Being conservative, there is about 1 bug per line of code in, say, Windows 2000.
Windows 2000/IIS has about 2,000,000 lines of code.
That gives Microsoft about 0.0432 seconds per line of code, if they stop and work on only Windows 2000/IIS.
I wish them luck.
Ahh yes, and here's where we have our fully redundant Linux cluster with an uptime of 2 years. If anything does ever go wrong, we have a Lego Mindstorms backup system in place. I think we can trust that the database will never be lost.
Production environments? That's what FreeBSD is for. >
Seriously though (and I was serious), most companies with experiences IT people aren't going to use anything that hasn't been thoroughly tested internally (like the latest releases of the kernel that have the new VM), but then most companies probably wouldn't use MySQL in a production environment either...though I would argue that it is more solid than most operating systems.
LOL, good argument. In my experience, C can do everything except certain very specific platform dependant things that ONLY assembly can do. What, really, can literally not be done in C that can be done in any other HLL?
Certainly, things can be done more easily, or faster, or more quickly, but as far as a specific programming task that C simply cannot do--I have yet to see one.
1) The better organization (and more sensical naming) are part of what make Java easier to learn, giving the programmer more time to master.
2) How many Joe Programmers do you know that can write their own garbage collector? What if that custom garbage collector has horrible bugs? When you need to write a program quickly and its stability and security are paramount, Java all but eliminates these worries. As far as Lisp and others having garbage collection--yes, that is true, but few programmers know Lisp (other than tweaking a little EMACS). Why is NT so popular for networking? For its merits? (HA!). It's much easier to find an MCSE than a Unix admin--and cheaper. The same goes with Java VS. Lisp, Smalltalk, or even C++.
3) This often isn't that big of a deal, but it is great for the 10% of the time that it is a big deal. As far as browser incompatibilities, sticking with basic Java 1.1 or 1.2 tends to be very safe, but I haven't delved much into that realm.
4) I wouldn't really mention that part if I were asked about the merits of Java, but to Java's credit its GUI libraries are part of the language spec. GTK+ and QT are definitely not part of the C spec, so it is not reasonable to expect a platform that can run C apps (all of them) will be able to run C apps that call GTK+ or QT functions. If you make a button in Java, that button will work simply by virtue of the fact that the platform runs Java. At least in theory...
5) Me too.
6) I know little of Javabeans.
Hope this helps a little.
Note that despite the above, I still prefer C or PERL for almost everything.
The Ultra-SPARC chips are faster at the same clockspeed period. They are server class 64-bit chips with integrated PCI controllers running on an OS whose core parts were largely written in optimized assembly language. You'll notice that Java runs faster on an AthlonXP than A Pentium 4 as well--not because the XP implimentation is better (in fact, it will probably be the exact same), but because the XP is a faster chip/Hz.
That said, Sun may have spent extra time optimizing the JVM for the SPARC... It makes sense for a hardware company to do so. I'd love to run my java apps on a Sun, but any decent maching startng at around $30,000 turns me off.
FreeBSD has long been one of the top performers--if not THE top performer--in TCP/IP. How they made a substantial improvement to code that has been maturing longer than the networking code for any other OS in common use today is amazing.
Can you imagine how skilled and bored the developers much be?
Unfortunately,it will take about 2 weeks for my 4.5CD to arrive, but all good things are worth the wait.
Good. I find those movies incredibly irritating. I suppose that it is bad that those who enjoy that sort of humor may not be able to, but I believe that this loss will prevent another decline in the average movie viewer's intellect.
I really like FreeBSD and X, both of which have BSD style licenses.
However, every time I think of Microsoft getting the benefit of the FreeBSD networking code that took years to polish, and not even giving credit, I feel ill.
I haven't used Linux in quite a while since discovering FreeBSD, so please forgive any incorrect Linux info here.
Updating FreeBSD is done entirely with CVS. With one command, you can update the source for everything including all applications installed through the Ports collection. (Well, technically it updates the list of files to download, but it works the same)
In one command, "make buildworld", you can then recompile everything on the base system using custom compiler flags for a large increase in performance. In another command, "Make installworld", it all installs and replaces the old stuff.
Updating an application installed via ports is as simple as:
"cd/usr/ports/catagory/NameOfApp"
"make deinstall"
"make reinstall"
While the updating system is nice, I personally feel that the biggest benefit of the system is that *everything* can be compiled easily with custom compiler flags. Running a system in which nearly everything was compiled for a 386 has always irritated me, because it was holding back the performance of the system. It doesn't matter for many programs (like vi, for example) but it surely does with many other applications. Who has the time or inclination to recompile everything manually?
That said, it does have a disadvantage. The kernel and the base system are very closely tied together. So closely, that if many of the base system's core files are improved in a CVS update, you not only "can" compile the full base system as well as the kernel--you *must* compile the whole thing. This takes a fairly long time. Fortunately, there are ways to speed this up immensely, but it still takes about 45 minutes even on a dual-AthlonMP with a 10,000RPM SCSI hard drive.
I know of many people that swore by apt-get and switched to FreeBSD after trying it--and I also know of many that have tried FreeBSD and went backto Linux. Usually because Linux is easier to use as a pretty graphical system since everything comes preinstalled on many distros.
...however... *My* Xfree86 and GNOME are compiled such that they are far faster than they would be on, say, Redhat.;-) You can do the same. Give it a try. If you don't like it--no problem! You aren't using Windows either way.
Tron was one of the first movies to use computer graphics.
/. is a computer community. See the connection?
Besides, I thought that it had a very interesting plot.
Actually, unless it is Tom himhelft doing the review, Tom's Hardware is widely known for fairly bad reviews. Take a look at:
http://www.tomshardware.com/mainboard/01q4/01112 6/ index.html
...which says that the KT266a chipset "blows away" the NForce, when the performance differences are right around 2-3%.
Overall, I still like the site and most pf the reviews, though.
Amen, brother!
Seriously, though. Babylon 5 DVDs are takingentirely too long to be released. You can already get the excellent soundtracks, but those just aren't the same.
Soundtracks have no plots.
I have a friend that purchased a Super-VHS VCR for the sole purpose of recording B5. If people are willing to do that, I am sure many more would happily fork out $100/season for B5 DVDs.
All in all, it's a pleasant package, well worth $799 for the hardware alone. And OEone claims this is not a "loss leader" price. I pointedly asked the company's finance v.p., Eddie Vlasblom, about this, and he assured me there was a "healthy margin" on each unit sold.
A healthy margin?! Hardware should be free and under the GPL!
Slashdot Mirror
Yes. I was kidding.
Windows 2000 seems to work pretty well. If anything, IIS needs more of Microsoft's attention. At least in the security department.
I actually have seen a few programs in which there were more compiler errors than lines, but they were, of course, quite short.
Interesting reaction. This was a joke--it is fairly obvious that you can't have 1 bug per line of code, or at least I thought it was.
BTW, I don't use Linux, so I guess that I am disqualified as a Linux zealot. Sheesh. Calm down.
Being conservative, there is about 1 bug per line of code in, say, Windows 2000.
Windows 2000/IIS has about 2,000,000 lines of code.
That gives Microsoft about 0.0432 seconds per line of code, if they stop and work on only Windows 2000/IIS.
I wish them luck.
Of course, rather than pirating such software, one should write their own or use free equivalents.
Software piracy really is bad and harmful, but as will all things there are exceptions:
Winzip:
How much should one person make off of a frontend for someone elses compression algorhythm?
Microsoftware:
Why not just make a donation to the Al Qaeda while you're at it?
or Incredibly trivial software like a cheesy password generator, or (OMG!) a random number generator that cost $15.00
Works fine. It's a bit slow, taking about 5 hours to compile the kernel. And it has that weird MCA bus...
I'd hate to be at the top of that elevator if the cable breaks...
>
Ahh yes, and here's where we have our fully redundant Linux cluster with an uptime of 2 years. If anything does ever go wrong, we have a Lego Mindstorms backup system in place. I think we can trust that the database will never be lost.
And here we have...
When you choose to play with bleeding edge software, you have to live with the caveats that come with that usage.
Production environments? That's what FreeBSD is for. >
Seriously though (and I was serious), most companies with experiences IT people aren't going to use anything that hasn't been thoroughly tested internally (like the latest releases of the kernel that have the new VM), but then most companies probably wouldn't use MySQL in a production environment either...though I would argue that it is more solid than most operating systems.
LOL, good argument. In my experience, C can do everything except certain very specific platform dependant things that ONLY assembly can do. What, really, can literally not be done in C that can be done in any other HLL?
Certainly, things can be done more easily, or faster, or more quickly, but as far as a specific programming task that C simply cannot do--I have yet to see one.
1) The better organization (and more sensical naming) are part of what make Java easier to learn, giving the programmer more time to master.
2) How many Joe Programmers do you know that can write their own garbage collector? What if that custom garbage collector has horrible bugs? When you need to write a program quickly and its stability and security are paramount, Java all but eliminates these worries. As far as Lisp and others having garbage collection--yes, that is true, but few programmers know Lisp (other than tweaking a little EMACS). Why is NT so popular for networking? For its merits? (HA!). It's much easier to find an MCSE than a Unix admin--and cheaper. The same goes with Java VS. Lisp, Smalltalk, or even C++.
3) This often isn't that big of a deal, but it is great for the 10% of the time that it is a big deal. As far as browser incompatibilities, sticking with basic Java 1.1 or 1.2 tends to be very safe, but I haven't delved much into that realm.
4) I wouldn't really mention that part if I were asked about the merits of Java, but to Java's credit its GUI libraries are part of the language spec. GTK+ and QT are definitely not part of the C spec, so it is not reasonable to expect a platform that can run C apps (all of them) will be able to run C apps that call GTK+ or QT functions. If you make a button in Java, that button will work simply by virtue of the fact that the platform runs Java. At least in theory...
5) Me too.
6) I know little of Javabeans.
Hope this helps a little.
Note that despite the above, I still prefer C or PERL for almost everything.
The Ultra-SPARC chips are faster at the same clockspeed period. They are server class 64-bit chips with integrated PCI controllers running on an OS whose core parts were largely written in optimized assembly language. You'll notice that Java runs faster on an AthlonXP than A Pentium 4 as well--not because the XP implimentation is better (in fact, it will probably be the exact same), but because the XP is a faster chip/Hz.
That said, Sun may have spent extra time optimizing the JVM for the SPARC... It makes sense for a hardware company to do so. I'd love to run my java apps on a Sun, but any decent maching startng at around $30,000 turns me off.
You would...Love Java, but poiters scare you?
Be brave, young Incon, for Java has no pointers!
FreeBSD has long been one of the top performers--if not THE top performer--in TCP/IP. How they made a substantial improvement to code that has been maturing longer than the networking code for any other OS in common use today is amazing.
Can you imagine how skilled and bored the developers much be?
Unfortunately,it will take about 2 weeks for my 4.5CD to arrive, but all good things are worth the wait.
It seems you are correct. Oops, my mistake. I misread the headline. I guess it's time to get some sleep.
Good. I find those movies incredibly irritating. I suppose that it is bad that those who enjoy that sort of humor may not be able to, but I believe that this loss will prevent another decline in the average movie viewer's intellect.
I probably shouldn't reply to this...
I really like FreeBSD and X, both of which have BSD style licenses.
However, every time I think of Microsoft getting the benefit of the FreeBSD networking code that took years to polish, and not even giving credit, I feel ill.
Sorry, but it sounds like you are an intelligent and fairly technically competant person. You would never fit in with the other politicians.
I haven't used Linux in quite a while since discovering FreeBSD, so please forgive any incorrect Linux info here. /usr/ports/catagory/NameOfApp"
;-) You can do the same. Give it a try. If you don't like it--no problem! You aren't using Windows either way.
Updating FreeBSD is done entirely with CVS. With one command, you can update the source for everything including all applications installed through the Ports collection. (Well, technically it updates the list of files to download, but it works the same)
In one command, "make buildworld", you can then recompile everything on the base system using custom compiler flags for a large increase in performance. In another command, "Make installworld", it all installs and replaces the old stuff.
Updating an application installed via ports is as simple as:
"cd
"make deinstall"
"make reinstall"
While the updating system is nice, I personally feel that the biggest benefit of the system is that *everything* can be compiled easily with custom compiler flags. Running a system in which nearly everything was compiled for a 386 has always irritated me, because it was holding back the performance of the system. It doesn't matter for many programs (like vi, for example) but it surely does with many other applications. Who has the time or inclination to recompile everything manually?
That said, it does have a disadvantage. The kernel and the base system are very closely tied together. So closely, that if many of the base system's core files are improved in a CVS update, you not only "can" compile the full base system as well as the kernel--you *must* compile the whole thing. This takes a fairly long time. Fortunately, there are ways to speed this up immensely, but it still takes about 45 minutes even on a dual-AthlonMP with a 10,000RPM SCSI hard drive.
I know of many people that swore by apt-get and switched to FreeBSD after trying it--and I also know of many that have tried FreeBSD and went backto Linux. Usually because Linux is easier to use as a pretty graphical system since everything comes preinstalled on many distros.
...however... *My* Xfree86 and GNOME are compiled such that they are far faster than they would be on, say, Redhat.
Tron was one of the first movies to use computer graphics.
/. is a computer community. See the connection?
Besides, I thought that it had a very interesting plot.
Bruce Boxleitner was also the second captain on Babylon 5. Now if only they'd release that series on DVD...
Actually, unless it is Tom himhelft doing the review, Tom's Hardware is widely known for fairly bad reviews. Take a look at:2 6/ index.html
http://www.tomshardware.com/mainboard/01q4/0111
...which says that the KT266a chipset "blows away" the NForce, when the performance differences are right around 2-3%.
Overall, I still like the site and most pf the reviews, though.
Amen, brother!
Seriously, though. Babylon 5 DVDs are takingentirely too long to be released. You can already get the excellent soundtracks, but those just aren't the same.
Soundtracks have no plots.
I have a friend that purchased a Super-VHS VCR for the sole purpose of recording B5. If people are willing to do that, I am sure many more would happily fork out $100/season for B5 DVDs.
All in all, it's a pleasant package, well worth $799 for the hardware alone. And OEone claims this is not a "loss leader" price. I pointedly asked the company's finance v.p., Eddie Vlasblom, about this, and he assured me there was a "healthy margin" on each unit sold.
A healthy margin?! Hardware should be free and under the GPL!