Tablets have industrial uses in specific vertical markets. They just haven't managed to break out into the general consciousness of a “mass market” until recently. The problem is that the digital hygiene product is not ready for industrial use. If it breaks when you drop it, its not ready for Industrial use. For most people its a toy just like its older sibling.

A couple of years ago, my son “rescued” a Transmeta based COMPAQ tablet from a recycling center. It has XP Pro on it and runs fine for about 2 hours on its new battery. When we put on the AC supply, it immediately booted up. I ran chkdsk and it booted every time there-after. Somebody had thrown it out. There must be something wrong with it?

My employer bought a bunch of used/refurbished tablet computers a few years ago and they promptly expired after a few weeks of operation. The ones that didn't, had major battery issues. The vendor only had the tablets, not the battery packs to go in them. When the physically flat NiMH batteries went electrically flat and failed, usually while charging, they would expand like plumped pillows! That put lots of pressure on the plastic box they were housed in and usually broke the plastic. Instant dead computer.

At least all of them had batteries that could be replaced by normal human users without any special tools or techniques.

Most of them did have Ethernet connectors and ran 100 Megabits/sec.

One thing all of these tablets shared, besides bad battery technology, was the lack of; a camera, a microphone and speakers, USB 2.0 connections, and built-in wireless Ethernet radios. Sound familiar?

There has been a crescendo of fruit-flavored sycophantic propaganda pouring forth from all the media outlets and I can't wait till it all dies down. Maybe then people will see what a crappy piece of hardware it is on just the specifications alone.

Tablets are great for certain uses. UPS and other freight companies probably couldn't exist without them. They promise to help revolutionize medical care in hospitals. Warehouses likewise are big tablet venues.

I work in an industry where tablets could provide a special “user experience” to crib a phrase from the bigger California gorilla. What I'm fearful of is that the fruit flavor of the week will rot and put off other potential OEMs from building their smarter and better functioning tablet devices.

You can return the disks you've rented in any other redbox. The movie disks have 2D barcodes printed on them so inventory control is automatic, no clerk required, practically foolproof operation.

There is even an applet you can download onto your i-Whatever to locate and reserve a copy of the movie you want.

As faster Internet access becomes more and more mainstream here, the utility of redbox is going to become much smaller. Most of the big rental stores have retreated under the onslaught of on-line Netflix and on-demand cable and satellite connected DVRs.

Obviously, the machines have some sort of Internet connection and therein lies the issue.

I have to wonder how difficult it is to embed a snooping, credit card reading, pile of malware into one of the redboxes? The systems are obviously PC or at least ARM based. Does the managing company maintain any sort of security surveillance on the systems?

Just something to think about the next time you swipe your credit card through the reader slot.

The first change is that it looks like the Patent Office is going to open up the patent approval process in a pilot program by allowing for a sort of peer review using evidence of prior art or prior patents to be submitted to disqualify a current patent application filing. http://dotank.nyls.edu/communitypatent/ http://www.peertopatent.org/getting_started http://www.uspto.gov/web/patents/peerpriorartpilot/

One of the articles listed above cites the fact admitted to by the USPTO is that there are approximately 1.2 Million patent applications pending. With 6000 inspectors, that translates to 200 applications each. Assuming that 18 hours per application (18 to 20 hours, cited by the USPTO) is required, that translates to 3600 hours each inspector. Assuming that each inspector gets 2 weeks of vacation each year, that translates into nearly 2 years' work by every inspector just to clear away what applications currently exist. It can be assumed that more applications keep coming in. Sounds more like a government jobs program.

The second development is that certain non-profit organizations have taken on activist postures and have filed challenges to patents based on documented prior art. In many cases the prior art cited is previous patents! There have been many patents challenged in the past but usually by corporations with a vested interest in the outcome. These challenges are filed to protect a patent they already own or to prevent a competitor from securing a prior dated patent invalidating a current patent application. http://w2.eff.org/patent

It seems that corporate IP owners think that patenting minor modifications or marginally innovative changes to products is a means of protecting their R & D investment. It would be interesting to do a cost versus profit analysis to determine if most patents are worth the money it costs to file them, to defend them in the application stage or later in the courts.

Following recent news stories concerning a number of software patents, especially in the last few years, it can't be possible that even 50% of the current software patents are worth the effort it takes just to file them. Considering that most algorithms implemented in software are simply just another way to do something already being done, even if not as efficiently, the patented process can be simply programmed around.

The obvious advantage for open source immediately jumps out for attention. If you don't patent the software in the first place, you don't have to pay a lawyer, his staff and the government to “protect” your software. Open source automatically operates like a peer review so the software will get progressively better as time goes on, assuming that the project remains active.

If the software is part of a product and if the product is successful, the software will likely still be of interest to other open-source programmers. Other programmers' contributions to your open source code become much like having multiple on-staff programmers. Your product gets improved for little to no additional costs. In either case, a successful product or not, no profit is lost to the leech-lawyer and to patent fees!

So if the proprietary manufacturers (and you know who you are) won't submit to peer review of their code, maybe we can force them to submit to peer review when they try to patent the stuff!

In addition there is a review here on reviews.ZDNet.co.uk of one of the items in the Electronic Design article. What really tweaks, is that a few weeks ago the Pogoplug device looked lame. After some thought and some research, now it looks like a pretty good idea. Using USB 2.0 or eSATA connected storage devices allows the NAS device manufacturer to stay out of the low-margin flash or hard drive business.

There are a number of differing implementations of Home-NAS-Server devices. One thing a lot of them have in common is a Linux or other open-source OS in the firmware. That becomes especially relevant when you compare the retail cost of the entire device (ranging from $69 to $149) with the cost of a Windows XP Home license ($89) or of a Windows Home Server license ($99). For practically the same price you can get the entire NAS-Server device based on open-source. Without open-source, the home NAS-multi-media-Server product would not be price competitive.

These low-cost devices take advantage of SOC (systems on a chip) designs to drop the price of network storage controllers way down low for appliance-type usage. For prior experience, think about how SOHO wired routers became wireless, became DHCP servers, added QoS, full NAT, and in-depth and state packet inspection firewalls, the whole time getter smaller and more integrated into fewer integrated circuit packages. Another example are all the low cost DVD players on the market.

The one firmware component or feature that needs to be added to the NAS-servers is some sort of file encryption to prevent hacks that bypass the firewall and end up landing straight on the WiFi-enabled NAS-Server. Some of the current devices might already be offering encrypted file storage. Most of the NAS devices implement NFS and CIFS to enable support for Linux and Windows clients. Encryption of the file can be handled by the OS but will likely limit the use to single users of one or the other OS and not allow for intra-operable usage or sharing.

Overall, these home/SOHO/small-business NAS devices are a fairly broad product category that will show more relevance especially as digital multi-media reaches complete saturation in the major world markets.

At work I have a DELL 755 (a desktop and on the Vista WHL) Core 2 Duo CPU with 4 GB of RAM running Vista and later Vista SP1 and it never ran right. Everything from browsing the hard drive to trying to navigate to a shared drive on the network was a painfully slow process. A lot of the time, the screen would go into a "white-out" and things would appear to come to a complete stop. Extremely annoying.

I dumped Vista and put Ubuntu 9.04 and later 9.1 on the exact same DELL with NO hardware changes and it has never even hiccuped once. It has run without ANY failures and only has required one reboot (outside of new kernel installs)in the year since I installed Ubuntu 9.04.

I asked some more questions and the chip sets are the same, the CPU clock speeds were within 10%. The only real differences were the motherboards obviously are different since they are in differing boxes and the power management settings were most likely different.

I've run into this before. Laptop users report no troubles (or at least very few problems) and desktop users hate Visaster. I have had some laptop users complain about Vista but when you look under the keyboard, the chipset is not a more recent model or the amount of RAM is less than 3GB.

So what's the difference? Laptops are the darlings of software programmers. Every programmer has to have the latest and greatest laptop made. A lot of them insist on Apple laptops. Is it possible that the OS Vista is predestined to run on the fastest Intel CPU and chip set available 4 to 5 years ago? What was running in all those laptops?

By the time the final builds got to the hardware testing teams, my bet is that there wasn't enough time to really subject the Visaster installs to really rigorous testing because the RTM date was too close. The focus was on laptop testing because most users are now operating laptops.

I also tried using VMWare's Workstation. Ease of use slightly favored VirtualBox although the feature set of VMWare Workstation is superior. For my little investigations either would work well enough.

Looking at all the desktop Linux distributions, I was interested in what special features or application focus in the distribution would enable or enhance certain applications or purposes. For instance, which Linux distros supported multi-media better than others? Would running a netbook version of a distro operate more efficiently in a VM than the full desktop version? Would a server version offer even better operation?

At home there's a new-to-me refurbished DELL GX620 with a Pentium D CPU and 4GB of RAM. Ubuntu and Windows operating systems reside on a 320 GB SATA drive. Storage is handled with a 500GB internal SATA drive and a 1Terabyte external USB2 drive. Currently the system dual-boots with grub into Ubuntu 9.1(default) with Windows XP Pro SP3 as the alternate.

I'm trying to figure out if there's any purpose or advantage to using a virtual machine architecture for a home server system with a heterogeneous mix of Windows and Linux clients. Lots of advertising being done by Microsoft, Novell, IBM and others make big deals out of supporting mixed systems usually with virtual servers. So far the only place it makes much sense are those situations where a server with little client traffic would otherwise sit idle a lot of the time. Or perhaps a VM server could be used to maintain some legacy application used by very few.

Intuitively there are some other potential advantages:

Increased security for the VM client is perhaps not so obvious. It should be possible to write firewall rules on both the host OS and the client VM to prevent any unwanted IP traffic in either direction. Remote log-ons could be forced to operate entirely within a VM without direct access to the host system. Virtual machines running applications that should not be allowed Internet access can be prevented from getting through the firewall yet still have access to the “local” workgroup.

Backup services could be handled by virtual machines configured specifically for each client. These backup servers could be run at any convenient time for the client instead of a block of time late at night. A small application can tell the server when the screen saver is running indicating an idle time and picking up where the backup last left off.

Print services although typically handled by shared printer services could be managed through virtual machines native to the client OS. This would allow the printer users to be printer administrators in the VM without giving them access to printer drivers or operating system privileges on the host OS.

Currently my experience shows that VM servers are in play at work since certain internal web servers or applications seem to always require a bit of “wake-up” time to get moving in the morning. It usually takes a few seconds, usually less than 10 for some applications to “go live”. If there is a interaction between the VM and an external client running on another system, the wake-up time will need to be dealt with by the client. Maybe use a software ping to alert the host OS to wake-up the VM, then wait and come back with an active function request.

Some things I've learned so far:

The nice thing about using desktop Linux distributions is that most of the desktop distros have related server versions. So you can go gui or not-gui for the specific VM. Try out the application with the graphical interface (if there is one) and later deploy it with the command line or server version of the application. Although a server version is optimized for typical server functionality, you're building something in the VM that is going to be much more function specific than the usual server distro.

At least for me, using Synaptic Package Manager instead of apt-get on the command line is a lot easier when starting with a full desktop image. Every time you remove something from the VM being prepped, you can more easily see what's left in the gui window.

A lot of times its easier to start with everything in the operating system image (usually the desktop image is fully loaded) and work backwards deleting the unnecessary items (actually un-installing them) until you've broken something. Then the answer is to put that one item back you've just deleted, you're done.

The alternative, assembling system functions from a blank slate, is that you start with an operating system image that might not work to begin with. You stumble your way through until you have added all the required dependencies just to get the system up and running.

Either way eventually you strip or build the operating system in each VM to the absolute bare essentials for the specific application or application group it will support.

Obviously the kernel needs to be in the OS image being built or minimized.

Generally the chosen Linux or Windows operating system will have a generic set of drivers for the virtual hardware to be supported by the VM. Hardware devices you want to “remove” needs to be turned off in the VM before the operating system software is installed. That doesn't guarantee that the drivers aren't installed, it just means the device will be disabled. An example is that Windows XP Pro will always install the floppy disk driver even when there isn't a floppy drive in the system and its been turned off in the BIOS.

Usually each VM will require at least some TCP/IP networking to communicate with or through the host. Each VM should have its own firewall rules.

An alternative to networking is to set the host up with storage that is shared with all the virtual machines. For the Linux VMs put the /home/username folders on it. Share the same folders as Everyone shares for the Windows variants virtual machines.

So far the best thing I've learned is that a bunch of VMs are the best way to try out different Linux versions.

I'm trying to become reasonably proficient in yet another programming language.

I've been writing an automated configuration program only to have the “customer” completely change his mind when the software was ready to be delivered. Now he wants to name the computers in a completely different manner. The program specifications now requires the user to be able to read AND type at the same time. Something I've never suspected possible by some of the field technicians, at least at the same time.

All of this requires relatively unrestricted Internet access.

The phrase “at the same time” has come up quite a few times in conversations, television and website advertising and of course blogs. Being on the trailing edge of technology at home has helped me ignore a lot of the commercial drivel out there. Ignorance is bliss. I was totally unaware that AT&T iPhones were unable to web browse and maintain a phone conversation at the same time and Android phones can do it. (Or do I have that turned around?) Big whoop-de-do. Obviously I don't own either.

Actually I'm not really sure why someone would want to do both simultaneously. Outside of jacking in a Borg-ish headset or going to speaker-phone and treating your nearby humans to your potentially private conversation, I'm not sure how you could actually do that physically on a small handheld device. The phone companies love it I suspect since they can charge you twice for the same airtime-minutes. Throw in texting and they can charge you three times. I know. I get the bills.

Simultaneous operations is a server's raison d'tre. Of course my 2 kids use the Internet at the same time in their rooms. And of course they watch or listen to 2 different websites, never the same thing, of course. However, I've noticed when I remove my earplugs, that they seem to at least listen to similar cuts of music over and over again and watch the same programs, if not the exact same episodes. The network connection gets maxed out quite often. A formula that a proxy server with storage might help relieve, since the old man has to share the connection with these 2 local mass-media black-holes.

Since they seem to often watch episodes in the same shows, I could possibly setup some sort of “subscription” and download the shows ahead of time when the network connections are semi-idle, midnight to next-day noon on weekends and 2 or 3 in the afternoon before they get home on weekdays since they're teenagers. That leaves the network connection more open when I want unrestricted access.

This is going to take some research. Proxy servers in the past made note of webpages visited and kept copies of those pages. I'm not sure how to trigger a download of a program that hasn't been seen previously but is accessible through the same page visited repeatedly. Call it a pre-proxy. Yes there are DVR type devices out there but its not just video I need to grab and temporarily store. Although I've seen predictable media behavior, they are teenagers.

* pronounced like "hmmmm".

The Quebec CTP 2011 is Windows 7 Embedded and I've been working with it trying to determine if there is anything my company might gain by using it instead of Windows XP Embedded. Not much.

However if Microsoft wanted to truly create something worth buying, they might consider the idea of merging Windows 7 Ultimate-Whatever with the Embedded product and allow customers the option of “building” or compiling images using the Quebec infrastructure. Combine that with a “license” to build up to 5 images, the customer then could put as much or as little as wanted into those 5 systems.

If Microsoft really wanted to make things work well for their stockholders, they would put it on a subscription model, allow customers to download upgrades for a fee to their licensed systems.

The Quebec ISO is practically a live DVD image. It will install itself and put what the customer wants on the system compile. It calls home to the mothership and gets registered. It requires an Internet connection.

Yes the customer of this particular version will need to be more knowledgeable than the typical Joe Sixpack BUT Microsoft does has a fairly high percentage of those since Windows has been around for a long time. I could see it being an attractive package for Windows wonks. It might even boost sales a bit by appealing to the techno-snobs and Windows fanbois. They'll just have to have it.

(You know if you ignore the word Windows, it sounds a lot like Linux, if you get rid of the stockholders.)

I have tried installing 9.1 using two methods. One was the on-line in-place upgrade from 9.04 to 9.1. The other was a full install ISO image. Neither method worked, both crashed on the next boot after installation.

At first I thought it is was an issue related to grub but that wasn't the problem. Increasing RAM size also wasn't the issue. The original 512 MB memory was twice as much as required minimum. One GB of RAM didn't help resolve the issue.

The really annoying thing is that a trashed install of 9.1 kills the previous installs like 9.04 on the same disk. Even selecting previous installs or the recovery options crash.

I suspect the Intel 845 chipset is the issue, in that I've installed Ubuntu 9.1 on later model chipsets and not had any problems. I've backed up a few steps, wiped the drive and re-installed 9.04 on it and everything is running again as well as it was 2 or 3 weeks ago. This particular system is my “server”. Lots of backed-up files on large hard drives. It cannot be unreliable.

The primary reason I have invested so much time and energy into Linux, Ubuntu in particular, is to wean myself and my household off Windows. Especially now. The price for Windows 7 shrink-wrap significantly approaches the price of a new hardware platform. The economics get even worse when talking about buying refurbished computers. Putting Windows 7 Home Premium on a salvaged computer system makes even less sense.

What has been damaged more than anything else is my trust in the Ubuntu programmers and test engineers. I had gotten to the point where I believed that installing anything from Ubuntu was not going to be an issue requiring lots of remedial work. It was stable, able to install on practically anything without crashing. That trust has been severely damaged.

Win 7 Embedded is an oxymoron like jumbo shrimp. The OS is so freaking fat that it really doesn't make much if any sense to use it as an “embedded” operating system. As a touch screen enabled bistro table “information appliance” yeah, I'll buy that idea. Something to put inside a handheld or portable device? No and no way. Putting it in netbooks with Intel Atoms, or Via C7's maybe, they'd be slow. Windows XP though would be a better choice, and Win XP Embedded even better.

I suspect that a large amount of the fat in Win 7 comes from supporting old, really OLD applications. As an example I found edlin.exe in the system32 directory. That in itself was funny since the original edlin was a 16 bit LINE editor in MSDOS. It pre-dates edit.exe, another MSDOS editor, also in the system32 directory. Adding notepad.exe and write.exe makes 4 text editors in one folder. Is that really necessary?.

(Before you jump me, yes I know edlin.exe is still in XP Pro etc. When was the last time you HAD to use edlin? Did you really want to?)

There are runtime packages for C, C+, C++, VB5 and VB6, old MFC etc some of them pre-date Win95. ODBC database connectors for Access 95/97 databases, dBase3, and Paradox. Support for OS2, its limited but there.

Iexpress.exe, an application-installer-packager from the Windows 3.1 era also has an system32 “update”.

Most of the Win7 fat though is semi-hidden in plain-sight. Portions of the operating system have to be written in such a way to support either the old applications directly or through the application compatibility add-ons Microsoft has patched onto the various versions of Windows. Ntvdm and wow (Windows on Windows) are examples of application patching, hosting or shims embedded into Windows. I appreciate the fact that Microsoft wants to support everything they have ever released (except maybe MS Bob!) but come on, can't that stuff be supported in a download and only on the users' systems that need it?

Since Microsoft has stopped supporting MSDOS, Win 3.1 & 3.11, and WIn9X directly, why continue to support them in the new operating systems?

Hook Application Compatibility into Windows Update and use that to download the appropriate packages to support the old stuff the user has to continue using. There already is a side-by-side mechanism setup for the DLL hell of previous Windows NT versions. Something similar can be done for the old stuff, especially the 16 bit stuff.

How about MS making their Virtual PC software into something that does the legacy support? Its an extension of the idea of Virtual XP Pro stuff going to be done in WIn7.

Think of all the plug-ins that users have to download when they go web-surfing all over the Internet. Its not like the users don't already download most of the junk on their computers already.

Most of the people I have had to fix their home computers don't make back-ups and lose their installation disks so when they go out and buy the latest version of Windows Whatever, they end up buying new software anyway. This new software seldom needs MSDOS and 16bit Windows support. So why leave it in the OS as part of the piles of detritus that hardly ever gets executed?

If you want another argument to remove this un-needed dross, think of system security. All of this old compatibility software sitting on the system has a very large and exposed surface to malware writers. At present most of the compatibility software hasn't been used much to attack the host 32 bit system but its an attack vector waiting to happen. If the software wasn't there reliably on every Windows system that would be one less way to hack into or around system security. In other words, the 16 bit legacy software is not on the system UNLESS the user downloads a compatibility package, until then it wouldn't present itself as such a tempting target for future malware. If the malware guys can't count on it being there for use then its not a viable means of attack.

For a customer's installation, we needed to buy a new rack mount server to replace a failed system. At first I was told that our newest software based on Java would be running on it. No problem. Just get Windows Server 2008 R2 since the server was coming from DELL. DELL recommends Windows (even when the customer doesn't want it). Actually Windows XP Pro SP3 would have sufficed. The new server would be replacing a previous 1U server that had been running Windows XP for nearly 6 years when it finally cratered. That's a phenomenal length of time for oilfield computer equipment.

I was mis-informed. The customer reports that they are using really old legacy stuff based on VB 5 & 6. Oh yuck. And that generates a problem. Turns out old VB5 & 6 software running old unsigned Active X components will not run on Server 2008 (“Visaster” server without the eye-candy). Not in WoW. Not in ntvdm. Not at all. I had message boxes popping out all over the place. I had to try to make it work. I was amazed that Server 2008 would actually let me install the software in the first place since the installer program was from the same era.

This is one of those situations where using a virtual machine could solve a problem caused by an incompatible operating system. I was tempted but I didn't bite the apple. This particular server system will be located on a drilling rig off-shore who-knows-where. The techs on the rig are fairly astute and have handled complex computer issues before. But asking them to learn a new technology just because the current OS cannot run what the old OS could is too much.

The system got “downgraded” to Server 2003 (XP Pro-like) and everything not only installed but ran without complaint. Since the system will be running behind at least 3 layers of managed routers and is on a private VPN supported intra-net it should be fine, I hope.

That kind of trade-off happens in the real-world. Microsoft knows that happens. You can't force your customers to upgrade their systems if they like what they have running. What would be wrong offering for sale old Windows operating systems with absolutely no warranty? If you read the current EULA you're not getting anything anyway from Microsoft even on the latest iterations or builds of Windows. As far as tech support goes, we don't even bother to call MS since it takes practically a “verbal hurricane” to elevate their attention.

What is wrong with selling old versions of their OS is that the old operating systems have so many ways to be compromised that botnets would be even worse than they are now. So in this instance I have to agree with Microsoft. Don't sell or install the old stuff but sometimes you don't have a choice.

I feel uneasy delivering something I know is broken and has been cracked by malware malcontents hundreds of times. When you “own” the operating system instead of just “leasing” it, you can better control the effects obsolescence, planned or otherwise, has on your corporate bottom line. I would have much more preferred to have installed a Linux based server for that very reason.

(endorsement) As to DELL tech support, I was pleasantly surprised. It took wading through about 6 or 7 phone voice menus but I did get some help. It turns out the Raid driver required was on a third disk. Quick and easy fix. (/endorsement)

Linus Torvald's says so too. (At least as reported by CNet. I tried to find the original quote. Probably only available through Conference proceedings.)

One of the biggest advantages of Linux is that you don't get trapped into one distro or version as with Windows. If you want a smaller footprint OS, you can find a Linux version that's a LOT smaller than Windows.

On VirtualBox, its like all other VMs from other manufacturers, if you can afford to throw more RAM at it, go ahead and do it since VMs are like JAVA, real RAM hogs. RAM is incredibly cheap right now.

The other area they like to gobble, is hard drive space. Usually 8 to 10 GBs at a time. A large hard drive is an advantage here. Intuitively putting the VM "virtual hard drives" on a drive other than the boot/system drive for the XP host should help performance.

Unfortunately by default VirtualBox will place the VM hard drive files in the Document and Settings User profile folder. Ugh. Be sure to put them some place else to prevent massive profile folder bloat.

When it comes to operating systems, especially Linux distros, trying as many as you like, yes you can. Create a new VM for each distro. Just remember that a running VM will grab RAM space. The more RAM you can give it totally uncontested, the faster the VM will run. So to speed up VMs, limit the number running at a time and give them as much room as is practical. That means if you aren't using it, suspend it or shutdown that specific VM.

Beware that the hard drive you have all of this running on is in good shape. What can't run or isn't running in RAM, is being swapped out to the hard drive. So drive swapping tends to increase rather remarkably. A high speed drive helps but its still way freaking slower than more RAM.

Each of the VMs will have a setting for the amount of RAM to allow the VM to use. That comes out of the total amount of RAM of the host. The host requires RAM space since its actually fairly busy responding to the services demands from the VM so don't totally choke off the host. Setting RAM below 512 MB on XP Pro SP3 system without shutting down a lot of the unneeded XP services is likely to generate a big performance hit to the VM.

When you get comfortable with Ubuntu (or another Linux distro) I find that running VirtualBox on Ubuntu as host, things runs faster than with VirtualBox running on a XP Pro host. Mostly because there isn't so much crap running like in XP Pro. But also because the OS can use the RAM above 3.7 GBytes like XP Pro can't in a 4 GB motherboard. This is an OS limitation based on a hardware issue. Might be different in differing OS or hardware. All my testing has been on Intel chipsets. Not a problem on 64 bit systems. The biggest argument I see for a 64 bit board is a VM.

Also another trick is to shut down as many services as you can in the XP host so that you can to give the VM more room. However do NOT shutdown networking, firewall, SMB, RPC, DNS, DHCP, workstation, secondary logon, error reporting, event viewer, or printer services. If you just stop the services instead of disabling them, when the system is rebooted they come back. There are others as well.

One special service ABSOLUTELY that needs to be turned off temporarily is Windows Update, especially if you have set it to auto-install. You don't want it trying to update while the VM is up and running! Especially since it seems practically ALL XP updates require a reboot.

You can shutdown dotNet related services, indexing service, help & support services, DCOM, Terminal Server like remote desktop and assistance, and anything related to dialup server (its not likely to want to stop) and you're not likely to need Hyperterminal!

In the case of Dialup Server and Telephone related services, if you aren't using a dialup connection, you can disable them but it requires a reboot to make it "stick". Just set them on disabled and then reboot. Its almost impossible to get the telephone services to shutdown from the MMC or Services applet. When the XP system comes up the services won't be running.

In the process window in Task Manager, you can also stop any of the software update processes, like Google, java (juschedule), realplayer etc. If you have a variation of SQL server or its agents or writers running in the XP host, shut them down. If the OS comes back and tells you that the process is vital, believe it and try shutting something else down. Anything with svchost attached to the process name is likely networking related so just leave it running.

If something stops working that you need to have running, just reboot the XP Pro host. As long as you haven't gone into the Services applet and DISABLED the service, it all comes back on the next reboot.

Please note that all of this "advice" is related to Windows XP Pro. I haven't much experience with Windows XP Home. In the areas of running VMs on XP Home, I've had no experience. I've considered XP Home the "runt of the litter" and haven't paid it much heed. My kids run it on their systems but that's about it.

BTW all of this advice works for everyday XP Pro operation or game playing even without a VM running!

I try to install as little software as possible on a computer's OS to keep the running speed to a maximum. In Windows there are limited things that can be done like killing the taskbar tray, shutting off unneeded processes, etc. Even killing the Explorer shell can give you a little speed boost of about 1% and about 15 to 20 Megabytes of more RAM space in XP Pro. Most applications using Win32 API calls will still draw windows and do the Windows Form library functions without having to run a full Explorer desktop.

I find that Ubuntu 9.04 runs the exact same applications as XP I need at work and at home for 85% of my work. Word Processor/office suite, a web browser and email client application. For those situations where I HAVE to run Windows, I'll run XP Pro in a VM hosted on Ubuntu. Its about as fast as Win7 without a VM on the same hardware.

The complaint I have with Windows OS that Linux addresses to a certain extent, is that I can strip out or NOT install big chunks of software that is more rightly defined as application layer software instead of the bloat the has driven Windows into the ground performance-wise.

Considering the speed increases in the hardware, the operating systems ought to be running 5 to 10 times faster than they do. Mr Kingsley-Hughes timings on the install/upgrades confirms to me that the Win7 operating system is too damn fat. (check out his column over at ZDNET.COM yesterday)

Years ago I did tests comparing an application suite software our company wrote running on Windows 2000 Workstation/Pro SP4 and XP Pro SP1 on the the exact same hardware, same RAM, hard drive etc. Win XP Pro ran the software 20% faster than Win2K.

A similar test I've done running our current application on Windows XP Pro and Win7 Beta on the exact same hardware revealed that XP Pro was faster by 5%! What happened?

I went through looking at what processes were running on the 2 systems. I tried to optimize the running processes in Win 7 to approximate what was running in XP Pro. I did speed up Win7 a bit but not enough to be statistically significant. Win7 was still slower than XP Pro. Both tests were done again on the exact same machine with 4GB of RAM, both Windows OS versions were 32 bit and the video and network drivers were Microsoft's device drivers. XP Pro was running SP3 and Win7 was running the Beta.

If you open the Services windows in both XP Pro and Win7 and compare, there are approximately 3 times as many services running in Win7. A lot of them do arcane tasks that have little service for data display or word processor-like applications. The trick is to figure out what can be shut-off.

Its obvious that I will have to repeat the test again once Win7 comes out. I was very surprised by the results. I expected Win7 to blow past XP Pro.

I installed a replacement hard drive and started the task with DELL's XP Pro re-install disk. I have images of the system but they are loaded with a lot of stuff I didn't really need for future work. So I decided to re-install everything from scratch. The DELL disk was slip-streamed with SP2 and I have a network install for SP3. So maybe its not really totally from scratch, close enough.

This time I'm running a clock on the whole process. I won't fault Microsoft for the bad drive so the clock starts at 4:23AM after the drive is installed and the system has started booting from the CD.

I've partitioned the 200 GB drive I installed into 2 sections. The first partition is NTFS for Windows XP Pro at 90GB. The second will be for Ubuntu 9.04. At this point, the second section is just unpartitioned space. There is a second drive, 320GB in size, that is formatted in a single partition as NTFS. Both drives are Western Digitals. The final drive complement is a CDROM\DVD RW and a CDRW drive. The computer is a P4, 2.8GHZ with 1GB of DDR RAM, a DELL Dimension 2400. It has the stock Intel video 845 family chip turned on as primary video adapter and a Nvidia TNT2 Riva PCI pcb with 64 MB of RAM as the second video adapter. Drivers for all of this are from the DELL web site. The Nvidia drivers are in SP3.

Base XP Pro system install at SP2 level is done by 5:08AM.

In the end what's installed on the system is SP3 from the network install package already resident on the second hard drive and approximately 60 updates downloaded “live” from Windows Update. Not all of the ones offered, but most. Total size of the first of 3 Windows Updates runs was 149 MB, including IE8. There were updates that loaded specifically for IE but I have no clue what size the files were. I refuse to install MediaPlayer 11 or dotNet framework 1.1. MediaPlayer 11 gets dumped because of the DRM issues. As to dotNet 1.1, nobody should still be writing to that piece of drivel. Other ignored patches related mostly to security patches appropriate for NT domain membership or for multiple language options I'll never need.

I also installed Mozilla FireFox 3.5, OpenOffice 3.1, Google Chrome, InfraRecorder, mspgcc (a gcc toolset for the TI MSP430) and Adobe Reader 9.1. All of these were package installs from the same second drive. With the exception of mspgcc, these are what I consider essential function tools for a typical desktop.

I used the open source tools simply because the Ubuntu install will get the same applications installed to match the XP.

Clock stops at 6:30AM. Total install time for Windows XP Pro SP3+ selected applications is 2 hours and 7 minutes. Actually quite fast for Windows simply because I “cheated” and already had all of the packages downloaded minus the SP3 updates.

Clock started again at 7:15AM. Now Ubuntu 9.04 is getting the same installment routine into the open partition on the IDE primary master drive (sda). The base system has finished installing by 7:30AM. I go into System/Administration/Update Manager and start it to download all the updates released for Ubuntu 9.04. It downloads 145 MB of updates and starts installing them.

Clock stops at 7:55AM. Total time for install is 40 minutes. Google Chrome for Linux isn't out of early Beta although I really don't need it on Ubuntu. Infrarecorder isn't necessary on Ubuntu since Brasero installs as part of the base install. The base install also includes FireFox 3.x, an Adobe Reader substitute, and OpenOffice 3.0. I haven't looked yet but mspgcc might be available in Linux.

I did get one giggle out of all the keyboard banging and mouse clicking this morning. There was an option to import my User settings from the Windows XP Pro partition. Not ever having done that before I clicked it mostly to see what it would do. It grabbed my desktop picture from the XP Pro Settings & Documents folder for my XP admin account and dropped it on the Ubuntu desktop. Ba-da-bing.

So for the price of: 1 hour and 27 minutes more of my time installing Windows XP Pro versus Ubuntu; 20GB of hard drive space used by Windows XP Pro versus 8GB of hard drive space in Ubuntu; no differences in functionality; potentially becoming XP-ground-zero for every malware-writing jackass in the world; I get the “equivalent” in Windows XP Pro?

No contest.

Ubuntu wins.

My previous comments about Control-Alt-Delete not working still apply. Control C and Control-Break don't do much either.

In HKLM\Software\Windows NT\CurrentVersion\WinLogon\Shell replace Explorer.exe with the BATCH file as described in Not-so-Stupid Nerd Trick, a previous blog. The second step is to NOT modify the setting in IniFileMapping. Third step is to have the system Auto-Logon as described previously. On the next reboot, the fun begins. Done.

With the system set this way, it becomes a single-purpose computer, an appliance as it were. It will not run any User programs other than what you've programmed into the batch file.

With this registry entry you've basically shackled the computer to a single task and that is exactly perfect for kiosks, public information displays, embedded systems tasks etc. It doesn't matter what the User settings are, the system never bothers to execute the the User programmed shell.

An interesting experiment would be to set the System Shell to be FireFox or Chrome. Consider it a taste of things to come, if you believe the Cloud enthusiasts.

You could set the "XP appliance" system up to be a printer server for other computers in the office. Just set the printer management software to be the application started by the batch file. Make sure the Auto-Logon user has printer management permissions.

The old XP computer becomes your dedicated printer management tool. Make sure to share the printer(s) out to all other users. This could be a way to use an old XP Pro license on a slow system that can't run Visaster or Win7. Dumping the printer tasks to an XP Printer will speed up the rest of the desktops when printing. This should be possible with any sort of printer. Or multiple printers connected to the XP Pro box.

BTW if you set up another account as admin that has remote logon capabilities, you can use one of the VNC tools to "manage" the system remotely or to break into the registry and hack the shell back to Explorer.exe. Remember to spell out the file extension, Explorer.EXE.

There are some "soft gotchas" with the bare metal VM concept. ( A soft gotcha is one that if you gave it a second thought or two, you would've known about it before asking the question.)

The obvious one is "How do you deal with all the freaking hardware drivers?" You limit the hardware selections supported.

1) The product is tailored for specific sets of hardware from one of the big three OEMs- DELL, HP and Lenovo or full Intel chipset boards. No AMD or other second source x86 need apply. A Pentium 4 isn't going to hack it either. Systems supported are high performance desktops and laptops.

2) It requires an Intel VT equipped CPU. That limits it to the Core 2 Duo or Quad CPUs. (I knew that was going to be a possibility.) AMD multi-core compatibility is still being worked on.

3) Video adapters are limited to Nvidia and Intel chip sets. No ATI video boards supported at this time. Well I didn't think that was going to happen with AMD owning ATI.

4) The software works with a client and server model. The client connects to the server that delivers the VMs to the hypervisor client. The client is on the Intel VT

5) The Server application has to be running on Windows Server 2008 running on just about anything, no dual or quad CPU necessary. No plans at present for a Linux server capability.

6) Management module resides on the server. Features in the VM and hypervisor can be tailored like levels of security or access to the hardware based on user groups or logon. Active Directory was very faintly hinted at. Each VM can be limited in access to the peripheral hardware on the client machine.

7) VMs are pre-designed or can be client designed as "generics" or as custom as wanted. All VMs can have the same or differing levels of access on each client.

8) "Restore points" or backups of the VM as modified by the client-user can be scheduled or initiated by the user.

9) VM images can be set as permanent safety images with incremental updates to operate as a fallback to wipe out malware in the VM.

10) New software release is planned this coming Monday or Tuesday.

Still got some growth to do but it looks really interesting for a lot of reasons.

The biggest is that if its done right, it eliminates the double-API conversions that have to happen between the OS in the VM and the OS on the host hardware. That tighter binding to the hardware should offer a really nice speed improvement.

It might spur some activity by others in the VM game to bring out their products. My experience with VMware on XP Pro was not a bad event but it was still slow on certain software applications.

In many ways, a stripped down Linux works fairly closely to the hardware level of this no-OS hypervisor. Maybe a group of Linux driver and kernel programmers are working on something like this?

What caught my eye was the concept that a "hypervisor" would run on the hardware of choice sans a hosting OS. In turn it would host as many sessions as the client wanted (within the physical & electronic limitations) with multiple VMs running the applications desired in each one, That would allow the simultaneous use of Windows and the Linux of the month in separate windows without any possible "leakage" from one to the other.

Hallelujah, somebody has seen the light! That is the way to do virtual machines. Give each major application or group of applications its own solitary-confinement cell with nothing the OS can use to get to the other cells. A virus or Trojan should not be able to hop the gap from one cell to the other because they won't know they're there.

Truthfully I do not know exactly how that can be done in total but it is the absolute best possible sort of solution with the recently released and soon to be digital d

Basically setting up a custom User Shell has been fairly well documented. It takes changes to 2 registry keys, one in HK_Current_User and the other in HK_Local_Machine. The Local_Machine key opens up Pandora's box and the Current_User key puts something in the box. I would suggest that the User account be a Limited Account, not an Administrator.

First log-on as the intended victim, uhh make that User and open Regedit.exe and change HKCU\Software\Microsoft\Windows NT\CurrentVersion\Winlogon “Shell” value to an executable BATCH file. This entry must be a RZ or a single string and must contain the full path if its not on the User's path variable. So something like this: “c:\folder name\surprise.bat”. The use of double quotes is suggested for paths that include folder names with spaces.

Surprise.bat can be any executable you would like to drive the User crazy with.

A one line batch file is all you need like: cmd /c “c:\folder name\surprise.exe”

Make sure to put whatever command line options you need inside the right-hand quotes. The batch file can be as long and complicated as necessary. Just make sure that all of the executables and other files necessary are where the batch file says they are. Once the batch file starts, the usual Control-Alt-Delete doesn't work. Control C will sometimes stop the program BUT it goes right back to the beginning and begins again.

Now log off and log-on again as an Administrator.

Now what makes it nuts is to set the System to allow custom shells. Start Regedit again, go to: HKLM\Software\Microsoft\Windows NT\CurrentVersion\IniFileMapping\System.ini\boot\shell

Change the Value to be:

USR:Software\Microsoft\Windows NT\CurrentVersion\Winlogon

Note that the string value starts with USR:Software.

Reboot and let the fun begin.

Every time the vic---User logs on, the program in the batch file starts. Every time the user manages to stop the program, it restarts. The magic cookie is the Batch file. Since the system considers a batch file to be part of the cmd shell it executes it. When the program is stopped it returns to the registry to restart the shell. Which is the batch file that just got terminated. This same behavior is what restarts Explorer when you've stopped the Explorer process in the Task Manager window or more likely it blows up all by itself.

This has some really nice properties for programming “endless loops” for slide-shows, media presentations in kiosks or powerpoint-like presentations. The advantage is that outside of pulling the power plug, there's no way to shut it off.

Add an AutoAdminLogon registry entry to the same WinLogon key in HKLM with a RZ value of “1”, put the DefaultUserName and DefaultPassword entries in the same key with the Victim's user name and password and you're all set. So the next time the power is plugged in, there it goes.

This is an excellent technique to use when scripting some sort of automated process that absolutely, positively has to be done exactly right. If the user manages to shut it down it starts right back at the beginning and runs all the way to the end. When the process has reached conclusion, put a shutdown.exe command in the batch file to shutoff the computer.

It might also be a way to torture terrorists. Put Windows Mediaplayer on it and force the terrorist to watch a SpongeBob SquarePants cartoon over and over again. Just don't give him earplugs or a keyboard and he'll tell you what you want to know in hours.

Since Thunderbird is a IMAP/POP client, its going out to the intranet and then back to the Internet connection on our Exchange Server, so I run a secure connection. We have a number of employees that are located all around the world so web access is necessary. I also get to take advantage of the spam filter since the IT guys have it set to NOT distribute spam to outside remotely logged on connections. Nice. My inbox is so nice and tidy now. Even though Thunderbird is using a POP connection, Exchange lets me keep folders on the server and local folders on Ubuntu without any complaints at all.

I was a little worried about compatibility but it seems to work without any complaint except one. I get a message on every refresh or connection that the server security license is not licensed to the Exchange Server. Yes that is true but even when I set Thunderbird to accept the license as a permanent exception, it pops the warning up every time. What really is the problem is that the license does not mention or cover the web connection I am using to get to the server, just the inside connections. So this message is popping up for all of our Internet connected employees as well. When I have some time I'll figure out how to deep-six the security pop-up on this one single connection.

Oh well. Its just another click. Working with Visaster has really inured me to the potential security threats represented by the pop-ups. (Not!) So far I haven't yet gotten nailed by a spoof pop-up (knock on my head!). I try to check my email when I can stay focused on it and not be distracted so I'll not just blindly click-through. However, I'm only sure about myself, not the other company employees. A number of them are situated where they have to use public Wifi or 3G connections to get to the Internet. Hopefully we won't get trashed by something dropped onto the Exchange Server by a remotely connected employee or a wireless man-in-the-middle.

One of the future IT fantasies would be a completely secure email system such that you could trust the connection, the message and the ID of the guy sending you the email. And likewise for him as well. For employees, using a variation of VPN or some secure tunneling technique takes care of the connection security issues but doesn't necessarily lend itself to expansion beyond the corporate structure. Vendors, sales reps, FAE's etc are a lot of the people I communicate with daily and for them there is no easy answer to increase security on email.

If I had a wish, it would be for the Internet gurus out there to invent a really secure email service and protocol that could be used for managed connections. Something besides set email filters to prevent unwanted email from getting dropped on your desktop. I could handle it if there was a way to automatically negotiate with the remote server and give them a secure keycode or token to allow an encrypted transmission to come through. PGP is good but its not automatic.

The email clients could have maybe four levels of security. SPAM or JUNK, the lowest could be just dumped daily if convenient. Unmanaged, would be the advertising that is of some use and maybe job related. Managed would be the email you wanted to get from vendors and the like and it would have the security features enabled. Corporate would be the company originated email and the most secure.

Another nice feature would be a way to dump forwarded email with spiritual, religious, and/or schmaltzy sentiments back to the originator AND all the people that forwarded it to the hundreds or thousands that got it before you did. (I did say it was a crabby engineer, right?)

Just some ideas.