Standards are always good. By having a consistant look and feel people can learn applicatins quickly. When I get a new program, I don't want to have to learn it, I want to know how I can use it to get things done. There is no reason for an office app to have a different look and feel from another office app. The good thing about windows is that one you learn one, you've learned them all. I can open up a complex program like Premiere and know how things are done. I know that under most new Windows programs a right click on an object will give me info relevent to that object. Under linux you can never be sure how a programer will do things. Take Moonlight Creator. It has an annoying "once you are in a mode you have to quit out before switching to another mode" Since no program I've ever used does that, I took me a while to get the hang of it. By having options, commands,etc in standerdized places all applications are easy to learn.
Actually you were trolling implying that people who use a GUI do not want to think. Secondly for many tasks a GUI can be just as fast and more efficient than CLI. Take a program. In Visual studio, I can add resources dependancies in a dialog box, and then every time I hit the hotkey to compile, it automatically links in the librarys I want. I compile many times in one program, and after the first time, I am saved the time needed to type out the command(and in windows there are a lot of.lib files to link in). Second, not everyone uses ONE app all the time. I switch between many programs (3D, webpage editing, avi editing, photoshop, etc) and for those people having a consistant interface is nice and lets you concentrate on your work instead of which command style the author of the program used. Humans are inherently visual and I don't know about you, but the only reason that I never use the Console mode is so I can launch my 1 xterm while looking at all the pretty KDE icons.
Actually yes I do. Prevasive multithreading is NOT only advantagous for SMP machines. the PII has 3 integer units and multithreaded programs allow them to be use more effiently. Second, Be has test where BeOS pushes 600 frames of 2D per second using a dual PII machine and BDirectWindow. It is much lighter weight and the UI is much more responsive than KDE Windows et al. And the APIs are very tightly integrated with each other,and they are even writing wrappers to make standerdized stuff more BeOS like. The API is completly C++ and a dream to work with. As for stability, 4.5 has never crashed on me, Linux has. And the crash took my filesystem with it. I was probably doing something wrong, but the fact remains that it happened. And if you don't belive the Be is a message passing OS read the developers section on their website. Lays all the APIs out nice and pretty.
As I remeber Windows DOES have memory protection. BeOS is a message passing OS, it beat Linux in speed, ease of programming and is just as stable. Message passing is NOT a bad thing, and it is actually a nice way to do an OS.
As I remember, most Linux drivers are not only in kernel space but compiled into the kernel. At least other systems load them outside kernel space (QNX)Besides, isn't the kernel space graphics in NT the reason it kicks Linux's ass in 3D? A proparly implemented driver in kernel space is just as stable and a lot faster than a user space driver.
Transmeta is not out yet, (in spec even). Wait, it would be just like Amiga to depend on them wouldn't it. I can see it now Amiga Press release Circa April 9 2000 "The AOE has recently switched its kernel to NT and will now be running nativly on the as of yet unspeced transmeta cpu. They still don't have an instruction set to show us but I firmly belive that it AOE will be out within six months."
Yea, but would you want your desktop environment running on top of an emulation layer? That would make it what GNOME on top of GTK on top of GDK, on top of Photon X emulation on top of Photon.
They would have to solder KDE based on QT and X onto a small slim window system, (photon) I think photon would tip over from the weight of it! Really though, Photon supports 3D, multimedia, and is really small, and no one wants a port of X to ANOTHER platform.
I think the Unix minimalism has gotten to the point where you need to use more than one app to do things. That leads to interdependancy which is bad (before you say it is not try installing X and KDE and compiling an X program on a computer that did not have X on it before. At one point I think I had to install Bison.) And one should not HAVE to use multiple programs to do tasks. People want to learn one program and have it suit their needs. However, Office apps can be bloated, I don't mind. When it extends to the OS + GUI, (ahhemm KDE, X) that is bad.
This test is perfectly fair. The point was not to compare Unix versions, but exactly what you said, which 15,000 box was best. Doesn't matter if it is a Power3 running AIX or a SPARC running Solaris, if it is easier to administer it is easier to administer, if it runs the benchmark application faster, then it is faster. No matter what the Workstation, the common ground is that they compete in the under 15K field, and their worth in that catagory can be measuered.
I have seen very few PII 400+es that do not have a plastic duct right over the cpu. And most graphics cards today have heatsinks +fans (Voodoo 3 3000 has a heatsink bigger than a low end PII one)
ha ha funny, if you came up with a good idea, would YOU want to share it and let someone else get the credit for the discovery. When people discover something, the news says one name. Sure you spent 70 years developing the technology to the 99% mark, but who ever gets that 1% at the end is the one who will be on the news. Open is not human nature.
can't use self selected reports he says
on
EDA: Unix vs. NT
·
· Score: 1
That is totally counter to what the Linux people say. All Linux has IS self selected reports. This was a real poll, which I think is a cut above the "I run NT at school and we tried to run quake on it while serving the network and it crashed. We have had the machine 2 days so that must mean NT crashes every 2 days!" which I see all over/. about linux.
A lot of games use a dedicated engine for their particular purpose. Take tribes, it has a custom hybrid engine that lets it do indoors and outdoors. However, games that cling to old engines, MSFS and Combat flight sim for example, are severly restricted in the kind of game they can do. An evolving design might work for an OS, but probably won't for an engine. I have nothing against people trying, but say the open source engine does catch on, developers will not innovate in game design, because the existing engine can't handle it. (Existing engines never seem to be able to handle the breakneck speed of the market) Then a developer takes advantage of this and develops a custom engine that blows away the old one, and makes huge amounts of money lisancing it. People abandon the old OSS engine and flock to the new one. (A free engine does no good if games don't sell) I just can't find a model where this would work.
It is interesting that anyone thinks this will be a good thing. Look at Bethesda Software. They kept using the same engine from Daggerfall to Redguard, and most of their games showed the age, along with a 1998 games that used the Descent engine! Good engine code can do a lot for a game, and giving it baggage (ala Unix) is just silly.
Do you think any game developer will open up Engine code? id makes a little bit for the quakes themselves and make huge amounts lisancing the engine out. I think the going price for Quake II was a few million, Lith-tech was a couple of hundred thousand and Unreal was also a million or so. The game itself is nothing compared to the money they make on the engine. Plus if the engine is opened up, people will get lazy and not write new ones. Carmack has motive to write Quake III becuase dozens of games will use the engine. If it is open source, why not just use an existing engine?
That is the only way to put it. The iMac lookalike is a much superior machine at about the same price. If you haden't noticed the G3 FPU performs at the level of a 233 MHz Celeron (Without Cache)in a 333MHz G3.
No I am not a Mac bigot, I can't wait for the G4, but the G3 is worse then x86.
Games are a big market consideration. You wouldnt think of it at first, but they do a lot to catapult a platform. And programming is not the only real app. You have 3D rendering, Photoshopping, Simulation, animation, etc and X does not cut it for that.
Not every on uses the console. I have my linux machine boot right up to KDE. When X dies, so does most of my work. You can say what you want, but editing documents, programming, drawing, is much easier under a GUI than console. With the GUI you can do a bunch of things at the same time without swaping between consoles to see them.
The reason that X is bad is not becuase of its features, but becuase of it's kludginess. The reason they guy said that Linux is for servers is becuase X is causing horrible problems with its bloat in local applications. Look at how people had to come up with LibGGI and DRI to bypass some of the X bloat. Most people don't need network transparancy, they don't want a windowing system that hurts local performance by using TCP/IP locally. A clean, light local windowing/graphics system is what Linux needs to penetrate the home/workstation market and thats what that person was lamenting about. And don't even think of saying that X is not bloated, it responds slower than Win2k on my computer (PII 412 64MB)
BTW. How is Solaris x86? I'm think about trying out their $10 home user thingy and installing GCC on it. anybody have that config and how would it run compared to say Linux on a PII 400 with 64MB
Controlled development is a very good thing! For a core thing such as an OS anyway. Its incredible that people say that this could be a bad thing. MS software is not bloated because it is controled, it is because they don't care. Plus they have such a wide group of people working on it that bloat is unavoidable. Look at BeOS, see the incredible cleanliness and standardness of the system. Look at the 30 person dev team. Look at Linux, look at the bloat of KDE and GNOME, look at the general hodge podge of APIs and the kludge that is X. Look at all the different groups who developed it. Controlled development is a good thing for things that need efficiancy. Humans would be much more efficiant if they had been developed in a controlled environment instead of the chaos of Darwin.
The cool thing about NT is that you CAN reinstall every month or two without problem. The reason is that with NT, most hardware is detected or their is a standard method of installing drivers. I have installed Linux atleast 7 or 8 times on different configurations, and NT is always easier. Think, with Linux, go through an automated install, then screw with pnpdump >isapnp.conf, then set up isapnp and have kerneld load the sound module at startup!!! Thats just to get sound working. Then screw around some more to get your modem. Then install Mesa, ALSA, and Libggi. Go download latest version of KDE or GNOME, compile and install those. Under NT I have a checklist of things to do and reinstalls (including apps) take less than 4 hours. An install of Linux takes at a minimum 8 to 12. Lastly, using windows I can change my host name with a few clicks, under Linux I have to figure out which text file to edit. Linux just isn't intuitive for that kind of thing.
Most libraries I have been to have the computers out in the open. How can someone in their right mind possibly look at porn in a library? Its not like librarians (or anyone else walking by the computer) are blind. I think this is just a "problem" that E. Dole is coming up with the work the voters.
In reply to the person who said you need to know your com port for windows, you don't. Windows correctly support plug and play. In fact in windows you don't need to know the irq, dma, io, etc. of your sound card. It doesn't matter if you can or not, why should you have to? Doing something harder does not make you anymore of a man it makes you stupid. And no you can't do image editing in GIMP. In my experiance it doesn't stand up to Satori or Photoshop. Last time I checked POVRAY read text instructions. I have yet to see decent 3d renderer for Linux, I haven't seen a good Web page editor (like hotdog, homesite, or Dreamweaver) for linux. Before you say that Linux already does everything better than windows, think a little.
Slashdot Mirror
Standards are always good. By having a consistant look and feel people can learn applicatins quickly. When I get a new program, I don't want to have to learn it, I want to know how I can use it to get things done. There is no reason for an office app to have a different look and feel from another office app. The good thing about windows is that one you learn one, you've learned them all. I can open up a complex program like Premiere and know how things are done. I know that under most new Windows programs a right click on an object will give me info relevent to that object. Under linux you can never be sure how a programer will do things. Take Moonlight Creator. It has an annoying "once you are in a mode you have to quit out before switching to another mode" Since no program I've ever used does that, I took me a while to get the hang of it. By having options, commands,etc in standerdized places all applications are easy to learn.
Actually you were trolling implying that people who use a GUI do not want to think. Secondly for many tasks a GUI can be just as fast and more efficient than CLI. Take a program. In Visual studio, I can add resources dependancies in a dialog box, and then every time I hit the hotkey to compile, it automatically links in the librarys I want. I compile many times in one program, and after the first time, I am saved the time needed to type out the command(and in windows there are a lot of .lib files to link in).
Second, not everyone uses ONE app all the time. I switch between many programs (3D, webpage editing, avi editing, photoshop, etc) and for those people having a consistant interface is nice and lets you concentrate on your work instead of which command style the author of the program used. Humans are inherently visual and I don't know about you, but the only reason that I never use the Console mode is so I can launch my 1 xterm while looking at all the pretty KDE icons.
Actually yes I do. Prevasive multithreading is NOT only advantagous for SMP machines. the PII has 3 integer units and multithreaded programs allow them to be use more effiently. Second, Be has test where BeOS pushes 600 frames of 2D per second using a dual PII machine and BDirectWindow.
It is much lighter weight and the UI is much more responsive than KDE Windows et al. And the APIs are very tightly integrated with each other,and they are even writing wrappers to make standerdized stuff more BeOS like. The API is completly C++ and a dream to work with. As for stability, 4.5 has never crashed on me, Linux has. And the crash took my filesystem with it. I was probably doing something wrong, but the fact remains that it happened. And if you don't belive the Be is a message passing OS read the developers section on their website. Lays all the APIs out nice and pretty.
As I remeber Windows DOES have memory protection.
BeOS is a message passing OS, it beat Linux in speed, ease of programming and is just as stable. Message passing is NOT a bad thing, and it is actually a nice way to do an OS.
As I remember, most Linux drivers are not only in kernel space but compiled into the kernel. At least other systems load them outside kernel space (QNX)Besides, isn't the kernel space graphics in NT the reason it kicks Linux's ass in 3D? A proparly implemented driver in kernel space is just as stable and a lot faster than a user space driver.
Transmeta is not out yet, (in spec even).
Wait, it would be just like Amiga to depend on them wouldn't it. I can see it now Amiga Press release Circa April 9 2000
"The AOE has recently switched its kernel to NT and will now be running nativly on the as of yet unspeced transmeta cpu. They still don't have an instruction set to show us but I firmly belive that it AOE will be out within six months."
Yea, but would you want your desktop environment running on top of an emulation layer?
That would make it what
GNOME on top of GTK on top of GDK, on top of Photon X emulation on top of Photon.
They would have to solder KDE based on QT and X onto a small slim window system, (photon) I think photon would tip over from the weight of it! Really though, Photon supports 3D, multimedia, and is really small, and no one wants a port of X to ANOTHER platform.
I think the Unix minimalism has gotten to the point where you need to use more than one app to do things. That leads to interdependancy which is bad (before you say it is not try installing X and KDE and compiling an X program on a computer that did not have X on it before. At one point I think I had to install Bison.) And one should not HAVE to use multiple programs to do tasks. People want to learn one program and have it suit their needs. However, Office apps can be bloated, I don't mind. When it extends to the OS + GUI, (ahhemm KDE, X) that is bad.
This test is perfectly fair. The point was not to compare Unix versions, but exactly what you said, which 15,000 box was best. Doesn't matter if it is a Power3 running AIX or a SPARC running Solaris, if it is easier to administer it is easier to administer, if it runs the benchmark application faster, then it is faster. No matter what the Workstation, the common ground is that they compete in the under 15K field, and their worth in that catagory can be measuered.
I have seen very few PII 400+es that do not have a plastic duct right over the cpu. And most graphics cards today have heatsinks +fans (Voodoo 3 3000 has a heatsink bigger than a low end PII one)
ha ha funny, if you came up with a good idea, would YOU want to share it and let someone else get the credit for the discovery. When people discover something, the news says one name. Sure you spent 70 years developing the technology to the 99% mark, but who ever gets that 1% at the end is the one who will be on the news. Open is not human nature.
That is totally counter to what the Linux people say. All Linux has IS self selected reports. This was a real poll, which I think is a cut above the "I run NT at school and we tried to run quake on it while serving the network and it crashed. We have had the machine 2 days so that must mean NT crashes every 2 days!" which I see all over /. about linux.
A lot of games use a dedicated engine for their particular purpose. Take tribes, it has a custom hybrid engine that lets it do indoors and outdoors. However, games that cling to old engines, MSFS and Combat flight sim for example, are severly restricted in the kind of game they can do. An evolving design might work for an OS, but probably won't for an engine. I have nothing against people trying, but say the open source engine does catch on, developers will not innovate in game design, because the existing engine can't handle it. (Existing engines never seem to be able to handle the breakneck speed of the market) Then a developer takes advantage of this and develops a custom engine that blows away the old one, and makes huge amounts of money lisancing it. People abandon the old OSS engine and flock to the new one. (A free engine does no good if games don't sell) I just can't find a model where this would work.
It is interesting that anyone thinks this will be a good thing. Look at Bethesda Software. They kept using the same engine from Daggerfall to Redguard, and most of their games showed the age, along with a 1998 games that used the Descent engine! Good engine code can do a lot for a game, and giving it baggage (ala Unix) is just silly.
Do you think any game developer will open up Engine code? id makes a little bit for the quakes themselves and make huge amounts lisancing the engine out. I think the going price for Quake II was a few million, Lith-tech was a couple of hundred thousand and Unreal was also a million or so. The game itself is nothing compared to the money they make on the engine. Plus if the engine is opened up, people will get lazy and not write new ones. Carmack has motive to write Quake III becuase dozens of games will use the engine. If it is open source, why not just use an existing engine?
That is the only way to put it. The iMac lookalike is a much superior machine at about the same price. If you haden't noticed the G3 FPU performs at the level of a 233 MHz Celeron (Without Cache)in a 333MHz G3.
No I am not a Mac bigot, I can't wait for the G4, but the G3 is worse then x86.
Games are a big market consideration. You wouldnt think of it at first, but they do a lot to catapult a platform. And programming is not the only real app. You have 3D rendering, Photoshopping, Simulation, animation, etc and X does not cut it for that.
Not every on uses the console. I have my linux machine boot right up to KDE. When X dies, so does most of my work. You can say what you want, but editing documents, programming, drawing, is much easier under a GUI than console. With the GUI you can do a bunch of things at the same time without swaping between consoles to see them.
The reason that X is bad is not becuase of its features, but becuase of it's kludginess. The reason they guy said that Linux is for servers is becuase X is causing horrible problems with its bloat in local applications. Look at how people had to come up with LibGGI and DRI to bypass some of the X bloat. Most people don't need network transparancy, they don't want a windowing system that hurts local performance by using TCP/IP locally. A clean, light local windowing/graphics system is what Linux needs to penetrate the home/workstation market and thats what that person was lamenting about. And don't even think of saying that X is not bloated, it responds slower than Win2k on my computer (PII 412 64MB)
BTW. How is Solaris x86? I'm think about trying out their $10 home user thingy and installing GCC on it. anybody have that config and how would it run compared to say Linux on a PII 400 with 64MB
Controlled development is a very good thing! For a core thing such as an OS anyway. Its incredible that people say that this could be a bad thing. MS software is not bloated because it is controled, it is because they don't care. Plus they have such a wide group of people working on it that bloat is unavoidable. Look at BeOS, see the incredible cleanliness and standardness of the system. Look at the 30 person dev team. Look at Linux, look at the bloat of KDE and GNOME, look at the general hodge podge of APIs and the kludge that is X. Look at all the different groups who developed it. Controlled development is a good thing for things that need efficiancy. Humans would be much more efficiant if they had been developed in a controlled environment instead of the chaos of Darwin.
The cool thing about NT is that you CAN reinstall every month or two without problem. The reason is that with NT, most hardware is detected or their is a standard method of installing drivers. I have installed Linux atleast 7 or 8 times on different configurations, and NT is always easier. Think, with Linux, go through an automated install, then screw with pnpdump >isapnp.conf, then set up isapnp and have kerneld load the sound module at startup!!! Thats just to get sound working. Then screw around some more to get your modem. Then install Mesa, ALSA, and Libggi. Go download latest version of KDE or GNOME, compile and install those. Under NT I have a checklist of things to do and reinstalls (including apps) take less than 4 hours. An install of Linux takes at a minimum 8 to 12. Lastly, using windows I can change my host name with a few clicks, under Linux I have to figure out which text file to edit. Linux just isn't intuitive for that kind of thing.
Most libraries I have been to have the computers out in the open. How can someone in their right mind possibly look at porn in a library? Its not like librarians (or anyone else walking by the computer) are blind. I think this is just a "problem" that E. Dole is coming up with the work the voters.
In reply to the person who said you need to know your com port for windows, you don't. Windows correctly support plug and play. In fact in windows you don't need to know the irq, dma, io, etc. of your sound card. It doesn't matter if you can or not, why should you have to? Doing something harder does not make you anymore of a man it makes you stupid. And no you can't do image editing in GIMP. In my experiance it doesn't stand up to Satori or Photoshop. Last time I checked POVRAY read text instructions. I have yet to see decent 3d renderer for Linux, I haven't seen a good Web page editor (like hotdog, homesite, or Dreamweaver) for linux. Before you say that Linux already does everything better than windows, think a little.