OSNews

1. Users don't change they computers because of a new OS, not in 2005. Most corporate users will keep Windows 2000 till hell freezes, it's just "enough". Same for Office. Microsoft revenue?

2. Hurt DVD decrypters and help Hollywood, well they slould do better films instead, but hey! Microsoft makes an OS that fucks its users so they can not make copies of DVDs they rent? Allright, what's next, a Windows Vista that can't be pirated? I like Linux and FreeBSD so that sounds as great new to me. Way to go.

Next year will be very interesting with Mactel and Vista maybe we see unexpected things on the os area.

As I've said before, it isn't MS that is the source of the DRM problem. That is the domain of the content producers/marketers. You want DRM to go away, stop consuming the crippled crap that gets shoveled at you.

As I've said before, it isn't MS that is the source of the DRM problem. That is the domain of the content producers/marketers. You want DRM to go away, stop consuming the crippled crap that gets shoveled at you.

While I generally agree, Microsoft is a partner of the 'content producers/marketers' in that MS benifits from implementing DRM as it allows them to differentiate Windows Vista from previous versions of Windows and non-Microsoft operating systems. Apple sees this, and is also jumping on board.

If both dropped DRM plans, things would be different.

The thing is, people have a great quality for poor quality, especially if high quality is much more expensive. Wide screen flat panel TVs and surround sound have been on the market for years, but the majority of people still buy 36 inch (or smaller) TVs and $100-$200 dollar stereos with ordinary stereo sound.

Look at it this way. If "disabled mode" is good enough, people will live with it and DRM hardware won't get any more traction than IBM's MCI "standard" and they'll lose out to cheaper hardware. If "disabled mode" isn't good enough, people will bawk the same way they did with region coding. Content producers with poor "disabled mode" content will lose out to ones with good "disabled mode" content, and DRM will fail like region coding failed.

MS dropping integrated DRM would make little impact in the long run. The media execs would still have the ****ed up notion that treating their paying customers like theives is a good method to raise sales.

If nobody does it, maybe Microsoft will buy a clue and stop releasing overly bloated OSs that do nothing but bother you while you're trying to work.

You must still be dreaming.

It's 6am here, and I can't get Windows to stop BEEPING at me. I don't even have sound drivers installed! I have the volume control turned all the way down. And its still BEEPING at me!

Better just go back to bed for a while .

You can cut the pc speaker wires inside your computer if you want to stop the beeping.

My sound card came with instructions on how to cut the wires to the speaker and solder them to the sound card for volume control (roughly 4 years ago). I could just imagine some unknowning consumer buying the card and thinking a computer upgrade required a soldering iron...

Or just disable it in the Device Manager. I believe you must show hidden devices to make it visible.

Someone suggested cutting the speaker wires...I say either simply unplug it, or if your a DIY person, put it on a switch.

Just what I was saying to a friend, I'm not going to buy a supercomputer just for a desktop environment to run on it. I use FreeBSD 6 with KDE now, and that's bloated enough for my taste. I don't need it to suck up more of my hardware because the hardware is meant for the occasional gameplay, not a flashier icon for my trashcan. The difference between MS and BSD was very clear when I had both Windows 2000 and FreeBSD 5 running in dualboot, BSD was quicker in use.

And all so you can get, mostly, the same neato animations you can get on a tenth the hardware with e17 or maybe (if it turns out good) KDE 4....
I'm actually a little bit annoyed to see the direction GTK people have gone with Cairo. If they can make it quick so that you can barely tell it's doing more work that's fine, but I don't want to need a gfx card to have good 2d graphics. I'm happy to buy a cheap gfx card for all my desktops, but my laptop is sort of froze where it is!

Yeah, it's ridiculous ... Just build yourself a render farm to play simple games. Like the bozo said RAM's cheap anyways, it's no issue.

I don't know who this feature is for. Is it for the content providers? Do they want a system where they can offer HD content that only 1% of the target audience can view? Hooray, the world of HD media is upon us! Quick, everyone go out and buy monitors that don't exist, 2GB of something called DDR3, a $300+ video card and a brand new hard drive, so that when Windows Vista comes out, you can pay an exorbitant amount of money to download a movie that you can't do anything with but play it on your computer monitor!!

No one has enough money for this. This is fascist, elitist, and anticompetitive. I posted a week or two ago about how DRM isn't evil, it's abusive DRM policies that are evil. Well, this is an abusive DRM policy, and it is pure evil.

There is nothing technically preventing someone with a 1920x1080 widescreen LCD from downloading an HD video in H.264 and playing it in full HD with mplayer, but it is not to be. Although, I did have a good time reading the code here:

http://mplayerhq.hu/cgi-bin/cvsweb.cgi/~checkout~ffmpeg/libavcodec/...

It's amazing that in spite of the effort the free software community puts into developing open implementations of open standards, commercial interests want to somehow render them useless. We can control how people use proprietary software, but not open source software, they say. When the commercial interests lose control of their target market's usage patterns, they get angry. Why doesn't MS listen to us when we say that we get angry when we lose control of the media we buy? The media consumers buy way more MS products than do the media producers. Why is MS listening to them?

Aero or Avalon or whatever the new graphics system is called accounts for increased demands for video memory. Instead of painting directly to the screen buffer when asked to do so, applications will now paint their windows into off-screen buffers, with the graphics system combining them on screen. At 1280x1024 in true color a buffer for only one full-screen window eats up 4MB.

But even so 256MB graphics memory seem a bit over the top. And God knows what those 2GB of main memory are for.

But even so 256MB graphics memory seem a bit over the top. And God knows what those 2GB of main memory are for.

It's short hand; they don't know what will be optimal.

* 256MB should do the trick for flashy video; 128MB is common, 128 * 2 = 256.

* 2GB main memory; If you need 512MB now, and Vista will have heavier requirements (2x), and 64bit is 2x32bit, you end up with (2 * 512) * 2 = 2GB.

I think they are right in the long run, though I would hope that they are doing some serious optimizations. The Linux kernel folks love scraping out another 4K here and there.

No, it's pretty much because while MS keeps piling features on the OS without addressing low-level problems, particularly (in this case) with regards to managing heap memory. I had an assignment a couple years ago where I needed to implement a my own malloc package that runs at least 150% as fast as linux's implementation for allocating memory. The challenge is to efficiently find not only a chunk of unallocated memory that is at least as big as the requested allocation size, but the chunk that leaves the heap as defragmented as possible. If you just allocate the first available chunk, the heap quickly becomes full of free memory chunks that are too small for the requested allocations. It seems as though a process is taking up a whole slew of memory, but it's really that a smaller amount of data is spread sparsely over a whole slew of memory.

The application layer is also full of culprits, for example the desktop search feature which indexes your data files. It's either storing the cache in main memory, or it's paging it out onto the disk. Windows doesn't really have a swap filesystem, so paging is inefficient on Windows. It would probably prefer to keep the large cache in memory if at all possible.

One thing is for sure, it's not .NET, because I don't think there's really any .NET applications shipping with Vista. I could be wrong, but at least there is no focus on ".NET" in Vista. I don't think this was the case since back when Longhorn was supposed to feature subscription-based application services running on remove servers.

good lord... is windows memory allocation really that far behind? I can't imagine they're still using the same ol' algorith they were back in the early multiprocessing days. they've got some pretty intelligent designers, i have trouble beleiving they're using the 'first available chunk' approach.

i know one thing that i can't stand is how windows paging works. it always takes over the system. obviously i know very little about what they're doing under the hood, but it seems like page faults are a killer on an otherwise fast machine in the windows world.

i would say it's all because of windows "services" and the new GUI it has.

More services and features means more complexity.

More complexity means a nearly exponential increase in the number of possible interactions.

All of that needs to be secured.

To secure any system or to run it in a stable manner, it is necessary to reduce complexity before checking for security holes...otherwise you drive yourself nuts.

XP is exceedingly difficult to secure. It can be done, but it is not at all easy to do. Vista looks like it will require quite an even higher level of commitment to concerned admins to handle it properly.

Expect those concerned and capable admins to be even fewer in number.

Expect that there will be substantial security issues that will take about 2 years post Vista release to resolve if -- and only if -- Microsoft takes security seriously this time around.

Microsoft is in risk of being crushed by the features they promote; they are not big enough or well organized enough to handle all the secondary issues. No single company or group is.

Are physical pages twice as big in x86-64? This would mean that there are half as many pages in x86-64 as in x86 for the same amount of physical memory. Therefore, this guy's statement *could* be true, but only in the absolute worst case where each process requests one page of memory and allocates half of it, on average.

64 bit memory paging requires a totally different strategy to handle.

a bare minimum 32bit table with 4kb per page is gonna run you about 4megs of memory.

a 64bit table using the same strategy is gonna be over 30 million gigabytes.

i can't remember the math offhand, but i remember it being covered in Andrew Tanenbaum's Modern Operating Systems.

your not simply doubling the address space by going from 32 bits to 64 bits, you're increasing it by orders of magnitude.

2^32 = 4000000000000
2^64 = 10000000000000000000000000

doh... i'm retarded. I had my calculator in hex when i did the 2^32 and 2^64... thought those numbers were coming out a bit too even.

need more coffee...

I'm not up on the specifics but amd64 has at least two more levels of page tables than ia32.

Sounds like a "strategist" at Microsoft is simply someone who's got no idea how thinks actually work.

He doesn't have to. If it is simple math for the uninformed (see my post a couple up with 2 * 512 * 2), it is less confusing. The point is that eventually people will have that kind of hardware. It is not at all unreasonable that by the time Vista settles down in 3 years, 2GB RAM and 256MB video will be common.

While it's true that pointers and registers double in size, most data stays the same size and code only grows by about 10% with x86-64.

Thanks for the details. I did a similar calculation back in the transition from 16-32bits. At that time, it was not as efficient, though we ended up changing some habits and corrected the problems. Now, compilers handle most of the dirty details.

While it's true that pointers and registers double in size, most data stays the same size and code only grows by about 10% with x86-64.

Is there any results on this?

In favor of more memory:
- All fixed address operations will require 32 more bits.
- 8 byte memory alignments eat som memory.
In favor of less memory:
- More registers will reduce the need for register spilling.
- 64 bit calculations require less instructions.

So: does anyone have any data?

In reagrds to whether 64bit code is about 10% larger than 32bit code... well, lets look at a major package for both architectures. This is for Fedora Core 4 Core RPMs.

gimp-2.2.7-4.i386.rpm
RPM size = 9885K
content size = 25M
gimp-2.2 binary = 3244464

gimp-2.2.7-4.x86_64.rpm
RPM size = 9998K
content size = 25.9M
gimp-2.2 binary = 3295472

difference in RPMs = 1.1%
difference in contents = 3.6%
difference in executables = 1.6%

I would say that 10% is being generous. Most programs will probably see very little difference. A glance at many different RPMs for 386 and x86_64 showed no major differences in size. I'm not saying 64bit will NEVER be considerably bigger, just that MOST programs won't be.

The idea is that we don't _NEED_ the upgrade.

Take a look at this guys screenshot, if you can't see it, you got a dumpy monitor

Wow that looks... terrible. Is the image blown up?

I think it is, after all the 30" Apple Display has a 2560 x 1600 resolution and HDTV max is like 1200 x 1000 (1080i)

The EyeTV can show multiple sizes including actual size.

Quartz 2D Extreme is still disabled by default in Tiger 10.4.2, even if Ars Technica anticipated that 10.4.1 sould have already made use of it.

You have to consider that the backing stores of the windows are stored directly in the the VRAM, so 256 MB are not that much.

Let's hope that Quartz 2D Extreme will be enabled by default in Mac OS X 10.4.3: more than 400 bugs are said to have been fixed.

Sorry, you're absolutely right. Couldn't screen images be compressed within the backing store, though? NVidia have had lossless compression at ratios of up to 4:1 since the 5900FX. As it is, an open Firefox window (984x713), pointed at OSNews.com, takes up 5.2MB. Assume that could be halved by texture compression, that allows you to have twelve windows in 32MB of VRAM.

I've got about 20 open now, and most of them are far less than 984x713.

It seems a little excessive. That said, I use Windows XP using the "Classic" theme, so I'm hardly MS's target audience. For those Windows users who want Exposé now, there's always TopDesk (http://www.otakusoftware.com/topdesk). With a bit of tweaking it works pretty well.

According to Ars Technica's technical article on Q2DE

(http://arstechnica.com/reviews/os/macosx-10.4.ars/14)

you should consider not only the backing stores of the windows: the rasterized bitmaps of the font glyphs and of the many GUI interface elements (which Mac OS X doesn't render in real time vectorially) are to be stored in the VRAM, too.

I know, that's why I capped my example at 32MB, I was still assuming the 64MB card I'd mentioned in my initial post.

Another thing, which I'd forgotten but which your link reminded me of, is the fact that Quartz Extreme uses a virtual memory system between VRAM and main memory for when VRAM fills up.

It looks like Microsoft has taken a more simplistic approach here. To a certain extent, I'd imagine that's because they're hobbled by backwards compatibility (neither Linux nor MacOS X still try to support software written in the mid-nineties), but I still think they could have done a better job. Frankly, from a technical standpoint, they appear to be falling behind MacOX X and - to a lesser extent - Linux/GNU/XOrg with this latest release.

This is hardly a unique situation for Microsoft, in the past they've been able to rely on their monopoly to carry them through. However this time around, they hype around things like Linux and Firefox is alerting people to the fact that they have a choice in these matters, and Apple in particular has finally got a line-up that caters for budget users. It'll be interesting how the next five years shape up.

"But how is it that MacOS X.4 can accelerate most of it's display (using Quartz Extreme) on a puny GeforceFX with 64MB of memory, when Vista needs 256MB of memory?"

Probably because you are used to crappy performance, 256MB would help quartz extreme aswell.

Probably because you are used to crappy performance, 256MB would help quartz extreme aswell.

256MB video would help just about anything.

The reason you need more video memory is to avoid the bottleneck caused by swapping video memory out to main system RAM. If you have 10 openGL apps each running simultaneously in ther own window then all their textures, shaders, buffers and other bits and bobs all need to occupy video ram. Vista basically makes all your windows behave as though they are OpenGl/direct3d apps.

Does this mean they're adding PCI Express to the requirements list?

Likely as a recommendation...as in some eye candy will not work or look poky without it. In 3 years, this will not be a big issue.

Quit whining. Your machine is plenty beefy to run Vista. It won't be as snappy as XP, but for checking your e-mail you'll think it's snappier.
The next beta will probably run very well on your machine!
You really should have held out for an NCQ drive though .

I happen to work on AIX, which has some interesting characteristics relative to Windows/Linux/Mac. For example, AIX 5.3 supports variable page sizes on a per process basis. Some applications perform better with small pages and some will large pages. Also, the AIX kernel is pageable. Parts of the kernel can be swapped out when they're not in use, and kernel pages can be distributed amongst multiple LPARs. Very neat stuff.

on a very nerdy level the AIX stuff sounds pretty damn slick.

No kidding. I mean, when you say "yea but what if you're a big image editor" well then sure you may need more memory. But just for Windows? Maybe IE with 50 web pages plus WMP with 8 movies playing; then maybe you'll need 2GB of RAM! But how many people have the screen real estate, and mental stability, to do all that at once?!

http://en.wikipedia.org/wiki/Quartz_Extreme

http://en.wikipedia.org/wiki/Windows_Presentation_Foundation

Ok, go read both of those. And please read carefully and understand:
Quartz Extreme is a texture compositor.
Avalon is a vector graphics system..

Avalon is significantly slower because it's significantly more useful...
They're comparable only at an end user level. And then it's a question of design ideas. Do you: Limit what applications can do for widget sizes or do you just let them do anything and make sure it still looks perfect on a low res LCD?

I *think* the next version of OS X is supposed to have much more similar functionality in Quartz. But IIRC they aren't going to do vector graphics, they're doing a large number of pre-rendered graphics.

That's why Mac's is so much faster. It's not doing nearly as much.

Don't forget that Vista Beta 1 comes without Aero. We'll see it only in Beta 2, so at the moment you just can't know how slow (or fast) Aero will be

Aero will be ready when Cairo will be ready because MS is copying a lot from it... really

> Don't forget that Vista Beta 1 comes without Aero. > We'll see it only in Beta 2, so at the moment you
> just can't know how slow (or fast) Aero will be

But then we are talking about what machine you need to run aero fast - and here the "hefty requirement" from the article is: you need a 128MB graphics card.

From my experience with WGF/Avalon: you cannot use it when the graphics hardware does not help you - just too much cpu usage.

That's the reason why the graphics in Vista has that XP compatible interface for older hardware or people who prefer the XP/W2K look and feel.

That's the reason why the graphics in Vista has that XP compatible interface for older hardware or people who prefer the XP/W2K look and feel.

The fact is - Why would anyone buy Vista when they can only get Windows XP in return? They may as well stick to Win2k or XP then.

Aero isn't the only feature:
Search capabilities + corrected user environment with respect to running as a proper limited user.
You know all that sudo stuff in Ubuntu and the admin password stuff in OS X? Microsoft's doing it in Vista....
Also, proper shut down algorithm's that do NOT allow program overrides. Supposedly, better hibernate support in other ways (not sure what that all means).

It's not so much amazing for 5 years work, but it's worth $200. Not worth $200 to me.. but it's worth $200 to most people .

The fact is - Why would anyone buy Vista when they can only get Windows XP in return?

On existing computers, true. There will be few upgraders beyond the initial surge of Windows advocates and people who think they must have it.

That said, for new systems, the situation is simple;

'Why would anyone buy XP when they can get Windows 2000?'

While there are plenty of fans of Windows 2000 that can't stand XP, none of them can buy a new computer with Windows 2000 on it.

The same will eventually happen with XP when Vista is out there. I give it 6 months; old stock gets purged, new stock is dual-loaded (pick XP or Vista on boot), and 6 months out XP is no longer an option unless you buy a seperate full copy of XP from the available stock.

Its getting harder to find unbloated hardware too, I don't do games so I figure the graphics memory to be 2k by 2k by 3 or 12 Mbytes max for my workstation. For 1600*1200 4Mbytes is enough at 16 bits. Any more than that I don't care about or won't pay for.

Just use older PCI cards with 1-4 Megs on it then. You can have those for free these days, and as old as they are, they just don't die, and are well supported by OS'es because they've been around for so long. Or go for the cheapest/most humble integrated graphics you can find.

PC architecture and Windows (and Linux too) architecture are all wrong, bloating faster than the hardware can keep up.

I agree with you. As the saying goes: "software expands to fill up the available hardware capacity". So true. I think it really depends on one's definition of an OS. To me, that's something like:

bare metal << low-level drivers << kernel << more or less hardware independent kernel extensions (virtual filesystem, low-level input/output API's) << libraries << middleware (high-level API's, GUI etc.) << user-visible applications << user

To me an OS consists of the "low-level drivers << kernel << kernel extensions" part. That should be small, and become even smaller as it's improved. Support for additional hardware/functions should follow a plug-in approach. To MSFT, I suppose an OS means 'everything between user and bare metal'.

See for instance how OS'es are installed on harddrive: in many cases, the user goes through configuration like language, date/time, hardware options, etc., then can sometimes edit a default set of installed components. Then a whole bunch of stuff is copied, a lot of auto-hardware detection is done, a lot of auto-configuring takes place, and one or more restarts follow to complete the process.

Modular approach: find harddrive partitions, have user pick one, copy something minimal that supports very basic safe defaults (just VGA/VESA, keyboard, USB mouse for example) onto it, configure bootloader if needed, and restart. Installing that way can be done in an instant, and if something goes wrong, there's few possible causes to investigate. Then, after succesful restart, the user can add/configure additional components and hardware support (audio, 3D video, game controllers, printers, power management, etc.). One by one if necessary, so that in case of trouble it's immediatly clear what component introduced a problem.

But hey, I guess this is ultimately a matter of 'simple but limited, works everywhere, guaranteed' vs. 'lots of things automagically supported out of the box, and a decent chance that it works on common/your hardware'. A matter of taste and convenience. Just pick what you like most.

I dunno, in my computer science classes all my programs (even 'hello world') came out to be 64kb (exactly, for some reason), on 'release' mode since we weren't taught how to use the debugging tools (just add printf statements to find your error). Of course, that was with Microsoft Visual C++ V6. Using the Bloodshed Dev-C++ program (which uses Mingw), I think they ended up more around 20kb.

And then there's Visopsys and MenuetOS, both of which I found on this website, that can STILL fit on a floppy.

Maybe it's the accelerated pace of hardware being produced that means people no longer make extreme optimizations like they used to do on Amigas and Atari STs. I mean, look at MenuetOS. It's coded in assembler, which isn't (apparently) used much any more. The hardware improves to compensate, so while Vista probably could be run at full features on 2003-era hardware, I doubt it will.

Yea, some crazy stuff. Like they've decided that managed languages have advantages that are worth the small speed cost. They decided that higher level languages that add in functionality are worth the cost in disk space. Crazy I know. We only have like 10 orders of magnitude more power and we're wasting almost a whole order of magnitude to make development more productive!

It's obvious to me that you're an expert in the software development field. Software truly is bloating at an unheard of rate; it's so bad in fact that PC turnaround times are as long as they've ever been! It's so bad that people who bought a computer to run Office 2000 on are only able to run the latest Office on it!
It's so bad that people are creating very high level managed languages which allow developers to focus on the logic more and their toolset less! And it's showing. Uptimes are up, changes in formats are down, and people are getting more done on their PC's!
Truly truly we should go back to the days when real men used assembly because they liked rewriting the same program for every system!

You realize that optimizing resources is using them right? You use them and then get out of the way when something important needs to happen....
This is why you'll see many Linux distributions with about 8 weekly cron jobs, including prelink. All of them take a whole bunch of time, but they all help speed something up majorly.

Did they really mean you should have your main RAM as DDR3? Is that even going to be available within the next two years?

Yeah, that's exactly like someone said about Apple's future - that they see the future as gadget-centered, not PC-centered, which will only act like a hub. Microsoft is really behaving childish and thoughtless here. Not to mention the monitors thing - they're just on crack I guess. Besides, if there's pure unecrypted video stream somewhere, anywhere, it CAN be decrypted.

Instead of playing with Exposée or Quartz Composer, try to work professionally with Motion 2: 64 MB of VRAM are simply not enough. The top models of PowerBooks nowadays carry 128 MB VRAM already and the Intel-based ones will come very probably with 256 MB VRAM at a minumum.

It's going to have vector widgets. Something Mac, IIRC, swore it'd never do. Mac still uses a bitmap display. Quartz, TMK, is a really kickass compositor.

This is why it works on a 32MB card that's not very fast.

You know that Cairo thing in GTK2.8? Microsoft's doing that in the graphics card. Picture this. Round buttons that look the same at 500,000 pixels^2 and 200pixels^2...

Also, they may want to do neat things to do with monitoring everything on your screen and things not on your screen. I believe Quartz only does that sort of thing when expose is activated?

IIRC the next version of OS X is supposed to have widgets done in Quartz using pre-rendered images. Which is a more efficient way to do it! It just makes a bit of a restriction on your widget sizes.

Pentium 4 630 I meant.

True, TPM isn't exactly appealing to anyone buying the software. They're thinking "but I don't want to pay to be restricted." And the response is "well if you don't you won't be able to do this and that." And then they say "but I can now." And the reply is "times change."
Then they think to themselves "man this sounds like I'm dealing with the mob or something..."

Ya know, I don't think that's fair. Here's why:
1.) Win95 was utterly unecessary in its day. You installed it so that you could have this "pretty" shell that was supposedly easy to use. You still used most of the same programs, because nobody switches the programs they use immediately. And slowly, over the next 2-4 years you started using the multi-tasking capabilities in the OS, but by the time you really needed them Win98 was out and so were fast Pentium II's.
2.) Win95 was actually pretty revolutionary in that it was a horrible set of speed hacks to get something that wasn't reasonable to happen on crap hardware. The thing was snappy, but horribly unreliable. I mean, put Win98 on a computer, then put Win95 on. You'd be amazed at the speed difference you'll see in the Windows shell.

Vista on the other hand isn't so revolutionary. It's really very evolutionary in that things don't so much change for the end user. Most of the cool stuff is additions or under the hood so to speak. Search is neat, but it doesn't mean you can't do everything the same way you used to. The new graphics will add usability, but the old usability is still going to be the same.

As for the hardware. I don't think they're saying you'll need a killer box (not for next year). They're saying you'll need a modern setup. And they're defining "modern setup" as:
A dual core/dual chip. (dual core being cheaper of course)
A graphics card with memory.
RAM - A bit of it, cause it's cheap people.
SATA 2 - Cause you might as well.

There's no reason that you'd need NCQ for Vista. And if there is, it's stupid and they shouldn't be using it as an excuse. Never yet have I heard of slower drive accesses meaning something was unusable.
More RAM? Well, it's typical to get 512 these days. So what's say 1-2GB next year? Pretty evolutionary.
Dual core? That's probably going to be almost all you can buy soon. Even Intel has admitted, implicitly, that netburst isn't viable today.
Graphics? Well it's about time! Bitmap screens are like an attempt to apply a character map to colored pixels: Not the best way to do it by far. Even graphics storage is, to some extent, heading toward using vectors. Obviously not for art and photo's, but for simple graphics. Well, widgets are pretty stinkin simple graphics!

I loathe Microsoft. But I really like a lot of the stuff I'm seeing in Vista. They seem to have left behind their days of doing things wrong to get a bit more features onto underpowered crap hardware. Now they're saying: Spend the extra $200 and we'll give you something useful!

I'm more concerned about two things:
1.) Driver stability from less than wonderful vendors. We've all had a bad graphics driver make a game play bad, or kick us out in the middle of a game. Then we used a recent driver and through that stupid driver cd away. Well, now your whole desktop will be dependent on it. But I think Vista has a fully able bitmap screen to fall back to..?
2.) Abuse of search. Search is great, until you change that document and it no longer contains the word "work". You've still gotta organize your stuff. But marketing people will try and say you don't to sell more copies..

But again. This isn't that beefy of hardware. Especially considering it's for PEEK performance. The only part of that which you don't need for PEEK XP performance is the graphics card.

People have been justifying a "supercomputer" on their desktop for years. Just look over on "/." at what happens when someone questions the utility of the latest geek toy.

Maybe when this is all over we can thank MS for giving us a desktop "supercomputer" running Linux and our kick-ass 3D interface.

Because we're certainly not going to get it waiting on everyone elses hardware to stop running Linux.

I buy the faster computer to get better load up times, better compile times, snappier interfaces, and so that I can actually do things like video and image editing.

I prefer my OS isn't the reason for the faster computer...
And Windows PC's have notoriously been built on vastly inferior hardware... So what are you blithering about anyway?

Not really... You're thinking of game distributors. I've run Windows XP on close to the base spec systems and it runs usably (barely).
I think the 2GB of RAM one of the surprising part. I'm not sure what the OS is doing to hog RAM. The OS itself should hog as little RAM as possible...

You dont realise the fact that many people on this planet do not have 500 USD to spare even in the next 10 years.

"Stop complaining, save your money and buy a new PC, you can get a new pc for around $500 nowadays"

And that PC will be absolutely useless for running Vista.

Look at those specs again, 256mb graphics card, dualcpu/dualcore cpu (so we're talking AthlonMP or intel equivilent, or better), 2GB of ram, you show me a system with all that in it for £500 let alone $500.

Any pc now that isn't bleeding edge now will run this system dog slow. And whilst those prices will indeed go down, if you run a company, can you justify the running costs of a dualcore AMD64 system, with all the trimmings, just so that your new system can run the programs your old system used (or equivilents thereof) only as fast, or slower then the previous system, whilst using maybe as much as 4 times the power?

Vista is an OS who's design is 5 years out of touch. If you hadn't noticed, cpus aren't getting that much faster any longer, they're getting more features to compensate for that such as dual cores, but even that will only get you so far, no, if you want to see real performance boosts on hardware now, its going to have to come through software, at least for the time being. And you're not going to get those boosts if your base OS has just absorbed your spangly new alienware as though it was the $500 pc you mentioned earlier.

M$ could really have innovated, started off with a whole new design from the ground up, and copied OSX with its sandbox for older software. I just see more of the same, but "prettier". I don't think consumers are going to be all that pleased at having their topend systems (from todays specs) running like a P3 450 128 ram using XP. Especially once they start seeing their electricity bills go up because of their dualcore/cpu SLI rigs power requirements.

No. The point is that by the time Vista is released, a system capable of running it smoothly will be that cheap.

What I find funny is how willing people are to abuse hardware companies, when a good number of the times, any problems that do occur, occur because of flaws in the basic architecture of the operating system.

Yes, there are dodgy hardware vendors out there who try to siphon off poor quality hardware, but the question I constantly bring forward, is how on earth FreeBSD, Linux, Solaris and all the other operating systems available on the x86, provide pretty good stability, exceeding that of Windows, on the same hardware - the very hardware that Windows fanboys claim that are the causes of Windows problems.

I ran a Dell XPS PIII 550Mhz, for around 3 years; Windows 2000, constant crashes from Explorer.exe, constant strange bugginess I couldn't put my finger on, I even did a full memory scan, which took some time, and everything was ok, did a check of the controllers, checked the connectors.

Once I found that my Matrox G550 was supported by XFree86 on FreeBSD, I instantly moved over, and compiled from the ports, KDE - a fresh breeze of stability once returned, I should have never left.

As for the solution for Microsoft - Solaris, the most scalable, stable operating system has now been opensourced, why not embrace that and build a GUI along with tools ontop of that? port all the drivers they own accross to it.

We can't run every cheap junky usb device on the shelves. We usually don't use the generally badly written commercial drivers (there's a lot more bad devices out there than good ones)..

You probably had a bad driver for your Matrox card.

How on earth does a 'bad driver' cause explorer.exe to crash?

The issue is with Windows NOT the driver; the driver itself is rock solid, just like the hardware - what the causes of the grief is the bugginess of Windows; the crap design, the unwillingness by Microsoft to accept that they got it wrong, and that they need to go back to the drawing board - and shock horror, means they have to do a MacOS X and base their next operating system on *NIX (and thus, go against 20 years of anti-UNIX rhetoric), then so be it.

Corrupting memory, for one thing.

How on earth does a 'bad driver' cause explorer.exe to crash?

Well not to be a dick but if you have no idea how a bad driver can cause explorer.exe to crash on a windows system you really have no authority to tell anyone that its the *bugginess* of windows that caused the problem.

Besides there were a few things you missed in your troubleshooting that may have honed in on the problem.

No point in me telling you what you missed or how a bad driver might crash explorer because as you stated, you already moved to another OS. Problem solved.

Cut the defenciveness; this weird behaviour occured on day one, and continued right till Sp4 installed - WHQL certified drivers, the lot.

If there was an issue with the driver, it would have atleast been fixed within the three years; all it tells me is that Microsoft would rather spend time souping up the GUI by adding unnecessary and confusing changes in the name of so-called "progress".

The best example of this change is in control panel in Windows XP, and the change to catagories - sorry, I worked at a help desk, and believe me, the first question I would get asked is, "how do I turn it back to my old Windows? I don't know where anything is?"

You sounded sane till this part...

M$ could really have innovated

Please drop the M$ and similar childness. It just makes you look like an idiot.

I have said it before, and I'll say it again...I WILL NEVER USE VISTA.

Why bother posting then ?

I highly doubt the specs put forth will be whats needed but seriously, if its an issue then run another OS people!

MS can't win with a good chunk of people. If they came out tomorrow and said they were scraping the entire OS and starting over people would "see they have a POS so they need to rewrite it"

If they work new software layers on what is there then people complain that they did not start over.

If the specs were low, people would claim there was nothing new and switch the debate to "it still costs X amount of money! I want it all free!"

They can't win for trying so if ubuntu works for some of you (its probably the best Linux Distro imho) then run it and be happy.

The short story is stop complaining. If the state of operating systems is so bad that you are that pissed off about it then get off your @ss and write a new OS.

I don't understand why it can't be done in disk driver?

It could...requiring that the CPU be tied up for the reorder. If the drive will do it anyway, you get that chunk of processing power -- and no impact on other apps -- for 'free'.

NCQ has been added to most current SATA drives these days. Look for it.

"I don't understand why it can't be done in disk driver?"

In the old days it could have been done because the disk geometry was easy enough to figure out.

With modern drives it's difficult (Even if possible) to find the real physical geometry and the drive screws everything up by silently mapping replacement sectors from reserved tracks when it detects a bad sector (All drives come with bad sectors). That combination destroys any chance the driver writer might have to order operations more efficiently based on geometry.