That's a great arguement, except not really. Windows rarelly fucks up some install, and if it does for you, you're trying to hard. The truth is, that you run the.exe, it reboots, you get on with your life. If you run funky programs on funky hardware, that's your own fault. I have used Windows ever since 3.1, and have yet to encounter a problem that requires me to hack the registry. My little brother fucks with his computer doing everything from Pokemon to photshop and he has yet to cause a problem like this. (Although he does have 50 items on his desktop that he can't figure out how to get rid of.) Windows fucking up is mainly an urban legend these days. If your hardware and software is old enough that it does screw up, you're probably knowledgable enough to fix it. If it isn't then it probably isn't screwing up/
I love the fact that you assume I don't use Linux. As of now I am running Mandrake 7.1 (for those of you who have read my other posts, yes I DO change distros that often) on ReiserFS running on a beta kernel running the beta nVidia GL drivers. I also have copies of Slack 7 and Suse 6.4 that I'll try out when I have the time. That said, I also like your cemment "the only thing that requires a kernel compile is a new hardware driver..." What other kinds of drivers are there besides hardware drivers? Plus, you also say that is only the case if you don't have it in your base distro. That is like saying Windows drivers install themselves. Which is true, but only if they come on the Windows98 CDROM. The truth is, that most people will not want to use older drivers, and for hardware that changes drivers frequently (graphics cards, sound cards, video capture cards, monitors in my case) a standard, easy method of installing drivers is needed. Second, about using ipchains, I still stand by the fact that it makes Linux harder. Even people who know enough about the system will go to the Mandrake or Redhat config tool to do that, and not being able to find it, will ask their nearest Linux nerd. What will they say? ipchains of course! Even Linux.com, which caters to Linux newbies, uses ipchains in its tutorial on networking. Ideally, of course, you're right. Practically, I'm right.
This has nothing to do with BeOS as an operating system. I have in the past extolled the virtues of Windows when it has deserved it or even Linux (gasp!) when it deserved it. My point is that Eazel says it is making Linux easy to use, and I say it is not. I'm simply providing BeOS as an example of an easy to use system. Many other OSs would have sufficed, but that is the one I am most familier with.
Who cares about this Linux thing. It is just UNIX. It just rips of everybody's ideas. KDE is just a clone of Windows, ad naseum. How many times have you heard that before? However, they are all false, as are yor comments about MS. MS integrated these things into the OS. In effect they pioneered these things on home operating systems much like Linux is a pioneer in making UNIX an easy to use cohesive environment for everyone.
You don't have to click/see file details. When you select the file, it pops up in a little Window in the side. That's a moot point however. Did Kideo integrate OLE into the viewer so that any OLE application could be used to provide previews in that box?
Those were still separate products. Saying MS doesn't deserver credit for integrating them into the OS is like saying that nobody in the Linux movement deserves any credit for anything. Most of Linux has already been done before, albiet by different systems. Windows already does 99% of the things GNOME and KDE are trying to do, yet they are billed as being pretty innovative.
And X still doesn't have anti-aliasing?
on
Cleartype In Depth
·
· Score: 3
Yea, I know it is kind of mean to harp on a system that is older than I am, but I think it is ridiculous that X is so far behind in modern display technology. (It is also ridiculous that we are still using it after so many years, but that's another story.) Coming from the all anti-aliased all the time world of BeOS, I notice that X has really poor font handling. If you read any of the FreeType docs, you'll see constant bitching about how X really doesn't provide any font support worth a damn. While X just recently got integrated truetype support, and is still far away from having anti-aliasing (which even Windows! has) Microsoft is introducing new technology to make fonts look even better! Aside from these small scope issues, this fact points to larger problems with the X architecture. It just wasn't designed to be extended cleanly. As such, X (at XFree86) still doesn't have genuinly usefull things like DPS and anti-aliasing. This is yet another reason that X needs A) A huge, major rewrite. B) To be replaced with something else. Berlin looks very promising. True it uses CORBA for its API services, which sacrifices speed, but even at the beta level that it is in, it already has nifty features like anti-aliasing support, and a totally cool imaging model. Also, it is designed to be extended cleanly.
Exactly how is Natulis going to make Linux easier to use? All it seems to be is a file browser/desktop shell along the lines of gmc on steriods. Sure it may make navigating around the system easier, but what about the inherent difficulty present in doing anything non-trival in Linux? Take a good hard look at the Mac community for a clue. Mac users aren't brain-dead neophytes who can't tell a close button from a minimize button. A great many are people who know their way around the system well enough, and can get basic things done. Rare is the Mac user who doesn't know what extensions are or who can't get basic services working on their Mac. There is a reason for this. The Mac is easy to use. Linux is not. All the great GUIs in the world will not make Linux easy to use. Take something like installing an extension. Under MacOS, this is trivial. Under Windows, installing a driver is similarly simple. But under Linux it often requires a kernel recompile! Take the new nVidia kernel driver. I've have run quite a few beta drivers in Windows, but none have required me to manually remove OpenGL files from the path! Installing basic services on Linux is similarly hard. Internet connection sharing is becoming a big thing, and it is easy on everything except Linux. In Windows you go through the wizard to set it up. It asks basic questions and it works. In BeOS it is even more trivial. You go to the graphical NAT config, give it the IP of your internet interface and your LAN interface, click "NAT On" and restart the net server. In Linux, it requires learning ipchains and its complex syntax. Sure it ends up to be three simple lines of script, but A) It takes hours to get those three lines, and B) How is the user supposed to know where to put the script? Even simple things like changing host names or IP addresses, or adding new hardware become a chore in Linux. This need not be the case. A GUI CAN do something about this. Take a look at BeOS. It has most of the features of BSD networking, yet its configuration panel has something like 3 address fields and a checkbox. Telnet and ftp servers can literally be set up with a click and a password entry. It might not be as powerful, but it sure is easy. Yet, the power is still there. Navigate up to "/etc" and behold! Network settings! A lot of stuff in BeOS works this way, and it is pretty cool to behold. I have no doubt that Eazel will be a cool shell. It will make it easier to navigate around the system, and will lower the learning curve, which is always good. However, stuff like that is trivial. Ease of use should permeate the system, allowing the user to do more as their skills increase. It takes a very well designed system to allow some one with no skill to easily learn the system, yet not constrain those who have mastered those skills. Linux has a lot of cool stuff in it and has the potential to become an OS that is powerful for the hackers, yet easy for the newbie. The beauty of Linux is its multi-faceted-ness and malleability. Yet, the very people who are working on making Linux easier to use don't seem to have the right vision. They just don't get it.
I find it incredible that people accuse Microsoft of not innovating and just stealing other people's ideas. take a look at the second screen-shot in the directory. For some reason, the file brower looks hideosly familier. Oh, wait. Its IE in a GTK+ dress! It even has the little thing on the side that gives a summery of the file. If the Eazel developers are worth their salt, it will also use Bonobo to do realtime previews in that box, cementing its similarity to IE. KDE does this too, Knoqueror is both a web browser and file browser. The integration between the browser and the file viewer is cool (and very useful) tech, and MS is the one that poineered it. They might make crappy products, but give credit where credit is due.
The system may bootstrap using the DOS kernel, but the Windows kernel is certainly not a virtual kernel running on top of DOS. That was the case in versions prior to 3.1. Read the book about PCs by Peter Norten. It is quite helpful and describes certain things about the Windows kernel. In particular, it says that in Windows95, the functions previously handled by DOS were moved into modules within Windows95. Whether or not these modules contain DOS code is irrelevant. They are within Windows 95. If what you say is true, that Windows runs as a virtual kernel over DOS, then most tasks that require access to hardware would have to go through DOS. However, Win32 rarely, if ever, have to go into DOS mode. They only times that happens is when you have a funky program or driver that runs in real mode. What your saying doesn't make sense from the way the Windows arch is done. The lowest level is the huge Window kernel. Sure it loads stuff, but those are modules, and drivers (being modules as well) can be considered part of the kernel. Additionally, Win32 programs run on a system-wide virtual Win32 machine. Win16 programs run in a single Win16 virtual machine, and DOS programs run in multiple DOS virtual machines that load DOS before running the program. The major 16 bit DOS legacy that Windows has is mainly the legacy from Win 3.1. Major parts of some of the Windows kernel modules, like user, are 16 bit, which leads to instability. PS> Programs and modules loaded into the kernel count as part of the kernel.
Actually, Windows works differently than Linux. Then kernel is located in a special file (c:\io.sys) on your hard disk. >>>>>> Not in any Windows version is the kernel io.sys. It provides some DOS functionality, but the Windows kernel resides in kernel32.dll and kernel32.exe.
All of the extra hardware functionality (USB, multimoniter, etc.) that you mention is tacked into the OS through a series of interesting things like normal and virtual device drivers. >>> All the buses are in the kernel. That's why decent USB support didn't appear until 98 and why NT4.0 never got Firewire support until MS patched it. TheyUSB drivers required a lot of hacking to get past the kernel, and the Firewire drivers were impossible to write without the cooperation fo the NT kernel. Multi monitor is also part of the kernel because that is under the control of the graphics system, which resides in the kernel.
It's sort of like a microkernel, only these drivers are accessed through a GUI (Windows) running on top of the DOS 7 part and the kernel. >>>> Wrong again. Contrary to popular belief, Win95 does not run top of DOS. Its mostly anti-microsoft propganda. True, Windows 95 has DOS embedded in it, but does not use it all that much when running Windows programs. If you run fully 32 bit programs, Win95 rarely goes switchs into real mode DOS. WinME will finally take DOS out altogether, although it will still suck. I mean even Win3.1 only used DOS for the file system! Second, no Windows is really a microkernel. They want to tell you that NT is a microkernel, but in reality, it has most drivers embeeded in the kernel, and one big Win32 system server. Hell, in Windows 2000, the hardware abstraction layer includes calls to DirectX! (Though there is nothing wrong with that. It might be acedemically incorrect, but if MS would just let the DirectX guys do NT, Linux would be in major trouble.) Win9x is even more monolithic. Everything from the graphics and some GUI functions to file systems run in the kernel.
Unlike with a microkernel, the kernel never actually touches the device drivers and things that Windows runs. Windows even has it's own virtual kernel that runs on top of the actual one! >>> You're confused. The Windows kernel has complete access to hardware drivers. It doesn't run on top of DOS, it uses some DOS code in the kernel. Its just like the Linux kernel in terms of closeness to harware, but while Linux is completely 32 bit protected mode, Win9x has some sections that are real mode. Also, there is no virtual kernel. I think what confused you is that Win9x has a virtual machine that runs all Win16 programs, and many virtual machines to run DOS programs. All 32 bit programs run without a virtual machine.
Some of the other things you mention (icons, IE) are actually in executable code in the GUI part of Windows and elsewhere, not through any interaction with anything resembling the kernel. >>>>>>>>>>>>>> True, icons are not in the kernel. Faux pas on my part. However, they are pretty close. All the routines to load icons and do graphics are in the kernel. Like I said, Win9x is SERIOUSLY monolithic.
X-Windows isn't part of the Linux kernel, and IE isn't part of the Windows/DOS 7 one. >>>>>>>>>> Stop sayiing Windows/DOS 7. There is no DOS 7 kernel in Windows, the virtual machine that runs on top of the Win9x kernel reports itself as DOS 7. Win32 programs never actually use that virtual machine. The DOS heritage that Win9x has is not that it runs on top of it, but that it uses a good deal of DOS code.
DirectX is weird. It is made up mostly of a large number of device drivers and some executable code, although there are more complicated things in there. >>>>>> DirectX is god. DirectX is mostly device drivers, that's true, that's what gives it the speed. Conceptually, DirectX is a set of COM objects that talk to the DirectX HAL/HEL. The hardware drivers make up the HAL (hardware abstraction layer) and emulators make up the HEL (hardware emulation layer) the other executable code is the stuff that orchastrates to whole thing.
These security fixes mostly update DLLs and stuff NEVER the kernel. Again, most of Windows is actually executables and libraries. The kernel is quite small and doesn't do a whole lot except interpret for these executables. >>>>> Whoa, that's kind of wrong. Windows networking is implemented in the kernel on Win9x, and in kernel mode servers on NT. True, they might not be in the same executable, but they are for all purposes part of the kernel. Again, I don't think you quite have the right idea about Windows. In all versions of Windows, the kernel is quite large. (Again, in NT the kernel32.dll is not that big, but a lot of stuff runs that is loading into the kernel.)
In short, the actual kernel has not changed much, other than moving it from two files (msdos.sys used to have part of it) into just io.sys. >>>> I'm assuming you're talking about Win9x here. The kernel is not io.sys or msdos.sys on any version of Windows. They are DOS modules loaded by the Win32 kernel to facilliate some operations.
USB and other nice things never directly interact with the kernel, but work through executable code (win.com and associated dlls and other files) that runs on top of the kernel and accesses hardware. >>>>> Wrong again. USB and stuff does work through the kernel. DLLs may provide support for the actual device, but USB is a bus and busses in Win9x are supported by the kernel.
Linux, on the other hand, integrates USB and such into the kernel, so it does not constantly crash because of the complex and unstable patchwork doing things DOS was never meant to do. >>>>> You're sentences make no sense. Integrating something into the kernel make it less stable, not more. (Ever wonder why NT used to be really really stable in 3.x before they moved graphics into the kernel?) DOS has nothing to do with it. There is no code in the USB services that use DOS.
You have a very well thought out response. The problem is that you are doing exactly what MS wants: seeing Windows as one big happy family rather than the confused mish mash it is. Dig a little deeper, and you'll discover why the model is insecure and why it crashes constantly. >>>>> You seem to be quite confused on what Windows is. I got all of my information from a BYTE article cirica launch of Win95 that detailed the architecture of Win95 (then Chicago.) (BTW. BYTE was THE nerds magazine. I've never since seen a mainstream mag that went into the kind of technical detail BYTE did. Read the one from 1993 about the new OSs that were coming out then. They talked about message passing and hardware abstraction layers like you had known about them forever!) True, Win9x is a mishmash, but the fact that it runs on top of DOS is just not true. Win3.1 did, but in Win95, everything was moved into a set of modules (such as USER32.exe and USER32.dll and GDI32.exe and GDI32.dll) which comprised the kernel. Some of those modules contained DOS code (at launch, GDI32.exe was largely 16bit Win3.1 code,) but that does not mean that Win9x runs on top of DOS.
Do you know how they managed the effiency? The common logic is that efficiency is inversly proportional to abstraction. The more abstraction, the less efficiency. That's why DirectX has a very small abstraction layer, and why (partially) why Windows GDI is so slow because it is abstracted so much.
You mean aside from USB, firewire, multimonitor, better plug & play, higher color icons, IE integration, improvement in memory management, harddrive organization optimizing, etc? By comparison, the changes from any release inside 2.2x is trivial. Plus, it did take many years to get from 2.0.x to 2.2.x so I don't think you can hold that over Microsoft. In addition, DirectX (a large component of Windows) was hugely overhauled in version 6 and 7 (near the release of Win98.) True, Win98 was not worth the extra 80 something bucks (I didn't buy it anyway. Actually, I've never bought a copy of Windows. Of course, I in no way encourage CD copying:) However, it was still much bigger than the changes between kernel versions. Also, take a look at Microsofts website sometimes. They post patches to security leaks quite often (much more often than the month or so it takes for a new kernel.) Sure Windows is unsecure, but that's because the model is flawed, not the slow pace of updates.
Some of Plan9's distributed computing stuff is cool, but it is not entirely new. Back around 1993, there was an OS called Chorus made by some french company. Chorus had a nifty distributed system where each machine (or "site" in chorus lingo) would run the microkernel (or "nucleus") Then, different machines would run the different servers needed. The kernel would either dispatch threads to these sites, or call the services of the servers one those sites. This seems to provide about the same features as Plan9, but seems less resource intensive than the huge levels of abstraction Plan9 has.
Sure Plan9 may be a nifty operating system for UNIX nerds, but what about speed? It seems that all programs communicate through a protocol, and that the interface to the system is hideously abstracted. Has anybody played with this system and tested its speed in relation to other OSs? ALso, this distributed computing thing sounds cool, but again, does the architecture translate into a tangible speed increase over systems like QNX?
You, my friend, are mistaken on several points. >>>>>>> Amazing, yet more wasted bandwidth on desktop systems. >>>>> ATA 100 is not wasted bandwidth. Sure no harddrive exists that can do a 100 (or even 66) MB/sec sequential transfer, but many exist that can do burst transfers over 50 MB/sec. These burst rates will only get faster in future drives. Harddrive transfers are full of bursts and sags and by capping the bursts, you lower transfer rate.
So tell me again why we are getting faster IDE speeds rather than moving to something like SCSI or IEEE 1394 (Firewire)? IDE blows if you're concerned with speed, 66 and 100mhz are fine if you're running several devices simultaneously but IDE doesn't do that does it? No. We'd be fine with PIO 4 since even 7200rpm hard drives can barely max out that bandwidth. >>>>>>>>>> We are getting faster IDE speeds rather than moving, because A) IDE doesn't suck, and B) We are already using it. Its like saying why not ditch UNIX because it is so cruffty. Sure there is better designs out there for OSs (I doubt anybody claims UNIX is the most academically perfect OS in existance) but people continue to use it because it is too much of a pain to start over. Plus, the alternatives suck worse. SCSI is expensive, and Firewire is half the speed of ATA/100 (400 mbps = 50MB/sec) and does even worse because it has trouble delivering that kind of bandwidth to a random access device. Lastly, PIO 4 would not be as good because A) It doesn't do DMA, and B) Again, 7200 harddrives can burst faster than that (hell, most can now do sequential transfers faster than that.)
Keeping IDE alive longer is pissing me off, why can't we move onto some more efficient standards? The Serial STS is an option, as is Firewire. >>>>>>> I'm not sure what serial STS is, but I do know there is a serial ATA standard in the works that should put SCSI to shame. Also, again, firewire is not an option. It is slower than even ATA66, and has trouble giving harddrives (or even ethernet cards. The VAIO PC has a firewire network interface and it doesn't stand up to 100mbps ethernet) bandwidth. Again, we're not moving to them because it is not worth the trouble (Rarely do ATA users have more than two harddrives, and ATA handles them fine.) You fail to see the point of ATA. It is not the standard that really holds ATA down (at least in ATA's target market of single to dual drives), because these days ATA is approaching SCSI in effeciency. The main problem are the drives. If you engineered an ATA version of the fastest SCSI drive available, it would be almost as fast as the SCSI version, and it would cost just as much. People don't move to SCSI because of the cost, so what have you gained?
Imagine if VA or LinuxCOmputers could ship FW enabled systems, they would make as much headway as Apple has with them. I bet many companies would even write up Linux drivers for their equipment since it is a viable platform that more people are migrating to. >>>>>> This comment shows two things. 1) You don't understand the point of ATA. ATA is not for multi drive monster servers, it is for a the workstation that has one or two drives in it. 2) You are delusional about the state of Linux. If companies didn't rush to develop drivers for Apple during its huge growth period during iMac season (not the mention the fact that there are many many more Mac users out there) what makes you think companies will support Linux?
Firewire runs at 1/2 the speed of ATA/100. Firewire is a serial link and thus measured in bits per second. It runs at 400 megabits per second, or about 50 MB/sec. ATA/100 is a parallel interface and measured in megabytes per second, and it runs at 100 megabytes per second. Barring that, Firewire is pretty crappy at this stage. Even a firewire network link (as featured in Sony VAIO PCs) cannot keep up with 100 megbits per second ethernet.
Even for single drivers, ATA/66 is not a useless upgrade. Sure there are no drives that will stream at 66 MB/s to disc, but there are a lot that will hit 55 MB/s bursts. Harddrive transfer is rarely a steady process. Instead,t he transfer rate has a lot of spikes and dip. By capping the bursts at say 33 MB/s, you're effectively lowering the average transfer rate.
I find it funny how the article says that this is another technology that Linux supported before Windows. Ignoring that fact that Windows had PCI, AGP, UltraDMA, Plug and Play, USB, Firewire, 3D acceleration, sound acceleration, SCSI, etc. all before Linux, I would like to point out why, for the near future, Windows will always have drivers for products before Linux. First, I would like to mention that this ATA100 support isn't an actual driver for a product, it is a driver adhering to a standard. Thus, this driver can be made before an actual product is on store shelves, because any controllers will adhere to the standard. Windows will always have the driver first, because no company in their right mind will release a PC product without Windows support. Barring a payoff from RedHat, the best a company will do for Linux is ship a driver in the box along with the Windows driver. The majority of the cases, however, involve a company releasing a Windows driver, people bitching, and the people finally going of to write their own driver. (In rare cases, the company may give up and just write the driver if the hardware is trivial enough or the company is desperate enough.) Thus, no matter what the hardware technology, Windows will have support first in the majority of cases, or at worst will have support at the same time. Now, when companies start shipping ATA100 controllers, Linux will take advantage of the ATA100 support built in, while MS can sit on its ass and know that company has already released Windows driver. To the end user, it wouldn't matter if Linux had support for ATA100 from version.99. To them, it seems that they have support at the same time.
44X CDROM drives in a cluster system? What are you smoking? Plus, why do cluster systems need high end graphics cards? I don't know of any GL implementations that can be clustered. (Get a WildCat 4200 for the main machine and be done with it.)
To tell the truth, I would rather have the news system that is hunky dory and polite. In my opinion, to many idealistic jackasses have made a mockery of the concept that "people deserve the news NOW." I think striving to deliver the truth, (or in this case not making an annoyance of yourself to mirror sites) is more important than getting the "scoop" on something. Slashdot has a lot of power. They cannot act like a small site does and post things without some forsight.
Slashdot Mirror
That's a great arguement, except not really. Windows rarelly fucks up some install, and if it does for you, you're trying to hard. The truth is, that you run the .exe, it reboots, you get on with your life. If you run funky programs on funky hardware, that's your own fault. I have used Windows ever since 3.1, and have yet to encounter a problem that requires me to hack the registry. My little brother fucks with his computer doing everything from Pokemon to photshop and he has yet to cause a problem like this. (Although he does have 50 items on his desktop that he can't figure out how to get rid of.) Windows fucking up is mainly an urban legend these days. If your hardware and software is old enough that it does screw up, you're probably knowledgable enough to fix it. If it isn't then it probably isn't screwing up/
Anal retentive little bastard aren't you? (Just kidding) I would do paragraphs, if I didn't have to type in this blasted little box!
I love the fact that you assume I don't use Linux. As of now I am running Mandrake 7.1 (for those of you who have read my other posts, yes I DO change distros that often) on ReiserFS running on a beta kernel running the beta nVidia GL drivers. I also have copies of Slack 7 and Suse 6.4 that I'll try out when I have the time. That said, I also like your cemment "the only thing that requires a kernel compile is a new hardware driver..." What other kinds of drivers are there besides hardware drivers? Plus, you also say that is only the case if you don't have it in your base distro. That is like saying Windows drivers install themselves. Which is true, but only if they come on the Windows98 CDROM. The truth is, that most people will not want to use older drivers, and for hardware that changes drivers frequently (graphics cards, sound cards, video capture cards, monitors in my case) a standard, easy method of installing drivers is needed. Second, about using ipchains, I still stand by the fact that it makes Linux harder. Even people who know enough about the system will go to the Mandrake or Redhat config tool to do that, and not being able to find it, will ask their nearest Linux nerd. What will they say? ipchains of course! Even Linux.com, which caters to Linux newbies, uses ipchains in its tutorial on networking. Ideally, of course, you're right. Practically, I'm right.
This has nothing to do with BeOS as an operating system. I have in the past extolled the virtues of Windows when it has deserved it or even Linux (gasp!) when it deserved it. My point is that Eazel says it is making Linux easy to use, and I say it is not. I'm simply providing BeOS as an example of an easy to use system. Many other OSs would have sufficed, but that is the one I am most familier with.
Who cares about this Linux thing. It is just UNIX. It just rips of everybody's ideas. KDE is just a clone of Windows, ad naseum. How many times have you heard that before? However, they are all false, as are yor comments about MS. MS integrated these things into the OS. In effect they pioneered these things on home operating systems much like Linux is a pioneer in making UNIX an easy to use cohesive environment for everyone.
You don't have to click/see file details. When you select the file, it pops up in a little Window in the side. That's a moot point however. Did Kideo integrate OLE into the viewer so that any OLE application could be used to provide previews in that box?
Those were still separate products. Saying MS doesn't deserver credit for integrating them into the OS is like saying that nobody in the Linux movement deserves any credit for anything. Most of Linux has already been done before, albiet by different systems. Windows already does 99% of the things GNOME and KDE are trying to do, yet they are billed as being pretty innovative.
Yea, I know it is kind of mean to harp on a system that is older than I am, but I think it is ridiculous that X is so far behind in modern display technology. (It is also ridiculous that we are still using it after so many years, but that's another story.) Coming from the all anti-aliased all the time world of BeOS, I notice that X has really poor font handling. If you read any of the FreeType docs, you'll see constant bitching about how X really doesn't provide any font support worth a damn. While X just recently got integrated truetype support, and is still far away from having anti-aliasing (which even Windows! has) Microsoft is introducing new technology to make fonts look even better! Aside from these small scope issues, this fact points to larger problems with the X architecture. It just wasn't designed to be extended cleanly. As such, X (at XFree86) still doesn't have genuinly usefull things like DPS and anti-aliasing. This is yet another reason that X needs
A) A huge, major rewrite.
B) To be replaced with something else.
Berlin looks very promising. True it uses CORBA for its API services, which sacrifices speed, but even at the beta level that it is in, it already has nifty features like anti-aliasing support, and a totally cool imaging model. Also, it is designed to be extended cleanly.
Exactly how is Natulis going to make Linux easier to use? All it seems to be is a file browser/desktop shell along the lines of gmc on steriods. Sure it may make navigating around the system easier, but what about the inherent difficulty present in doing anything non-trival in Linux? Take a good hard look at the Mac community for a clue. Mac users aren't brain-dead neophytes who can't tell a close button from a minimize button. A great many are people who know their way around the system well enough, and can get basic things done. Rare is the Mac user who doesn't know what extensions are or who can't get basic services working on their Mac. There is a reason for this. The Mac is easy to use. Linux is not. All the great GUIs in the world will not make Linux easy to use. Take something like installing an extension. Under MacOS, this is trivial. Under Windows, installing a driver is similarly simple. But under Linux it often requires a kernel recompile! Take the new nVidia kernel driver. I've have run quite a few beta drivers in Windows, but none have required me to manually remove OpenGL files from the path! Installing basic services on Linux is similarly hard. Internet connection sharing is becoming a big thing, and it is easy on everything except Linux. In Windows you go through the wizard to set it up. It asks basic questions and it works. In BeOS it is even more trivial. You go to the graphical NAT config, give it the IP of your internet interface and your LAN interface, click "NAT On" and restart the net server. In Linux, it requires learning ipchains and its complex syntax. Sure it ends up to be three simple lines of script, but A) It takes hours to get those three lines, and B) How is the user supposed to know where to put the script? Even simple things like changing host names or IP addresses, or adding new hardware become a chore in Linux. This need not be the case. A GUI CAN do something about this. Take a look at BeOS. It has most of the features of BSD networking, yet its configuration panel has something like 3 address fields and a checkbox. Telnet and ftp servers can literally be set up with a click and a password entry. It might not be as powerful, but it sure is easy. Yet, the power is still there. Navigate up to "/etc" and behold! Network settings! A lot of stuff in BeOS works this way, and it is pretty cool to behold. I have no doubt that Eazel will be a cool shell. It will make it easier to navigate around the system, and will lower the learning curve, which is always good. However, stuff like that is trivial. Ease of use should permeate the system, allowing the user to do more as their skills increase. It takes a very well designed system to allow some one with no skill to easily learn the system, yet not constrain those who have mastered those skills. Linux has a lot of cool stuff in it and has the potential to become an OS that is powerful for the hackers, yet easy for the newbie. The beauty of Linux is its multi-faceted-ness and malleability. Yet, the very people who are working on making Linux easier to use don't seem to have the right vision. They just don't get it.
I find it incredible that people accuse Microsoft of not innovating and just stealing other people's ideas. take a look at the second screen-shot in the directory. For some reason, the file brower looks hideosly familier. Oh, wait. Its IE in a GTK+ dress! It even has the little thing on the side that gives a summery of the file. If the Eazel developers are worth their salt, it will also use Bonobo to do realtime previews in that box, cementing its similarity to IE. KDE does this too, Knoqueror is both a web browser and file browser. The integration between the browser and the file viewer is cool (and very useful) tech, and MS is the one that poineered it. They might make crappy products, but give credit where credit is due.
The system may bootstrap using the DOS kernel, but the Windows kernel is certainly not a virtual kernel running on top of DOS. That was the case in versions prior to 3.1. Read the book about PCs by Peter Norten. It is quite helpful and describes certain things about the Windows kernel. In particular, it says that in Windows95, the functions previously handled by DOS were moved into modules within Windows95. Whether or not these modules contain DOS code is irrelevant. They are within Windows 95. If what you say is true, that Windows runs as a virtual kernel over DOS, then most tasks that require access to hardware would have to go through DOS. However, Win32 rarely, if ever, have to go into DOS mode. They only times that happens is when you have a funky program or driver that runs in real mode. What your saying doesn't make sense from the way the Windows arch is done. The lowest level is the huge Window kernel. Sure it loads stuff, but those are modules, and drivers (being modules as well) can be considered part of the kernel. Additionally, Win32 programs run on a system-wide virtual Win32 machine. Win16 programs run in a single Win16 virtual machine, and DOS programs run in multiple DOS virtual machines that load DOS before running the program. The major 16 bit DOS legacy that Windows has is mainly the legacy from Win 3.1. Major parts of some of the Windows kernel modules, like user, are 16 bit, which leads to instability.
PS> Programs and modules loaded into the kernel count as part of the kernel.
Actually, Windows works differently than Linux. Then kernel is located in a special
file (c:\io.sys) on your hard disk.
>>>>>>
Not in any Windows version is the kernel io.sys. It provides some DOS functionality, but the Windows kernel resides in kernel32.dll and kernel32.exe.
All of the extra hardware functionality (USB,
multimoniter, etc.) that you mention is tacked into the OS through a series of
interesting things like normal and virtual device drivers.
>>>
All the buses are in the kernel. That's why decent USB support didn't appear until 98 and why NT4.0 never got Firewire support until MS patched it. TheyUSB drivers required a lot of hacking to get past the kernel, and the Firewire drivers were impossible to write without the cooperation fo the NT kernel. Multi monitor is also part of the kernel because that is under the control of the graphics system, which resides in the kernel.
It's sort of like a
microkernel, only these drivers are accessed through a GUI (Windows) running on
top of the DOS 7 part and the kernel.
>>>>
Wrong again. Contrary to popular belief, Win95 does not run top of DOS. Its mostly anti-microsoft propganda. True, Windows 95 has DOS embedded in it, but does not use it all that much when running Windows programs. If you run fully 32 bit programs, Win95 rarely goes switchs into real mode DOS. WinME will finally take DOS out altogether, although it will still suck. I mean even Win3.1 only used DOS for the file system! Second, no Windows is really a microkernel. They want to tell you that NT is a microkernel, but in reality, it has most drivers embeeded in the kernel, and one big Win32 system server. Hell, in Windows 2000, the hardware abstraction layer includes calls to DirectX! (Though there is nothing wrong with that. It might be acedemically incorrect, but if MS would just let the DirectX guys do NT, Linux would be in major trouble.) Win9x is even more monolithic. Everything from the graphics and some GUI functions to file systems run in the kernel.
Unlike with a microkernel, the kernel never
actually touches the device drivers and things that Windows runs. Windows even
has it's own virtual kernel that runs on top of the actual one!
>>>
You're confused. The Windows kernel has complete access to hardware drivers. It doesn't run on top of DOS, it uses some DOS code in the kernel. Its just like the Linux kernel in terms of closeness to harware, but while Linux is completely 32 bit protected mode, Win9x has some sections that are real mode. Also, there is no virtual kernel. I think what confused you is that Win9x has a virtual machine that runs all Win16 programs, and many virtual machines to run DOS programs. All 32 bit programs run without a virtual machine.
Some of the other things you mention (icons, IE) are actually in executable code in
the GUI part of Windows and elsewhere, not through any interaction with anything
resembling the kernel.
>>>>>>>>>>>>>>
True, icons are not in the kernel. Faux pas on my part. However, they are pretty close. All the routines to load icons and do graphics are in the kernel. Like I said, Win9x is SERIOUSLY monolithic.
X-Windows isn't part of the Linux kernel, and IE isn't part of
the Windows/DOS 7 one.
>>>>>>>>>>
Stop sayiing Windows/DOS 7. There is no DOS 7 kernel in Windows, the virtual machine that runs on top of the Win9x kernel reports itself as DOS 7. Win32 programs never actually use that virtual machine. The DOS heritage that Win9x has is not that it runs on top of it, but that it uses a good deal of DOS code.
DirectX is weird. It is made up mostly of a large number of device drivers and
some executable code, although there are more complicated things in there.
>>>>>>
DirectX is god. DirectX is mostly device drivers, that's true, that's what gives it the speed. Conceptually, DirectX is a set of COM objects that talk to the DirectX HAL/HEL. The hardware drivers make up the HAL (hardware abstraction layer) and emulators make up the HEL (hardware emulation layer) the other executable code is the stuff that orchastrates to whole thing.
These security fixes mostly update DLLs and stuff NEVER the kernel. Again, most
of Windows is actually executables and libraries. The kernel is quite small and
doesn't do a whole lot except interpret for these executables.
>>>>>
Whoa, that's kind of wrong. Windows networking is implemented in the kernel on Win9x, and in kernel mode servers on NT. True, they might not be in the same executable, but they are for all purposes part of the kernel. Again, I don't think you quite have the right idea about Windows. In all versions of Windows, the kernel is quite large. (Again, in NT the kernel32.dll is not that big, but a lot of stuff runs that is loading into the kernel.)
In short, the actual kernel has not changed much, other than moving it from two
files (msdos.sys used to have part of it) into just io.sys.
>>>>
I'm assuming you're talking about Win9x here. The kernel is not io.sys or msdos.sys on any version of Windows. They are DOS modules loaded by the Win32 kernel to facilliate some operations.
USB and other nice things
never directly interact with the kernel, but work through executable code (win.com
and associated dlls and other files) that runs on top of the kernel and accesses
hardware.
>>>>>
Wrong again. USB and stuff does work through the kernel. DLLs may provide support for the actual device, but USB is a bus and busses in Win9x are supported by the kernel.
Linux, on the other hand, integrates USB and such into the kernel, so it
does not constantly crash because of the complex and unstable patchwork doing
things DOS was never meant to do.
>>>>>
You're sentences make no sense. Integrating something into the kernel make it less stable, not more. (Ever wonder why NT used to be really really stable in 3.x before they moved graphics into the kernel?) DOS has nothing to do with it. There is no code in the USB services that use DOS.
You have a very well thought out response. The problem is that you are doing
exactly what MS wants: seeing Windows as one big happy family rather than the
confused mish mash it is. Dig a little deeper, and you'll discover why the model is
insecure and why it crashes constantly.
>>>>>
You seem to be quite confused on what Windows is. I got all of my information from a BYTE article cirica launch of Win95 that detailed the architecture of Win95 (then Chicago.) (BTW. BYTE was THE nerds magazine. I've never since seen a mainstream mag that went into the kind of technical detail BYTE did. Read the one from 1993 about the new OSs that were coming out then. They talked about message passing and hardware abstraction layers like you had known about them forever!) True, Win9x is a mishmash, but the fact that it runs on top of DOS is just not true. Win3.1 did, but in Win95, everything was moved into a set of modules (such as USER32.exe and USER32.dll and GDI32.exe and GDI32.dll) which comprised the kernel. Some of those modules contained DOS code (at launch, GDI32.exe was largely 16bit Win3.1 code,) but that does not mean that Win9x runs on top of DOS.
Do you know how they managed the effiency? The common logic is that efficiency is inversly proportional to abstraction. The more abstraction, the less efficiency. That's why DirectX has a very small abstraction layer, and why (partially) why Windows GDI is so slow because it is abstracted so much.
You mean aside from USB, firewire, multimonitor, better plug & play, higher color icons, IE integration, improvement in memory management, harddrive organization optimizing, etc? By comparison, the changes from any release inside 2.2x is trivial. Plus, it did take many years to get from 2.0.x to 2.2.x so I don't think you can hold that over Microsoft. In addition, DirectX (a large component of Windows) was hugely overhauled in version 6 and 7 (near the release of Win98.) True, Win98 was not worth the extra 80 something bucks (I didn't buy it anyway. Actually, I've never bought a copy of Windows. Of course, I in no way encourage CD copying :) However, it was still much bigger than the changes between kernel versions. Also, take a look at Microsofts website sometimes. They post patches to security leaks quite often (much more often than the month or so it takes for a new kernel.) Sure Windows is unsecure, but that's because the model is flawed, not the slow pace of updates.
Some of Plan9's distributed computing stuff is cool, but it is not entirely new. Back around 1993, there was an OS called Chorus made by some french company. Chorus had a nifty distributed system where each machine (or "site" in chorus lingo) would run the microkernel (or "nucleus") Then, different machines would run the different servers needed. The kernel would either dispatch threads to these sites, or call the services of the servers one those sites. This seems to provide about the same features as Plan9, but seems less resource intensive than the huge levels of abstraction Plan9 has.
Sure Plan9 may be a nifty operating system for UNIX nerds, but what about speed? It seems that all programs communicate through a protocol, and that the interface to the system is hideously abstracted. Has anybody played with this system and tested its speed in relation to other OSs? ALso, this distributed computing thing sounds cool, but again, does the architecture translate into a tangible speed increase over systems like QNX?
You, my friend, are mistaken on several points.
>>>>>>>
Amazing, yet more wasted bandwidth on desktop systems.
>>>>>
ATA 100 is not wasted bandwidth. Sure no harddrive exists that can do a 100 (or even 66) MB/sec sequential transfer, but many exist that can do burst transfers over 50 MB/sec. These burst rates will only get faster in future drives. Harddrive transfers are full of bursts and sags and by capping the bursts, you lower transfer rate.
So tell me again why we are getting faster IDE speeds rather
than moving to something like SCSI or IEEE 1394 (Firewire)? IDE blows if you're concerned with speed, 66 and 100mhz are
fine if you're running several devices simultaneously but IDE doesn't do that does it? No. We'd be fine with PIO 4 since even
7200rpm hard drives can barely max out that bandwidth.
>>>>>>>>>>
We are getting faster IDE speeds rather than moving, because
A) IDE doesn't suck, and B) We are already using it. Its like saying why not ditch UNIX because it is so cruffty. Sure there is better designs out there for OSs (I doubt anybody claims UNIX is the most academically perfect OS in existance) but people continue to use it because it is too much of a pain to start over. Plus, the alternatives suck worse. SCSI is expensive, and Firewire is half the speed of ATA/100 (400 mbps = 50MB/sec) and does even worse because it has trouble delivering that kind of bandwidth to a random access device. Lastly, PIO 4 would not be as good because
A) It doesn't do DMA, and B) Again, 7200 harddrives can burst faster than that (hell, most can now do sequential transfers faster than that.)
Keeping IDE alive longer is pissing me off, why can't we move onto some more efficient standards? The Serial STS is an option, as is Firewire.
>>>>>>>
I'm not sure what serial STS is, but I do know there is a serial ATA standard in the works that should put SCSI to shame. Also, again, firewire is not an option. It is slower than even ATA66, and has trouble giving harddrives (or even ethernet cards. The VAIO PC has a firewire network interface and it doesn't stand up to 100mbps ethernet) bandwidth. Again, we're not moving to them because it is not worth the trouble (Rarely do ATA users have more than two harddrives, and ATA handles them fine.) You fail to see the point of ATA. It is not the standard that really holds ATA down (at least in ATA's target market of single to dual drives), because these days ATA is approaching SCSI in effeciency. The main problem are the drives. If you engineered an ATA version of the fastest SCSI drive available, it would be almost as fast as the SCSI version, and it would cost just as much. People don't move to SCSI because of the cost, so what have you gained?
Imagine if VA or
LinuxCOmputers could ship FW enabled systems, they would make as much headway as Apple has with them. I bet many
companies would even write up Linux drivers for their equipment since it is a viable platform that more people are migrating to.
>>>>>>
This comment shows two things.
1) You don't understand the point of ATA. ATA is not for multi drive monster servers, it is for a the workstation that has one or two drives in it.
2) You are delusional about the state of Linux. If companies didn't rush to develop drivers for Apple during its huge growth period during iMac season (not the mention the fact that there are many many more Mac users out there) what makes you think companies will support Linux?
Firewire runs at 1/2 the speed of ATA/100. Firewire is a serial link and thus measured in bits per second. It runs at 400 megabits per second, or about 50 MB/sec. ATA/100 is a parallel interface and measured in megabytes per second, and it runs at 100 megabytes per second. Barring that, Firewire is pretty crappy at this stage. Even a firewire network link (as featured in Sony VAIO PCs) cannot keep up with 100 megbits per second ethernet.
Even for single drivers, ATA/66 is not a useless upgrade. Sure there are no drives that will stream at 66 MB/s to disc, but there are a lot that will hit 55 MB/s bursts. Harddrive transfer is rarely a steady process. Instead,t he transfer rate has a lot of spikes and dip. By capping the bursts at say 33 MB/s, you're effectively lowering the average transfer rate.
I find it funny how the article says that this is another technology that Linux supported before Windows. Ignoring that fact that Windows had PCI, AGP, UltraDMA, Plug and Play, USB, Firewire, 3D acceleration, sound acceleration, SCSI, etc. all before Linux, I would like to point out why, for the near future, Windows will always have drivers for products before Linux. First, I would like to mention that this ATA100 support isn't an actual driver for a product, it is a driver adhering to a standard. Thus, this driver can be made before an actual product is on store shelves, because any controllers will adhere to the standard. Windows will always have the driver first, because no company in their right mind will release a PC product without Windows support. Barring a payoff from RedHat, the best a company will do for Linux is ship a driver in the box along with the Windows driver. The majority of the cases, however, involve a company releasing a Windows driver, people bitching, and the people finally going of to write their own driver. (In rare cases, the company may give up and just write the driver if the hardware is trivial enough or the company is desperate enough.) Thus, no matter what the hardware technology, Windows will have support first in the majority of cases, or at worst will have support at the same time. Now, when companies start shipping ATA100 controllers, Linux will take advantage of the ATA100 support built in, while MS can sit on its ass and know that company has already released Windows driver. To the end user, it wouldn't matter if Linux had support for ATA100 from version .99. To them, it seems that they have support at the same time.
Actually, MS writes drivers for things like system chipsets (which commonly include the ATA interface) and things like monitors, etc.
44X CDROM drives in a cluster system? What are you smoking? Plus, why do cluster systems need high end graphics cards? I don't know of any GL implementations that can be clustered. (Get a WildCat 4200 for the main machine and be done with it.)
To tell the truth, I would rather have the news system that is hunky dory and polite. In my opinion, to many idealistic jackasses have made a mockery of the concept that "people deserve the news NOW." I think striving to deliver the truth, (or in this case not making an annoyance of yourself to mirror sites) is more important than getting the "scoop" on something. Slashdot has a lot of power. They cannot act like a small site does and post things without some forsight.
Sorry, it's hard. The damn default is +2.
Oh, I get it. They should put that on the moderation, or make 1 the default instead of 2.