Dell has taken the unusual step – for a PC vendor of its size – of toning down its sales pitch for Microsoft’s Vista operating system and warning businesses of the migration challenges that lie ahead for them. The step is particularly unusual because one of the issues the hardware vendor is warning business about is the extra hardware they will need to buy.
Dell is just telling its corporate users to watch ahead before jumping and I think it is okay, I’d like that from my provider. For most home users changing from XP to Vista may not be as difficult or problematic than for large company networks, although it will be expensive indeed.
Anyway, if you take a look around at the computers that are being sold right now, most of them should not have any problem running Vista at all. Another thing is the installed base, that probably won’t be worth to migrate.
After all, for me the key thing will be software compatibility. If this one is properly achieved (which is not the case to my knowledge nowadays) it will be just a matter of time for Vista to take on the new computers at both corporate and home world.
I’d rather see exchanging XP for Vista as a downgrade right now, so I see Dell advice as a good idea.
It’s not about trolling or hating Microsoft or the so, it is just that it is expensive, unstable, and incompatible with a lot of the installed software base as well as with a lot of hardware, especially older one.
Apart from those things, I don’t like it, but this is a 100% subjective thing.
Yup.
Well, that’s not Microsoft’s fault. It’s the problem with the drivers because some vendors obviously didn’t notice the release.
That’s true, but most more or less recent software will run.
*I* think that the problem is that the people do not NEED Vista. It doesn’t have the features that would justify the hassle to upgrade all comps.
Seconded.
I doesn’t matter if it’s Microsoft’s or the hardware vendors’ fault. People want something that works. This isn’t an attack on either but someone has to take responsibility and must be held accountable.
Although I don’t have any experience with Vista apart from a quick test of Beta 2 in VMware more than a year ago I do have experience with various XP SP2 installations where various certified drivers either included in Windows or from hardware vendors are conflicting up to the point where the system is too unstable for use just after a fresh install.
The way drivers are included in the Linux kernel means that although not all drivers for everything under sun are included, many are and they are tested together constantly so conflicts between drivers are resolved much easier and faster because all drivers can be modified at once if necessary.
I am certain the NT kernel would be a lot more stable if Microsoft adopted the same policy of developing and maintaining drivers alongside the core of the kernel itself. As long as the kernel itself and all drivers are not tested together crashes may happen.
Having had a discussion on this with a relative who has been an MCSE since NT 4.0 but who has also worked with OpenVMS (VAX/Alpha), HP-UX and Solaris, he came to the same conclusion. He also runs my Slackware based operating system at home and it serves primarily as a virtualisation base to run all kinds of Windows releases on, since XP proved too unstable a few years ago.
I doesn’t matter if it’s Microsoft’s or the hardware vendors’ fault. People want something that works. This isn’t an attack on either but someone has to take responsibility and must be held accountable.
I would agree with you, however, I notice that if a driver for Linux is either missing altogether or incomplete, people are quick to point the finger at the hardware vendor. But when the same scenario happens under Vista, some of these same people will scream that it’s all Microsoft’s fault.
Frankly, I believe it is the fault of the hardware vendors in both cases. Btu in the end, if stuff doesn’t work, then the reason really doesn’t matter as far as in end user is concerned. It either works or it doesn’t. If it doesn’t, then somebody needs to fix it.
Frankly, I believe it is the fault of the hardware vendors in both cases. Btu in the end, if stuff doesn’t work, then the reason really doesn’t matter as far as in end user is concerned. It either works or it doesn’t. If it doesn’t, then somebody needs to fix it.
Actually, in the case of an incomplete, in-kernel, open source hardware driver, with Linux you’d be justified in (mostly) blaming the kernel developers. Not that that happens often, and this is where open source really comes into its own – because the Linux kernel devs have control over the kernel, because they have control over open source drivers, and because they demand high quality in their software, you don’t get the situation where high-quality, production code suffers because of some crappy 3rd party driver. Of course, mistakes can happen, and I doubt pre-release code is that stable, but the only time I have ever seen a kernel panic on Linux that wasn’t due to my own stupidity passing the boot flags was, indeed, a case of a crappily-implemented or unmaintained part of the kernel. Since that experience with ReiserFS3, I have since sworn off ReiserFS!
So the lack of (or extraordinarily poor) support of webcams, winmodems, scanners, and wifi cards in linux is something the kernel devs should be held accountable for? I don’t believe that, and I doubt you do either.
The big problem with the linux way of having drivers as kernel modules is that any time the kernel changes, all the drivers have to change too. This could mean a recompile, but it could also mean more substantial changes. Windows has a layer of abstraction which gives vendors a stable ABI to work with, which means that when the kernel gets patched, drivers still work (the majority of the time anyways)
The linux way works when drivers are opensource, but the linux way makes commercial support pure hell. Since being business friendly is what windows is all about, it would make no sense to go that way.
“So the lack of (or extraordinarily poor) support of webcams, winmodems, scanners, and wifi cards in linux is something the kernel devs should be held accountable for? I don’t believe that, and I doubt you do either.”
Is that true!? Your spreading little lies again. I have a scanner here a Cannon IDE 30. I have a wifi card here ;a zyxel 202. Both work, if fact the software in Gnu is better(read not trash) than that bundled with the scanner or in Vista, and connecting my wifi to my router;gateway;modem, is simpler process on Gnu for my wifi card. In fact my router;gateway;modem uses Gnu.
I actually chose the wifi card from http://linuxwireless.org/ compatibility list.
The world has evolved past linmodems now, but they were pretty rubbish, and a hardware modem was the only way to go, but many linmodems are supported.
I know nothing of webcams, but I suspect there are also compatibility lists. In fact there was an article that one individual had wrote webcam support for over a thousand webcams.
Is it the fault of Gnu if it is unable to support these devices!? Absolutely especially if they are given access to proper hardware specifications, and a real working relationship with the company involved. The reality is an awful lot of hardware seems to work better on Gnu despite this. Oddly you can buy everything 100% compatible products of *all* those devices *together* for less than the cost of vista home basic oem.
I do find it funny that you lie that you have to recompile the kernel for new devices, when very few compile their kernel at all. I believe Distributions like Suse; Ret Hat; Fedora; Ubuntu all come with a recompiled version…with everything
Microsoft do not even have this excuse or any other for poor hardware support in Vista. Its not like its been in beta for like forever, or hardware/software companies were surprised by the Vista launch, or the fact that Vista was released November last year. When will it come out of beta.
BTW making loads of cash is what Microsoft is all about.
Edited 2007-07-05 20:16
First of all, GNU has nothing at all to do with Linux drivers. Secondly, I wasn’t arguing that Linux does not in fact solve all the worlds problems, or that vista is not in fact, the anti-christ.
What I was saying is that the Linux way is technologically very good, but it doesn’t work well with commercial support. The reason there are not many closed source drivers in the Linux world is that, as I said, it is a support nightmare. A business would literally have to release a new driver every point release of the kernel. This is a non issue for OSS drivers, as they just get a recompile along with everything else.
Edited 2007-07-05 20:19
“First of all, GNU has nothing at all to do with Linux drivers. Secondly, I wasn’t arguing that Linux does not in fact solve all the worlds problems, or that vista is not in fact, the anti-christ. ”
The only thing you can say is the “FSF” hasn’t done a lot of code in GNU’s kernel, and it hasn’t. I am not talking about a kernel. I’m talking about the OS.
The only think you are arguing is that Vista’s still being in beta, and having poor hardware support is excusable, which its not. You say its not the fault of Microsoft and *try* and use Gnu’s kernel as an example of poor hardware support when Gnu has both excellent hardware support even with limited access to hardware; 70,000 employees, Billions in the bank…and lets face it being a Monopolistic OS.
You continue by perpetuating the lie that on Linux you have to *compile* stuff to get things working in the kernel, and you lie that common devices don’t have support under Gnu.
Edited 2007-07-05 20:44
The GNU kernel is called Herd, and it has been in development for so long it has kind of become a joke. And you were talking about a kernel, because we were discussing Linux’s driver model. For someone who is such a huge fan of Linux, you really need to get your facts straight.
I didnt argue anything about Vista. I am not talking about vista. I am talking about using a HAL for everything as opposed to having drivers run in kernel space. I didn’t say it had poor hardware support, I listed several common types of devices that linux devs have had trouble supporting through no fault of their own.
Please do a better job reading my posts before flaming me.
Binaries don’t magically compile themselves, they have to come from somewhere. In the OSS world, the source is there so it is no problem, in the commercial world, that task is placed on individual vendors.
As for common devices, it is hard enough to reverse engineer hardware you dont have specs for. It is insanely hard when half the processing is done on the hardware, the other half is done with software which is embedded in the drivers, as is the case with the stuff I mentioned.
“The GNU kernel is called Herd, and it has been in development for so long it has kind of become a joke. And you were talking about a kernel, because we were discussing Linux’s driver model. For someone who is such a huge fan of Linux, you really need to get your facts straight. ”
I think your talking about “Hurd”. http://www.gnu.org/software/hurd/hurd.html and If the kernel was the *only* part of the *modular* operating system I use. You might have a point, unlike Vista users I have a *choice* between kernels, based on my *my* requirements. I simply chose one that comes under a GNU license. That just happens to be the best.
I am perfectly willing to discuss the Open development development model that attracted Linus to use the Gnu License in the first place over choosing a more proprietary License but you only seem interested in spreading lies about Gnu compatibility, and making poor excuses for Vista’s shameful hardware support.
Binaries do come from somewhere, normally from what is know in the Gnu World as a Distribution, which in essence provides a bundled choice of a variety of applications; kernels; desktop etc etc, and normally includes a package manager to add your own. All the main Distributions OpenBSD; Ubuntu; Fedora; Novell etc come in binary form. Open source just allows you see the source, but it does not mean you have to compile it yourself. In fact these distributions are made up of Free software as well, this does not mean you don’t have to pay for it, but means you can run, copy, distribute, study, change the code. It does not force you to do things however. They also include applications *only* available in binary form.
Often hardware specifications are *available*; Work on Gnu’s kernel is predominately done by companies(read commercial) with Hobbiests only producing 3.5% of the work. Please to not lie about its development. In fact you should support those companies that, support open-source, becuase it improves *your* chance of this hardware having continued support in future, continuous improvements; compatibility etc etc. Otherwise you end up with major regressions or your hardware not being supported at all like under Vista.
Edited 2007-07-05 21:34
There you have it. Hurd is the kernel part of the GNU operating system. We are not talking about GNU, we are talking about Linux.
I think you have a complex or something.
I AM NOT TALKING ABOUT VISTA
how many times do I have to say it?
That is a really long explanation that has nothing to do with the discussion that was going on in this thread about abstraction layers vs kernel modules for driver development.
I am sorry if English is not your first language, I will try to phrase this another way since you seem to have completely misunderstood what I was saying.
The Linux Way is to have drivers as part of the kernel. This has several benefits, but one of the negative sides is that any time the kernel changes at all, the drivers have to change. For open source, this is no problem, as most of the time it just needs a recompile to a few lines changed. This is not a big issue for open source drivers, as the person who decides to maintain the code does the work. On the other hand, businesses who choose not to open their code have a much more difficult time, they will need spend their resources on maintaining their driver against an ever changing kernel. Since drivers are only there to help hardware sell, they are making no real money off of this work.
On the flip side, the windows way is to have the drivers plug into an abstraction layer which does not change throughout the life of the os. That means they are only coding against a single ABI, not an ABI that changes even during bug fix patches. Hardware companies love this, as they only have to do the work once. That is why whenever a new version of windows comes out, all the bargin bin hardware that people had by companies which no longer exist suddenly end up with stuff that will never work with any future version of windows. It is also a big part of the reason that vendors are much happier to support windows, as it takes significantly less longterm work (the other reason being the whole 90% of the world using the platform thing.)
Both models have advantages and disadvantages. My comment is that it would not make sense going the linux way for windows, not that the linux way is inferior at all. In fact, there are plenty of instances where hardware is supported under linux and will be until the end of time, while on the windows side it stopped working in the XP era (the SANE project has plenty of examples of this). It is far more difficult to reverse engineer the hardware, but once it is done it is done forever. Again, tradeoffs.
Last time, I was not talking about what you seem to think I was talking about. My view of the technology world does not consist of some grand battle between good and evil, and there are plenty of times that I will comment on something that has NOTHING to do with the superiority or inferiority of windows or linux.
I am 100% certain I’m talking about GNU OS if you wish to discuss a kernel either Hurd , Linux, BSD kernel, Solaris kernel which can all be part of GNU OS I will be more than willing to do so.
I’m glad we are finally focused on the development model. I’ll simplify it for you. Gnu kernels can come out as often as two and a half months, and development is transparent; meaning smaller regressions and they are fixed, greater improvements; faster innovation, all from having regular updates; better security; improvements from similar drivers often have *mutual* benefit to drivers of a similar type. Any device made available to a Gnu kernel can work *forever*
The propriety kernel like that of Microsoft Vista means the kernel only changes every 5 years, with no maintenance for overall security; stability, and when a change is made everything stops working or suffers major regressions, bug filled drivers remain do, as we see from Vista hardware support is still not on par with its predecessor and has been out since *last year*. I believe that one of Gnus kernels is having 3rd release since the launch of Vista this week.
There are no trade-offs.
Please focus on the topic in hand rather than creating lies about Gnu OS.
Edited 2007-07-05 22:34
As far as i know the bsd kernel isn’t GNU nor can it considering GNU stuff is GPL’d and BSD has its own license
“As far as i know the bsd kernel isn’t GNU nor can it considering GNU stuff is GPL’d and BSD has its own license”
OpenBSD is a distribution not a kernel, look through the included packages. There are more CopyLeft Licenses like that of GNU, than that of BSD, and GLP is by far the most used license in the List.
abiword (2.4.6)
alsa-lib (1.0.14a)
amarok (1.4.6)
apache (1.3.37)
bind (9.4.1)
cups (1.2.11)
db (4.5.20)
dhcp (3.0.6)
emacs (22.1)
evolution (2.10.3)
firefox (2.0.0.4)
gcc (4.2.0)
gimp (2.2.15)
glibc (2.6)
gnucash (2.0.5)
gnumeric (1.6.3)
gtk+ (2.10.13)
httpd (2.2.4)
iptables (1.3.8)
jre (6u1)
k3b (1.0.2)
kdebase (3.5.7)
kdewebdev (3.5.7)
koffice (1.6.3)
libgnome (2.18.0)
linux (2.6.21.5)
Package
module-init-tools (3.2.2)
mono (1.2.4)
MPlayer (1.0rc1)
mysql (5.0.41)
nautilus (2.18.3)
OpenOffice.org (2.2.1)
openssh (4.6p1)
openssl (0.9.8e)
perl (5.8.8)
php (5.2.3)
pidgin (2.0.2)
postfix (2.4.3)
postgresql (8.2.4)
Python (2.5.1)
qt-x11 (4.3.0)
samba (3.0.25b)
seamonkey (1.1.2)
sendmail (8.14.1)
thunderbird (2.0.0.4)
udev (113)
vim (7.1)
webmin (1.350)
xfce (4.4.1)
xine-lib (1.1.7)
xinetd (2.3.14)
xorg-server (1.3.0.0)
The point is an OS is more than a kernel. The kernel albeit an important part of the OS. In the GNU world we have a choice of kernels, becuase GNU is a modular OS, and those come under many different licenses.
Edited 2007-07-06 02:41
Why not? What has changed? Was the code made more simple hence more bugfree? Were all promised advantages implemented we’ve hoped for? Have companies become so much richer to throw their money to half baked product?
Why not? What has changed? Was the code made more simple hence more bugfree? Were all promised advantages implemented we’ve hoped for? Have companies become so much richer to throw their money to half baked product?
Yeah, I don’t see the time honored practice of waiting for SP1 for any MS OS changing anytime soon.
Mine has. It’s called waiting for SP2.
I think it is more waiting for drivers and other software than Windows itself.
Hi! This is a friendly reminder from Dell. Just so you know, if your business plans on paying Microsoft tons of money to upgrade to Vista, you should also consider paying us tons of money too. Because, um. . .you need more hardware or we’ll make something break on you. That’s not a threat. We’re your friend! Therefore, it’s a warning
.
Vista seems to work in Dell’s favor here
.
Somebody might want to try out spell check…
Yeah, looks like they misspelled “downgrade”.
Maybe till SP2; SP1 is near and diffenetly they will fix bugs in that release and nothing about performance will be addressed.
Vista still has almost all of previous generation bad inheretance of bugs and shortcommings; and it adds more a layer of performance degradation, so the solution to this is to begin a new OS development based on one of the Unixes available out there; that way they could be 100% sure of success on the long run, if that what they provision.
I cannot put vista on any of my fairly old hardware (P4@3 GHZ, 1GB RAM DDR Dual channel, GF660GT) because it would feels like my PIII @800Mhz system. There is no killing feature(s)in vista that encourage me to install it.
Happy with my Linux Unbuntu/ OSX/Windows Server2003/WXP environment I have.
“so the solution to this is to begin a new OS development based on one of the Unixes available out there…”
From where and whence did the notion arise that *every* OS must be based on Unix? Unix is not the be-all and end-all of OS design. It is not the case that any OS that does things differently than Unix is doing those things wrong by definition. It’s not the case that advancement in OS design stopped ~35 years ago with the birth of Unix. Even back in the day, there was much debate regarding Unix vs VMS. And there were OSes like TOPS-20/TWENEX being run by major corps and universities. There was no wide consensus that Unix > everything. (DEC’s mismanagement killed off TOPS-20 and VMS, thus ceding the enterprise OS field to the Unixes, but Unix’s victory over those OSes wasn’t based on a consensus that Unix simply blew those OSes away technically.)
Remember, Mac OSX 10.0 was horribly slow (much slower than is Vista relative to their respective predecessors, Mac OS 9 and XP) and had many deficiencies, despite being built on Unix. Mac OSX improved with versions 10.1, 10.2, and finally 10.3 (the first really good version), just as Vista will improve with SP1 and SP2.
(Mac OSX being built on top of Unix wasn’t out of some belief in the greatness of Unix, but was a side effect of being built on NeXTStep, which ran on Unix and NT at the time. Apple didn’t go with the NT version of NeXTStep because that would result in a dependence on Microsoft, but the NT version of NeXTStep ran just as well as did the Unix version.)
(Speaking of Mac OS X, I would’ve preferred that Apple’s Copeland project succeeded, so that their OS wouldn’t be yet another Unix, and there’d be a third type of OS commonly in use (besides Unix-based and NT-based).)
Anyway, there’s nothing wrong with the NT-kernel. Replacing the NT kernel with Unix kernel buys nothing, and is in fact a step backwards. But if Microsoft wants to make a new OS, I’d prefer that they build on top of Singularity rather than the ~35-year old OS design that is Unix.
Edited 2007-07-05 18:11
“… so the solution to this is to begin a new OS development based on one of the Unixes available out there …”
Well, yes and no. It’s just that no one is paying attention.
http://plan9.bell-labs.com/plan9/
“(Speaking of Mac OS X, I would’ve preferred that Apple’s Copeland project succeeded, so that their OS wouldn’t be yet another Unix, and there’d be a third type of OS commonly in use (besides Unix-based and NT-based).)”
I totally agree. But reality was knocking. It was ahead of it’s time.
http://lowendmac.com/orchard/05/1108.html
http://www.igeek.com/articles/History/CoplandAndNeXT.txt
A/UX
http://www.aux-penelope.com/index.htm
hylas
It’s not the case that advancement in OS design stopped ~35 years ago with the birth of Unix.
Certainly not! Look at GNU/Linux, FreeBSD and Solaris.
I would’ve preferred … there’d be a third type of OS commonly in use (besides Unix-based and NT-based).
Me too. But OSes that follow in the traditions of BeOS or Amiga may one day be mature. As may SkyOS. In the mean time, there is a very wide selection of Unix-based OSes and variations to enjoy.
I doubt that “replacing the NT kernel with the Unix kernel…[would be]…a step backwards”, since although Unix was originally designed 6 years before the kernel that NT reimplements, it HAS been developed further. And some things just work better in Unix than VMS/WNT (a single filesystem tree, for example). Besides, if anyone else (including Microsoft), in the past, now, or in the future wishes that their operating system had become a multi-vendor standard, then they should have taken the trouble to either open source it, or licence it to other vendors cheaply whilst making sure that all versions remained compatible.
Had an Athlon64 3200+ (s939), 4x256MB DDR 400Mhz, ATI X1300 and Vista ran like a charm, even the beta ones. Even played games quite normally (Quake 4 for instance).
Although I admit there were minor hickups sometimes opening windows, but nothing all to bad.
There are three things that can really hurt vista on older hardware:
1) 5400 rpm HardDrives
2) Crappy motherboards and drivers (mind you I have seen plenty of motherboards which cause 100% CPU usage even under clean installed XP during idle -> some sort of physical error on some cunductor probably…)
3) Slow RAM (< 400 Mhz).
In my experience you get roughly the same usablity for the following RAM ammounts:
256 XP = 512 Vista (do turn off AERO here
)
512 XP = 1024 Vista
1024 XP = 2048 Vista
But if it were as you tell, which is the killer application that makes you need Vista to “run like a charm” on your system that would not run under Windows XP as if it had twice the memory, which would run at least like let’s say one and a half charm ?
For me it’s mainly system-wide search and filtering (i am in publishing and it speeds work up very much), quick sleep/wake, new mediacenter and multimedia functionality, Explorer with giant thumbnails (I find files much quicker this way), Wi-Fi connectivity without constant nagging, and many little improvements like per-app volume (invaluable!) drag-n-drop customization of favorite folders in Windows Explorer. Those aren’t killer features, but simplify things a lot. There is no known virus for Vista yet, and it’s as fast (or slow?) as XP on both my computers (not exactly top notch – A64 3k+, 1,5 GB RAM and GF6600 and laptop P-M 1,7G, 1 GB RAM and X700). I have no problems with HW (even specialty like Cotour Shuttle for my video editing) and most SW. Your mileage may vary.
I also like how it doesnt look like garbage, how everything across the board feels smoother and snappier then xp, how the networking is no longer worse then third party reverse engineered solutions, and how you dont have to drill through the start menu anymore.
My favorite improvement hands down is explorer, took them about 20 years, but ms has finally put out a good file manager.
For what it’s worth, with Microsoft’s TweakUI you could adjust thumbnail size up to 256 pixels square in Windows XP. However, Explorer is extraordinarily slow at it, so it wasn’t worth it for me — there was plenty of room for improvement, and Vista no doubt takes advantage of better image handling and resizing.
I wonder what you guys use for HW that you constantly run into problems with Vista? I have several niche devices (tablets, shuttles, tuners, HD cams) and it all works. The one device which keeps nagging for drier is my Nokia bluetooth, but it works anyway without it.
“I wonder what you guys use for HW that you constantly run into problems with Vista? I have several niche devices (tablets, shuttles, tuners, HD cams) and it all works. The one device which keeps nagging for drier is my Nokia bluetooth, but it works anyway without it.”
Without discussing Trusted computing paths, performance etc etc. I’m surprised that bluetooth does not work for *you* on Vista, as bluetooth has a *common* interface.
Edited 2007-07-05 20:44
We’ve had no problems with upgrading to Vista at the company where I work. I quite like Vista myself and this is coming from someone who uses Macs at home.