorganize the memory map of running processes from time to time, but AFAIK this wasn't a garbage collector.)
The SUN Java GC does use heap compaction techniques, but this is not required by the JVM specification.
In every correct application, heap space grows until an equilibrium is reached. The allocated memory may be fragmented, but new objects merely fill the "holes" left by previously released objects.
Sometimes, an object is not deallocated after all references to it are cleared. This is a memory leak. A GC serves to find and release these objects. (You can still have memory leaks in a system with a GC: if the program keeps references to objects that are no longer used.)
In a long-running program, those memory leaks will have a larger impact on the footprint of the process than those holes.
I have been using the 2.4 kernel on my home system since the beginning. The early releases were quite stable, but there were issues that required an upgrade.
But then, each new release seemed to deteriorate. System lock ups, broken drivers, you name it. Some kernels even refused to boot. I was lucky I never lost any data.
Around 2.4.13, things had become so bad that I decided to skip a few releases. And after the next crash, I reverted to 2.4.9 (which still had problems, but it was bearable - I was lucky I still had it: I bumped into random errors while compiling a new kernel).
In 2.4.16 and 17, things finally became stable. At last. I hope Marcelo Tosatti keeps stability as the main priority of his job as kernel maintainer.
The 2.4 series has had more than its share of brown-paper-bag bugs. Up to the point that 2 releases didn't even compile properly.
Switching VM implementations in a stable kernel series has been one of Linus' greatest mistakes. I hope people learn from this, and that some kind of QA will be installed.
Under CP/M, it's even easier to undelete a file: all metadata was stored in the directory. (Flat file system - only a root directory.)
When a file was deleted, a single byte - the "user number" was changed. If you wanted to undelete a file, you only needed to restore that byte (which would be 0 for the single-user version).
Slashdot Mirror
Short answer: no.
If you do sign an NDA as part of a job contract, it expires as soon as your job assignment ends. You are free to use whatever knowledge you have inside your head. (Taking along actual source code with you is another matter. In that case, you are actually employed, and the NDA is in effect.)
(IANAL, so don't take this as legal advice.)
If you really want to know, ask your doctor. He will be able to tell you what the effects of HGH are, whether or not you can legally obtain it. In most countries, it would be classified as medication and thus subject to several regulations. Ordering this product from a spammer could be illegal.
If your doctor decides it would have beneficial effects on you and gives you a prescription, use that to obtain it, and not buy from the spammer. After all, you don't know what they will sell you, and when you have already paid, it is too late.
Whatever you do, do not buy anything from a spammer. (You don't want them to use you as an example to show that "spam works", do you?)
Personally, I wouldn't believe them. After all, if it really works, would they have to resort to spam to sell it?
Microsoft's strategy has always been to bind everything to the Windows environment: develop on Windows, run on Windows. That is one of the reasons they went after Java: Java would allow people to easily develop programs that don't run on Windows.
Occasionally, they have made software that ran on other platforms: Office for Mac (mainly to avoid anti-trust issues), .Net for *BSD.
From MS' point of view, it doesn't make sense to create Linux software. They'd lose their main advantage: expertise of the Windows platform. Not because of "hidden APIs", but just because their application developers are Windows developers (rather obvious). It would take a lot of time before they know Linux well enough to develop for it.
On the other hand, with Microsoft, nothing is impossible. It wouldn't be the first time they'd change their policy. Just don't bet on it.
On the gripping hand, even if they would produce Linux software, it won't be open source (as noted by other posters). Shared Source seems likely, BSD licence for a few parts. GPL is out of the question. (As noted above, nothing is impossible, but Microsoft releasing GPL'd code is extremely unlikely.)
if I'm silent while driving, I'm not sulking (I'm concentrating on driving)
Sounds familiar. My girlfriend insists on having conversations in the car while I'm driving. That is terribly distracting.
Even worse are the "alarm signals". Ever had someone sitting next to you yell "STOP!" or "WATCH OUT!"? It gets to you. (Most of the time, there's no danger at all. And when there is something really dangerous, I saw it and I don't need someone to scream about it.)
I'm looking forward to the day she has her driver's licence, then I can let her drive and feel for herself how difficult it is. (My mother has the same kind of reactions, and her husband always lets her drive.)
LOL
This reminds me of a story I heard a few years ago:
My friend was a network admin at the time. One morning, he gets a message from a remote site: a control unit was down. He checks, and sees everything is OK. Next morning, another report about the same control unit. Looking more closely, my friend notices that the machine has shut down at 7:45, and came back 15 minutes later.
So he activates all logging on the control unit. Next day: logging stops at 7:45 for no apparent reason. Machine reboots at 8 o'clock.
My friend decides to go to the machine and monitor it in person. Next morning, he goes to the remote site, locates the control unit and waits.
7:45. Nothing happens. 8:00. Still everything is all right. Next day (when my friend is back on his regular place), machine goes down at 7:45.
My friend goes back to the scene of the crime and now he hides in the closet. 7:45. A man walks up to the control unit, pulls the power cord and plugs in an electrical razor.
My friend was so astonished that he just stayed in the closet watching this man shave himself.
w
The colon on the first line is an older version of the #! line, but only works for sh. And of course, `w' is one character less than `ls'.
On systems that automatically use /bin/sh on unknown files, the smallest possible shell script is:
wYes, a single character.
He was no dummy, but he carried the misbelief that a computer virus was something that occurred naturally, like an influenza virus, or herpes.
You might like this story I got today.
True story: a while ago, my girlfriend's PC was acting strange, and she thought it was a virus infection. To me, the symptoms didn't seem like a virus. After looking around in the system settings, I figured out that the proxy server was fubar, so I disabled it, hereby solving the problem.
I still remember the Y2K transition. We started on Monday, and everything was ready by Fridak. And the next week we did the months.
Then we started to suspect that something might go wrong in 00...
I got an AMD ad when I read the article. Thought it was pretty ironic.
LOL. You forgot a few things: the screen displays green characters on a black background., with a huge blinking cursor.
And the monitor displays 40 characters per line.
This new OS featured in User Friendly some time ago.
Yes, his name is Dave Cutler. IIRC he quit working for M$ in discust about the time NT 3.5.1 was released (please correct me here).
IIRC, MS hired him back to work on W2k.
The Win cmd shell is nothing like VMS.
True. The kernel underneath is similar - some of the internal structures have the same names (that's Cutler's heritage), but the user interface is different. Even the kernel has a different design: WinNT is a microkernel, while OpenVMS is monolithic.
Correct, at one point in time, Digital's marketing division decided to rename it to OpenVMS. The change was made retroactively for all versions.
Digital vision on the word "Open" was "support all industry standards". Technically, a lot of things were changed as well. (TCP/IP, POSIX support, DECNET V that ran on top of TCP/IP instead of it's own transport...)
This name change happened at the same time as the introduction of the Alpha AXP, causing many people to believe this caused the change. The new name applies to all VMS versions, even the ones that only run on VAX. (IIRC, Alpha support was added in version 6.)
Will there be an 'official' 2.4.19 kernel image package for Debian Woody?
No. Woody is stable now, meaning that no new packages can be added, and updates are only allowed in case of very serious problems.
Just take the package from testing (Sarge) when it appears.
True.
When MS removed their JVM from Windows, it opened up new possibilities for Java:
So singling out a single pkg format is just ridiculous.
I don't think the poster specifically meant to target RPMs, but included them simply as an example.
All distributions I know supply the MD5 sums for their packages (whether .rpm, .tgz, .deb or other). It's up to the user to check them (and as the poster pointed out, they probably don't).
Signing/checking with public key signing is much much better.
Debian packages are signed with the maintainer's key. Unfortunately, the standard configuration file disables the checking of this signature.
noexec disables the execution of binaries FROM that drive. Programs can still affect it
In most cases, users do not have write access outside of their home directory. If they are stupid enough to save the program to disk, they won't be able to execute it. (Of course, you would need the noexec on /tmp and /var/tmp as well to make this effective.)
This is obviously not an option for a development system: programmers would not be able to test their work. (Supplying a directory where they can run programs would nullify the advantage of the noexec environment.)
The best comment is the code.
Actually, it is the worst comment.
Most code tends to include a lot of error handling stuff. In C, you can end up with more than half of your code devoted to error handling. (C++ and Java are better, because of exceptions.)
If you read such a program, you'll have a hard time figuring out what it does in the general (non-error) case.
Cobalt used to sell Linux rack servers before they were bought by SUN. I believe SUN plans to migrate those products to Solaris, but I'm not sure whether this has been done already.
(I rent a site on one of those boxes myself.)
Control-Shift-L pops up a window with more options (like open link in new window.)
At my current assignment, we are implementing a web site generation system, backed up by the SilverStream e-portal CMS.
The system will be used to create a web site, based on the information in the CMS database. The whole system is implemented using standard JSP components (including struts), replacing the e-portal GUI.
There are plans to do an Interwoven roll-out at this company (mainly for political reasons). But even then, our code will be used because Interwoven doesn't have the capacities for on-line content delivery (at least not in the way we need it). In this scenario, content would be transfered between Interwoven and Silverstream for the online delivery.
At my own company, they use Silverstream as well for a CMS, but there they use Silverstream's interface.
I understand that this is directly from the FHS.
Not true. This is what the FHS says about /usr/local:
Before Windows 98 or 2000, setting up a TCP/IP network was not so easy.
Either you had a DHCP server (small chance in a home network), or you'd had to assign an IP address to each PC.
IPX on the other hand, "just works" out of the box for small networks (like a typical LAN party). Simply connect a couple PCs, and they'll figure out the network themselves. (This has always been one of the strongest selling points for Novell.)
IMHO, that is the real reason for IPX games.
Windows 98 and later have a feature to autoconfigure a TCP/IP network, in much the same way as IPX. Novell has switched to TCP/IP for Netware servers as well, making IPX largely obsolete.
I think you're mixing two concepts here: garbage collection and heap optimization.
A GC finds all unused objects and marks them as free. Heap optimization shuffles memory around to reduce fragmentation. Both systems can be implemented independently. (E.g., older versions of MacOS would reorganize the memory map of running processes from time to time, but AFAIK this wasn't a garbage collector.)
The SUN Java GC does use heap compaction techniques, but this is not required by the JVM specification.
In every correct application, heap space grows until an equilibrium is reached. The allocated memory may be fragmented, but new objects merely fill the "holes" left by previously released objects.
Sometimes, an object is not deallocated after all references to it are cleared. This is a memory leak. A GC serves to find and release these objects. (You can still have memory leaks in a system with a GC: if the program keeps references to objects that are no longer used.)
In a long-running program, those memory leaks will have a larger impact on the footprint of the process than those holes.
I have been using the 2.4 kernel on my home system since the beginning. The early releases were quite stable, but there were issues that required an upgrade.
But then, each new release seemed to deteriorate. System lock ups, broken drivers, you name it. Some kernels even refused to boot. I was lucky I never lost any data.
Around 2.4.13, things had become so bad that I decided to skip a few releases. And after the next crash, I reverted to 2.4.9 (which still had problems, but it was bearable - I was lucky I still had it: I bumped into random errors while compiling a new kernel).
In 2.4.16 and 17, things finally became stable. At last. I hope Marcelo Tosatti keeps stability as the main priority of his job as kernel maintainer.
The 2.4 series has had more than its share of brown-paper-bag bugs. Up to the point that 2 releases didn't even compile properly.
Switching VM implementations in a stable kernel series has been one of Linus' greatest mistakes. I hope people learn from this, and that some kind of QA will be installed.
Under CP/M, it's even easier to undelete a file: all metadata was stored in the directory. (Flat file system - only a root directory.)
When a file was deleted, a single byte - the "user number" was changed. If you wanted to undelete a file, you only needed to restore that byte (which would be 0 for the single-user version).