Windows 8-Rumors and Announcements

[hym 358]

The ultra-cumbersome and overused Registry needs to be replaced or become history.

Microsoft has already done this. Starting with Windows Vista, the Registry is a two-tiered system. There is a Windows-wide Registry that can only be accessed by Windows and a select few Redmond-certified software vendors. Everything else is forced to write to a new "registry" that is tied to the user account. You can scramble that "registry" all you want and it does nothing to the main Windows Registry.

[A Nonny Moose 8,262]

The ultra-cumbersome and overused Registry needs to be replaced or become history.

Microsoft has already done this. Starting with Windows Vista, the Registry is a two-tiered system. There is a Windows-wide Registry that can only be accessed by Windows and a select few Redmond-certified software vendors. Everything else is forced to write to a new "registry" that is tied to the user account. You can scramble that "registry" all you want and it does nothing to the main Windows Registry.

I haven't used anything since XP. What they have done is a start, but the "user registry" should be discontinued a.s.a.p. If applications want to keep initialization files, let them create one with the standard extension .rc

[jacksunny 741]

The latest leaked build hints at improved mulit-moniter support. source

[A Nonny Moose 8,262]

The latest leaked build hints at improved mulit-moniter support. source

Phooey. Just supports hardware manufacturers.

How about multi-desktop support. I can switch desktops and the running programs don't even realize that they've lost focus. Seems to be an X11 feature.

[hummer0328 6]

The latest leaked build also seems to use a new filesystem, something codenamed: Protogon. Perhaps, it's an updated form of WinFS, the fs that was originally slated for Vista. Apparently, it is content aware, a bit like XFS, and HDD placement is sorted and based on file type, keeping similar files close together. MS hasn't released any info, so all this is pretty much speculation at this point.

As far as the Windows Registry goes, it's now very similar to the way Linux configuration files work. The system registry is equivalent to the /etc folder (owned by root:root, everybody can read, only root can write). The user registry is similar to the hidden /home/user/.* folders in Linux (owned by user:user). There are far more differences than similarities though. MS just needs to learn the value of KISS (Keep It Simple Stupid). That would help them quite a bit.

[jacksunny 741]

Microsoft has revealed that Xbox LIVE will be built in to Windows 8. link. I hope this is an optional feature as it will be useless for those of us who don't even have an Xbox.

[A Nonny Moose 8,262]

Now, if only their revised file system can avoid checker boarding the disks, they may have got somewhere. The idea of a file system being content aware is yet another mistake. Extensions are garbage as far as I am concerned, and only useful as a visual aid. The O/S can easily look at the first sector of a file to find out what it is without any symbolic guff. One of these days, Microsoft will get out of kindergarten. Maybe someone should give them a subscription to Software Practice and Experience.

Having an Xbox driver is just another example of Microsoft's tendency to incest. One hopes it is plug and play, because there is enough junk in their kernel now. Sounds like yet another mistake driven by the marketing division.

Saw a commercial for a "new" Intel chip last night and was surprised they spent the money on a TV ad. Is this the opening shot for needing an I5 processor to run Windows 8?

  Edited June 11, 2011 by A Nonny Moose  

[hym 358]

Microsoft has revealed that Xbox LIVE will be built in to Windows 8. link. I hope this is an optional feature as it will be useless for those of us who don't even have an Xbox.

It's not simply about whether you own an Xbox. It's part of a larger strategy to integrate Windows-based games with the Xbox Live system so you can play a game on your Xbox and then turn around and play another game on your PC and the Xbox Live service is aware of both. Makes it possible for people who don't want to buy an Xbox to play online with their Xbox-using friends. (At least I am assuming that is theoretically possible under this system.)

[Alejandro24 575]

...

Saw a commercial for a "new" Intel chip last night and was surprised they spent the money on a TV ad. Is this the opening shot for needing an I5 processor to run Windows 8?

This remember me an article that says the next version of Windows, Win8, will not be compatible at all with the ARM architecture of Intel precessors.

[A Nonny Moose 8,262]

...

Saw a commercial for a "new" Intel chip last night and was surprised they spent the money on a TV ad. Is this the opening shot for needing an I5 processor to run Windows 8?

This remember me an article that says the next version of Windows, Win8, will not be compatible at all with the ARM architecture of Intel precessors.

Now this might just be the final event in the suicide of Microsoft, if true. Sooner or later people are going to get sick of pay and pay. New hardware to go with new O/S. Sucks.

[hym 358]

Are you sure about that? I just pulled this bit of info from another website:

Microsoft is making sure ARM Powered Windows 8 works exactly like on x86.

I find it hard to believe that MS would allow the software developers to shoot the company in the foot that much.

[Shadow Assassin 336]

Having an Xbox driver is just another example of Microsoft's tendency to incest. One hopes it is plug and play, because there is enough junk in their kernel now. Sounds like yet another mistake driven by the marketing division.

Like just about every other driver (aside from the core drivers) it probably will be optional. You don't need to install it if you don't want it, Microsoft have traditionally been pretty good about that sort of thing. If you want it, install it via Windows Update... and every gamer will probably end up installing it anyway if they play any recent game.

Besides, a business workstation with the Xbox driver? It's just not needed.

In any case, I'm sure this will be a supplement to DirectX.

[A Nonny Moose 8,262]

I have always believed that RISC processors give up a lot to CISC processors because it is like putting smaller tires on your car. You have to run the engine at much higher revs to go the same distance. And unless you adjust your speedometer it will lie like a rug.

I really haven't followed this much since I retired, but can some one who is current please summarize the current state of things? What are the current standard benchmarks showing for the two competitive concepts?

The last RISC processor I knowingly used was a DEC PDP-8 which had only 8 instructions. In the old days, this was real challenge, and you had to write either in machine code or assembler (PAL). Of course, with the advent of the microprocessor this machine has faded into the mists of pre-history. Working with one of these machines really makes you think algorithmically. It had only one arithmetic instruction and the store instruction cleared the register. For example, to multiply by ten you loaded the value, shifted it left 2 bits then added it in twice. To do a decimal arithmetic add you created an excess six addition routine. Guess why I prefer CISC machines? Cost is no longer a factor in hardware, eh?

[Ocram's Razr 964]

I was born in 1991 AD, I grew up with technology around me and Windows 95 was released before I was old enough to use a computer. I like using computers but I am no expert and I have thoroughly decided that I prefer to learn Chemistry (biochemistry, biology, and botany) far more than computer programming. I know enough to fix or upgrade my computer and compile open source C++ games. That is all I want to know.

[hym 358]

I have always believed that RISC processors give up a lot to CISC processors because it is like putting smaller tires on your car. You have to run the engine at much higher revs to go the same distance. And unless you adjust your speedometer it will lie like a rug.

There was a very practical reason for the adoption of RISC design philosophies. Most compilers at the time couldn't take advantage of what CISC offered, and RISC design theory stated that you could achieve greater speed if you could simplify things enough that you could run them faster than CISC processors were capable of performing.

I really haven't followed this much since I retired, but can some one who is current please summarize the current state of things? What are the current standard benchmarks showing for the two competitive concepts?

Realistically, CPU design has advanced to the point that the RISC vs. CISC debate is pretty much pointless.

[A Nonny Moose 8,262]

I always felt it had died the proper death some time ago, and am glad to note you also think so. Of course, compilers now can indeed optimize for any given processor. The days of the simple context free language has not only arrived, but code generators are now surpassingly good. The days of easy use of YACC and LEX have passed and the newer versions of the same idea have grown up.

One of the internal developments at GE when I was with them was a program called MALT or Multidecor Assembly Language Translator. You could hand it the architecture of a processor and it would generate an assembler for it. Very much like NELIAC, the Naval Electronics Laboratory International Algol Compiler, which could be made to accept ALGOL programs in any language. The demo was in Hawaiian with ALOHA for BEGIN and PAU for END, and so on. It was an interesting toy.

I think the only non-context free language left is COBOL XX, which has grown up considerably from COBOL 60. It now quite happily handles OOP, and most data base interfaces are either in COBOL or PL/1 where a string slinging language is needed. PL/1 was a spectacular failure after the death of the Multics project, even though IBM tried to keep it alive. There is just too much COBOL code in the world that has to be maintained and upgraded for it to go away. However, writing a COBOL compiler is akin to witchcraft.

When I was working at the U. of Waterloo in the 1970 time frame, COBOL programming was in the Arts faculty, and all other languages were in the Math faculty.

More and more users are getting to the point of treating their computers like a home appliance, and rightly so. In fact, most home appliances these days have embedded computers that the consumer doesn't need to be aware of. Our friend in Seattle is completely correct. Even for a scientist these days, programming is out. It has become a specialist trade, and the do-it-yourself phase of computing is going the way of the Dodo.

  Edited June 14, 2011 by A Nonny Moose  

[jacksunny 741]

A Microsoft VP has hinted at an August 2012 release date. link

[A Nonny Moose 8,262]

A Microsoft VP has hinted at an August 2012 release date. link

I see. Just enough lead time for Christmas presents.

[jacksunny 741]

Build 7989 has been leaked and has the beta fish on the boot screen indicating that the beta release could be soon!!!

[A Nonny Moose 8,262]

For the truly adventurous, a beta release of W8 should help to shake the bugs out of this box, but don't bet that the geeks who play with it do ordinary things. The ordinary bugs will still be there on the release.

[hummer0328 6]

Now MS is touting 20 sec cold boot times. It would sound impressive, if it wasn't just a form of hibernation that closes all non-essential programs and then writes a limited system image to a swap file/partition. Honestly, that isn't anything to be bragging about.

My system resumes from true hibernation faster than that, and that's usually with a browser, terminal, music player, etc, open. Actually, my system almost does a true cold boot just as fast (22 sec). That's from hitting the on button, sitting thru approx 7sec of my BIOS post --> very responsive desktop. And that's with loading all 400MB of my browsers catch/profile, 27MB of DNS cache, + 500MB of very aggressive system/app cache all into RAM. It makes for an insanely responsive system. Most people swear I'm using a solid state drive. No, just a standard 5400rpm HDD.

Of course, the time I will save will never be greater than or equal to the time I spent customizing my system. Custom and highly minimalized kernels don't compile themselves.

But anyways, reports say Win8 beta runs faster Win7 does, on the very same computer. So that sounds promising!

  Edited June 22, 2011 by hummer0328  

[A Nonny Moose 8,262]

You certainly have a right to happy with our response with that customization. However, some of us, me for example, don't have the high anxiety to make it worth the bother. I run with all the junk that UBUNTU 11.04 drops on you together with a dual boot scenario. Since my production system is not the first on GRUB's list at the moment, I have to manually select it, which really extends the boot time. All in all, I guess it take a couple of minutes to get up. It is certainly long enough after GRUB for me to go get a cup of coffee. Then I have a rather extensive list of applications that start by default. Last time I tried a hybernate it failed, but that was on a previous verson. So, I guess I'll try it again.

How's the cold start for your server? I'll bet it has to do a few handsprings to get going with all the SELinux stuff to set up. I don't bother with ACLs and CACLs because I have a well-isolated single-user machine, and stuff like remote login is turned off. I haven't had any activity on my firewall nor my virus checker except when called to look at e-mail attachments I didn't trust, since I've been on Linux. I suspect the hacking community isn't really interested in our little niche very much. Do you log attacks on your server?

I still get some SPAM, but my SPAM filters are a little overaggressive, and often flag stuff that I have to unflag. And occasionally I get some junk that gets through once. I set a specialized filter to get those. It is easy with Thunderbird to create a custom filter, and I just do it when needed. I never see these guys again. My ISP Internet guy tells me that they get regular attacks and usually stop most of them. Recently there were a few e-mails that came in disguised as being from them, and I called them of course. We had a nice talk, and now when one of these comes in I expose the full header, and you can usually find out that the last proxy is not even in the country. Most of these guys just are not good enough to hide their total footprint.

Well, back to microshaft: If they are faster it is because they have cleaned out a lot of fat. I wonder if they will include their new file system on this release. Anything would be better that the New Technology File System (NTFS). It is just a grown up FAT system, and it is, indeed, fat and sloppy. Whatever they do, they will still have to support the old structure for a few years. The Linux system quite happily supports several competing systems, and I gather you can have several of them in the same tree operating seamlessly.

I wonder what their actual cold boot time is on initial installation after everything is loaded?

[hummer0328 6]

All in all, I guess it take a couple of minutes to get up. It is certainly long enough after GRUB for me to go get a cup of coffee.

Hopefully that should be fixed by 11.10 with Systemd replacing the current init process. By itself systemd cut my boot time (grub --> gdm) from 19 sec down to 13.

Last time I tried a hybernate it failed, but that was on a previous verson. So, I guess I'll try it again.

Yeah, I always had hibernation issues with Ubuntu as well. I just started using Tuxonice (kernel patch) that adds tons of features to hibernation and uses LZO compression. Wish Ubuntu would add something similar to their kernel.

Edit: try adding the resume hook to your grub. best way is: resume=/dev/sdaX (or whatever your swap partition happens to be). If using grub2:

[code]# nano /etc/default/grub #or vim[/code]

find the line:

[code]GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"[/code]

and add the resume argument:

[code]GRUB_CMDLINE_LINUX_DEFAULT="quiet splash resume=/dev/sda2" #sda2 in this case[/code]

save, and of course, if using ubuntu, don't forget to:

[code]# update-grub[/code]

it should be in the kernel by default, but sometimes you have to force it like this. If that doesn't work, then it's most likely a hardware incompatibility.

How's the cold start for your server? I'll bet it has to do a few handsprings to get going with all the SELinux stuff to set up.

It's not as bad as you might think, since it doesn't have to load a desktop environment. But I rarely reboot it anyways.

Do you log attacks on your server?

Sure do. Not that they got very far. Made me glad I had good security measures in place though.

The Linux system quite happily supports several competing systems, and I gather you can have several of them in the same tree operating seamlessly.

ext2,ext4,reiserfs,xfs all running on the same system , just check out my fstab

[code]devpts /dev/pts devpts defaults,noatime 0 0 shm /dev/shm tmpfs noatime,nodev,nosuid 0 0 tmpfs /tmp/.chromium tmpfs defaults,noatime,nodev,nosuid,mode=1755 0 0 tmpfs /tmp/.pdnsd tmpfs defaults,noatime,nodev,nosuid,mode=1755 0 0 tmpfs /tmp tmpfs defaults,noatime,nodev,nosuid,mode=1777,size=2G 0 0 /dev/sda4 / ext4 defaults,noatime 0 1 /dev/sda5 /boot ext2 defaults,noatime 0 1 /dev/sda6 /var reiserfs defaults,noatime,data=writeback,notail 0 1 /dev/sda2 swap swap defaults,noatime 0 0 /dev/sda3 /media/Files xfs noatime,logbufs=8 0 2[/code]

I wonder what their actual cold boot time is on initial installation after everything is loaded?

From what I've read, similar to Win 7. And I guess the new touch interface in completely optional and targeted towards the touch device market, with the normal desktop interface resembling the Aero interface. So they aren't starting from the ground up, for the most part. Maybe MS learned from the mistake that was Vista after all.

My guess is MS never stopped development of WinFS, but decided it was too much of an alpha version to be released with Win7 beta (aka vista). Protogon seems to be faster than NTFS in real world throughput (not that thats hard), possible live defrag, very content aware, and the FS takes up less space. My guess is that Protogon is the main reason beta testers are reporting that Win8 feels faster than Win 7. But I guess only time will tell

  Edited June 22, 2011 by hummer0328  

[A Nonny Moose 8,262]

Thanks muchly.

I tried hibernate on this system, and it worked quite well except for the GRUB intervention due to my dual boot situation. With your help, I can now avoid that. No matter how long you've been in the cave, the system comes back up in your screen saver. They really want a password and I am not averse to this. It is much quicker than a cold boot.

I'll start testing 10.10 when it gets to about the beta 2 stage.

There has been a fix in fglrx. I can now run SC4 in full screen with hardware or software rendering which earlier caused a system lock up. I had to use the kernel commands to recover from one of those. So far I've had only one hiccup that made the i/o system stop responding, and I haven't been able to reproduce it, so I'll keep trying. When the normal keyboard and mouse drivers quit, you are really in a rare problem. So I leave the kernel stuff enabled.

Wonder if we should start a Linux thread?

[hummer0328 6]

There has been a fix in fglrx. I can now run SC4 in full screen with hardware or software rendering which earlier caused a system lock up. I had to use the kernel commands to recover from one of those. So far I've had only one hiccup that made the i/o system stop responding, and I haven't been able to reproduce it, so I'll keep trying. When the normal keyboard and mouse drivers quit, you are really in a rare problem. So I leave the kernel stuff enabled.

Huh, that sounds odd. Never had a problem like that before, well one that wasn't my fault. Although, 2.6.39 seems a bit unstable at the moment, as I've gotten a couple of random kernel panics during boot that disappeared with a subsequent reboot. I've noticed it sometimes hangs for about a min during shutdown as well. My guess would be bugs in an attempt to repair the notorious power consumption kernel regression. But never a problem like that.

Say what you will about intel's integrated GPUs and their <sarcasm>amazing OSS drivers</sarcasm>, but at least they work out of the box on linux with little to no configuration. I just spent a day trying to help a friend with his new laptop and ATI GPU. WOW! What a headache that was. Kernel Mode Settings were causing glitches, so we had to turn that off and configure everything manually.

Wonder if we should start a Linux thread?

What, and stop hijacking other peoples threads? Yeah, might not be a bad idea!

  Edited June 23, 2011 by hummer0328  

[jacksunny 741]

The latest rumour is that Windows 8 could be released as early as April 2012.

[jacksunny 741]

Steve Balmer will release the public beta version of Windows 8 at CES 2012. My link

[A Nonny Moose 8,262]

Steve Balmer will release the public beta version of Windows 8 at CES 2012. My link

Deadlines like that are driven by marketing. That means the box of bugs has to be firm by the end of the previous month. Rushing software out the door, and all the associated hype, is what makes for immediate patches after release. If I were going to run that O/S I would wait for SP1 at least.

You can't debug a program by throwing a mob at it any more than you can have a baby in one month by throwing nine women at it. They just get in each other's way, the Librarian can't keep up, and they start trading fixes without going through the procedures. When in danger or in doubt, run in circles, scream, and shout.

The Librarian (for want of a better name) is the coordinator who does QA and adds any proven changes to the product. No matter what you call it now, the Chief Programmer Team concept works.

[Munchiez 74]

Sounds interesting. I however would like to stick to Macintosh. I will look into W8, I started to laugh becuase I just realized my friend bought a new computer with windows 7, pretty funny.

[A Nonny Moose 8,262]

Sounds interesting. I however would like to stick to Macintosh. I will look into W8, I started to laugh becuase I just realized my friend bought a new computer with windows 7, pretty funny.

Don't laugh. He has a system that mostly works. I feel sorry for the guys who get stuck with Windows 8's initial release. Remember, they are still talking beta in 2012. Reminds me of the young lady who was married three times and was still intact.