I don't play computer games, so I might be out of the loop on this one, but why does speed matter so much? What can a 1 GHz computer do for the average user that a 500 MHz computer can't? I'm sure there are lot's of answers like "faster SETI", "faster compiles", "faster ray-tracing", "quake @ 1600x1200" etc. But most people don't use their computer for that.
Most people just want to interact with the internet and create different kinds of documents. Their "power app" is playing DVDs. The only thing that keeps their CPU below 99.5% idle is "Clippy" dancing at the bottom of the screen.
What people really need, more than CPU power, is good, easy to use software. Apple tries to provide that. If you don't belive me about the importance of good software, look at the success of the Palm Pilot vs. the failure of Windows CE.
Keep in mind that OpenBSD isn't really very old. These classified sites have been around for a while and it seems unlikely that OpenBSD could become better than what they had before in a just a few years.
Fundumentally, OpenBSD is still UNIX. Remember the Orange Book codes? You don't see many UNIX systems past C2 because they really weren't designed for it. The A's and high B's belong to operating systems designed with security in mind from the start. Not insecure OSes that have been patched up.
Not to say OpenBSD isn't secure, it's just not at the right level of paranoia.
One of the problems with microkernels is that they can require a lot of context switches. There is hope though:
The ineffiencies of context switching will disappear as multi-threaded CPUs are developed.
This isn't a single system image. There are many separate copies of AIX, each controlling a small number of CPUs. Programs communicate between these systems using a technique called "Message Passing".
I worked, briefly, in a large company where Netware was pretty entrenched. Everyone loved it. It took twice as many NT servers and administrators to do the same thing that Netware did.
The admins said, "We'll never get rid of Netware! NT sucks." They were half right. Just before I left they started knocking out Netware boxes and installing NT everywhere.
You may be surprised to know that there is still at least one company that makes money by selling and maintaining old PDP-11s. In that sense, there will probably be a market for Netware for a long time to come. Many company's won't want to chuck their Netware boxes just because they are out of style. But how many new customers does Novell get each year? I don't think they will ever bounce back the way Apple famously did.
I don't have a good background in this so I have a question:
How can you detect these watermarks? When (if) SDMI is implemented will there be a way for users to check a file and see if it is "clean" or can only the person who created the watermark detect it?
In the early 90's F1 cars were jam packed with computers. Before the FIA banned it, the layout of the course was fed into a computer. The car had an active suspension that would lean into the corners, etc. Your mother could have been an F1 champion in one of those cars.
A computer should be able to get a better lap time than a human, but that's not the same as winning in F1. The technology of the cars has a lot to do with winning. (Look at what happened to Irvine when he left Ferrari for Ford!) Driving in traffic would probably be beyond the computer.
A question that I would have is: who would do better in an unusual situation? Say the car starts spinning, or something on the car isn't working 100%?
If Corel implements ".NET" on Linux it will be on Microsoft's terms. So you better get out the magnifying glass and the lawyer to read the licensing agreement.
It would be nice to see free support for Linux clients, but I hope someone (not me!) comes up with the server end for Linux. That would really tick Microsoft off.
I think it will be a small machanical human for the AIBO to play with. Only a niche group of AIBO's will want one though. The rest will just get real humans.
How about loadable virtual memory modules? It seems difficult, but what is the alternative?
A forked kernel will make it difficult to write new features for both. This will mean that people will just write for the kernel they want to use and ignore the other one. Of course code could be ported by someone, but it's kind of a duplication of effort.
If linux does fork it should be the 8MB 386 crowd that goes off and does their own thing. Computers aren't getting any less powerful, even embedded ones.
Re:When will the world accept that x86 is a deaden
on
Pentium IV Problems?
·
· Score: 1
You have to take into account the cost of re-programming for a new architecture. It's fine to throw out x86 if you have an option for people whose critical app (windows for most people) only works with Intel.
As for your request for a clean architecture, I wouldn't get my hopes up. Besides, your C compiler should hide all of the annoying parts from you. Unless you are in the buisness of writing C compliers, I don't see how you can even tell what kind of architecture you're running on. (Unless you notice that you FP performance sucks.)
Consider IBM's S/390 architecture which is much older than x86 and isn't showing any sign of being eliminated or rethought.
I agree. When I first heard about Eazel there was a lot of "revolutionary" talk. Now it turns out they're making YAFM. That's nice, but I don't see any evidence that this is leap forward. I'm glad they're around and making this software, but I was hoping for exciting new things that no other OS would have. It turns out we're just catching up to Windows.
quite a few of the same interface mistakes Microsoft made
I think one problem is that it's hard to agree on what a good interface is. I think one of the biggest problems that technical people have with Microsoft is that it only offers one iterface and claims that it is good for everyone. It seems to be aiming at users who are somewhere above a complete novice and somewhere below a professional. They end up alienating a lot of new users and it's really frustrating for UNIX people who are used to having a sophisticated interface to have to dumb down to Windows. I don't think everyone needs the same interface.
Hopefully Eazel's file manager is just the first step to a more user friendly UNIX. I personally think the the Mac is too complicated for most people, so UNIX is right out.
I think we're doing a good enough job of loosing the battles we're currently fighting without looking around for more unfair fights.
This is like a nest of hornets colluding to sting someone. You might get some pricks in, but if you piss the person off enough, they'll be coming after your nest.
I think we need to be fighting these battles in the courts so that we won't have to worry about complicated ways to avoid the laws. The courts are the ultimate authority, and no matter how dumb the laws are, they're still law.
There are lots of ways to be disruptive. It might be cool if programmers organized a strike to protest some of this. I don't think that annoying corporations with programs is a good winning strategy. It's more like a holding action.
Anyway, it's clear that the U.S. government isn't on our side, so we may want to use this upcoming election to send a message to them. I know that Ralph Nader of the Green Party has said positive things about Free software and is opposed to the kind of corporate strongarming that is going on right now. If you know of other candidates who are on our side, speak up and let everyone know about them. If you get a chance, ask your local candidate about the DMCA, DeCSS, and other important geek topics. Even if we can't remove the establishment from office, showing them that we're an angry, voting demographic will cause them to pay attention.
Please, take a few bucks out of this month's paycheck and send it to the EFF. They're fighting the hard legal fight for us.
RT is useful in subtle ways. For example, it would make burning a CDROM easier. You could say, well I'll just double the speed of my CPU and it will keep up. But then what if you decide to buy a 2x CDROM burner, now you have to push twice the data and you're back where you started.
To say that we will just have to wait for computers to get faster and then we won't need decent software is making the assumption that the things that people want to do with computers won't get any more complex or resource intensive in that time.
This is why companies like Microsoft can't win against free software. If Microsoft wanted a real-time version of NT they would have to put togeather a huge team and it would be a risky, complicated project. With Linux, somebody decides they want a pre-emptable kernel so they just do it; and everyone benefits. (Not to belittle the effort it took to make this kernel.)
Now other projects, like GNOME, are gaining a 'critical mass' of interest from developers and industry. I expect progress will be made by leaps and bounds.
It's a really handy feature to be able to log on to system that you're not in front of. System administrators know this, and they probably benefit the most, but the feature can be handy for users too.
The example from my work is compiling and simulation. If I have to perform some task that is too big for my workstation to efficiently handle, I can use a server that has much more memory. I don't need to walk over to the server room (I don't even know where it is, and I don't care), I don't need to have 2 or more computers crammed in my office. I just have to ssh to the server and it acts just like I'm there. I can use dozens of servers at once without having to get up from my chair. Since this is a CLI, I can even automate the whole thing. I'm sure that windows has a way to do something similar if you bought the right 3rd party software package, but I'm sure it would be less powerful than what comes with UNIX as default.
The other way this feature is handy is when I'm working on something with another person and I need them to access something for me. Let's say that a file of theirs isn't group readable and we need to use it. I don't have to logout, I don't have to wait while they run back and fix it, they just type su and fix the problem.
I can ssh from home to work and have an identical environment to the one that I have at work.
The last reason is kind of a UNIX thing, so it really doesn't apply to NT, but I bet NT would benefit from this if it's graphics weren't in the kernel. 90% of the time, if a UNIX system is frozen it is X and not the OS that is locked. Fixing this problem is possible with a telnet to the frozen machine. On NT you would just give up.
Windows NT security may be fine-grained, but it's CPU resource allocation is not. It basically gives the user in front of it the whole thing, whether they need it or not. NT's ability to switch between users seems to be more a limitation of it's GUI and registry rather than the kernel itself. I've used Terminal Server. I was impressed and disappointed at the same time. It is a step in the right direction, but it is hampered by the need to retain compatibility with poor decisions from the past.
Slashdot Mirror
more likely the "anti-aliased" look comes from the screenshots being scaled down.
Most people just want to interact with the internet and create different kinds of documents. Their "power app" is playing DVDs. The only thing that keeps their CPU below 99.5% idle is "Clippy" dancing at the bottom of the screen.
What people really need, more than CPU power, is good, easy to use software. Apple tries to provide that. If you don't belive me about the importance of good software, look at the success of the Palm Pilot vs. the failure of Windows CE.
Fundumentally, OpenBSD is still UNIX. Remember the Orange Book codes? You don't see many UNIX systems past C2 because they really weren't designed for it. The A's and high B's belong to operating systems designed with security in mind from the start. Not insecure OSes that have been patched up.
Not to say OpenBSD isn't secure, it's just not at the right level of paranoia.
We've recently been reminded about WINE too. If software doesn't come to Linux, Linux can always meet them halfway.
One of the problems with microkernels is that they can require a lot of context switches. There is hope though: The ineffiencies of context switching will disappear as multi-threaded CPUs are developed.
This isn't a single system image. There are many separate copies of AIX, each controlling a small number of CPUs. Programs communicate between these systems using a technique called "Message Passing".
The admins said, "We'll never get rid of Netware! NT sucks." They were half right. Just before I left they started knocking out Netware boxes and installing NT everywhere.
You may be surprised to know that there is still at least one company that makes money by selling and maintaining old PDP-11s. In that sense, there will probably be a market for Netware for a long time to come. Many company's won't want to chuck their Netware boxes just because they are out of style. But how many new customers does Novell get each year? I don't think they will ever bounce back the way Apple famously did.
Apparently some writer was having a slow week and got desperate for a story.
How can you detect these watermarks? When (if) SDMI is implemented will there be a way for users to check a file and see if it is "clean" or can only the person who created the watermark detect it?
A computer should be able to get a better lap time than a human, but that's not the same as winning in F1. The technology of the cars has a lot to do with winning. (Look at what happened to Irvine when he left Ferrari for Ford!) Driving in traffic would probably be beyond the computer.
A question that I would have is: who would do better in an unusual situation? Say the car starts spinning, or something on the car isn't working 100%?
Before he dies he'll ask for some water.
It would be nice to see free support for Linux clients, but I hope someone (not me!) comes up with the server end for Linux. That would really tick Microsoft off.
I think it will be a small machanical human for the AIBO to play with. Only a niche group of AIBO's will want one though. The rest will just get real humans.
A forked kernel will make it difficult to write new features for both. This will mean that people will just write for the kernel they want to use and ignore the other one. Of course code could be ported by someone, but it's kind of a duplication of effort.
If linux does fork it should be the 8MB 386 crowd that goes off and does their own thing. Computers aren't getting any less powerful, even embedded ones.
As for your request for a clean architecture, I wouldn't get my hopes up. Besides, your C compiler should hide all of the annoying parts from you. Unless you are in the buisness of writing C compliers, I don't see how you can even tell what kind of architecture you're running on. (Unless you notice that you FP performance sucks.)
Consider IBM's S/390 architecture which is much older than x86 and isn't showing any sign of being eliminated or rethought.
I don't know what you're whining about. This is America at it's finest. If we didn't have advertisements, how would we know what to buy? Geez.
quite a few of the same interface mistakes Microsoft made
I think one problem is that it's hard to agree on what a good interface is. I think one of the biggest problems that technical people have with Microsoft is that it only offers one iterface and claims that it is good for everyone. It seems to be aiming at users who are somewhere above a complete novice and somewhere below a professional. They end up alienating a lot of new users and it's really frustrating for UNIX people who are used to having a sophisticated interface to have to dumb down to Windows. I don't think everyone needs the same interface.
Hopefully Eazel's file manager is just the first step to a more user friendly UNIX. I personally think the the Mac is too complicated for most people, so UNIX is right out.
At the rate clock speeds are rising, you'll be able to get the speed you want soon, without overclocking.
This is like a nest of hornets colluding to sting someone. You might get some pricks in, but if you piss the person off enough, they'll be coming after your nest.
I think we need to be fighting these battles in the courts so that we won't have to worry about complicated ways to avoid the laws. The courts are the ultimate authority, and no matter how dumb the laws are, they're still law.
There are lots of ways to be disruptive. It might be cool if programmers organized a strike to protest some of this. I don't think that annoying corporations with programs is a good winning strategy. It's more like a holding action.
Anyway, it's clear that the U.S. government isn't on our side, so we may want to use this upcoming election to send a message to them. I know that Ralph Nader of the Green Party has said positive things about Free software and is opposed to the kind of corporate strongarming that is going on right now. If you know of other candidates who are on our side, speak up and let everyone know about them. If you get a chance, ask your local candidate about the DMCA, DeCSS, and other important geek topics. Even if we can't remove the establishment from office, showing them that we're an angry, voting demographic will cause them to pay attention.
Please, take a few bucks out of this month's paycheck and send it to the EFF. They're fighting the hard legal fight for us.
To say that we will just have to wait for computers to get faster and then we won't need decent software is making the assumption that the things that people want to do with computers won't get any more complex or resource intensive in that time.
Now other projects, like GNOME, are gaining a 'critical mass' of interest from developers and industry. I expect progress will be made by leaps and bounds.
The future looks bright.
I better go update my resume!
Can you imagine what kind of mess this thing would make if it collapsed?
to know that after all we've been through, we still have something to fight about.
The example from my work is compiling and simulation. If I have to perform some task that is too big for my workstation to efficiently handle, I can use a server that has much more memory. I don't need to walk over to the server room (I don't even know where it is, and I don't care), I don't need to have 2 or more computers crammed in my office. I just have to ssh to the server and it acts just like I'm there. I can use dozens of servers at once without having to get up from my chair. Since this is a CLI, I can even automate the whole thing. I'm sure that windows has a way to do something similar if you bought the right 3rd party software package, but I'm sure it would be less powerful than what comes with UNIX as default.
The other way this feature is handy is when I'm working on something with another person and I need them to access something for me. Let's say that a file of theirs isn't group readable and we need to use it. I don't have to logout, I don't have to wait while they run back and fix it, they just type su and fix the problem.
I can ssh from home to work and have an identical environment to the one that I have at work.
The last reason is kind of a UNIX thing, so it really doesn't apply to NT, but I bet NT would benefit from this if it's graphics weren't in the kernel. 90% of the time, if a UNIX system is frozen it is X and not the OS that is locked. Fixing this problem is possible with a telnet to the frozen machine. On NT you would just give up.
Windows NT security may be fine-grained, but it's CPU resource allocation is not. It basically gives the user in front of it the whole thing, whether they need it or not. NT's ability to switch between users seems to be more a limitation of it's GUI and registry rather than the kernel itself. I've used Terminal Server. I was impressed and disappointed at the same time. It is a step in the right direction, but it is hampered by the need to retain compatibility with poor decisions from the past.