http://en.wikipedia.org/wiki/Deep_Freeze_(software)

Make two partitions, one for system, other for My Documents. Install everything, set it up, update it, make a backup, freeze the system and tell them. I do not like antivirus, but you do as you please, just remember to install them to unfrozen partition.

"You can do everything you want, but the computer will reset everything as it was after restart."

Hope it helps...

This is a classical example of how you just have to take the time to read the paper and think about it because the cited text and other sources that are not behind pay-wall are just not clear enough.

I have had put my head in number of MR scanners, with field strengths of 0.3T, 0.5T, 1T, 1.5T, 2T, 3T, and 7T, with no nistagmus, but that simply does not mean anything in context of this paper.

You should have in mind that the maximum magnetic field gradients, and their speed are regulated by law.
Interesting question that this paper is offering: are functional-MRI results sometimes tainted by effects of magnetic field gradient while the person is placed inside the device?

As one involved in medical profession, I am witnessing a slow but steady advance of robotic and similar "automatic"/"AI" technologies helping us more and more each year.

I for one, welcome the advent of robotic helpers in OR, ER, radiology or other part of medical profession.
I do believe it will take a decade or more before we can use automatic detection on our x-rays, CT/MRI scans or blood-work. But, to be honest, "robotic" or better said "automatic" diagnosing is already here.

Take a look at you ECG strip/record. It already contains an opinion of the machine.
There are several systems that can make a diagnosis from mammograms with better results than many overworked, underpaid, outsourced (or local) medical professionals (doctors).

These are just two examples, but use your search engines and find for yourself that there is a considerable number of independent (and other) groups that make great efforts to make a software that can give a diagnosis from images, samples or other information about your health.

It is just a matter of time when your portable sensor device, attached to skin or in your hand, will measure some molecules in your breath, sweat or blood, and offer an advice. It will be cheaper, it could not be fooled as easily, and it will be available anywhere, anytime.

Maybe I am dreaming, but take a good, long look at your mobile phone, and then replay to this post.

Back in the 1990-s my friend was working at one institution where they had 3T mass spectrometer. The machine developed a error and it would work for two-three weeks, and then it would die. Each servicing was taking tens of thousands of dollars, so my friend started tearing the machine apart one day.
.
After his boss went nuts, and basically, everybody left the building trying to get him of campus, he discovered that one lens for laser was of incorrect type, went to optics shop, bought new one for $1, replaced the lens, and repaired the machine. Original lens was too large and it would simply flip out of the socket after even a gentle nudge.
.
Company that made this device it is long gone now, but this machine still works without even single quenching.

Two weeks ago, one Sony printer started to drop packets (5-10% - nothing fancy), and generally, refused to print films (35x43cm). It is connected to a Cisco switch, same as about 20 other devices that work fine, thank you for asking. Interestingly, printer's web page works fine, printer reports all test are o.k. but switch keeps bringing errors and no films.
No changes in settings or connections were made to printer or switch, but it simply stopped working.
.
The switch costs about $3,000 and printer about $10,000. Bummer.
.
After IT guys combed through this problem, printer was disconnected, and sony service contacted. No luck with them either.

Then, one MD needed an hardcopy baldly, and tried to connect printer back to ethernet socket, but could not find UTP cables of sufficient length. So he took $10 SOHO switch, two UTP cables and connected printer back to network. ...you guessed it.. it works fine now.

We also had similar expirience with home routers, but then tried monowall, pfsense, ip cop, and mikrotik. All of them work nice, have more or less user frendly web interface (or something similar), and also differ in price (monowall, pfsense, and ip cop are free). In all cases we were serving a mixed wireless-wired network of 50+ users, using NAT, DNS, firewall, port forwarding and some other features, depending on "router/firewall/whatewer" software mentioned.

At hardware side, we tried several hardware configurations, from 125MHz ARM-based routers, to 333MHz celeron or over 2GHz AMD processor-based PC's, and maybe most interesting was an Alix board with 500MHz AMD Geode x86 processor. Runs at low power, it's small, and gives all advantages (and other things :) ) of a PC. Trying several homer routers (Linksys, Buffalo, Planet, TP-link, etc) proved what other posts already pointed out - they are good-enough for aDSL lines, and speeds up to 10Mbps. Nevertheless, several Thompson and Siemens routers performed badly (instability is their middle name), but they are out of your league anyway and some of them are not available any more. Worst firewall in our experience came from Microsoft (ISA), and while being stable, it introduced huge packet delay and a number of "features" that made us bitter many times.

We also tried several Cisco routers and firewall, and to say the truth, were not impressed by what you get for the price, as beforementioned solutions provided same or better level of service for much less money. I don't say that they suck, but just that they are some kind of reference, so we tried them.

For last 2 years we settled with 1.6GHz AMD Turion based PC with 4 network cards, and one wireless card, 512MB RAM, system is on 256MB CF card, running one of mentioned software packages, while logging is done on separate machine. Going with CF (notice that nothing gets written to it) instead of HDD, provided us with increased stability, as hdds do fail more often. Good UPS is also a plus.

So many replays, but I can't see the right one. Here is my configuration:

Motherboard: Alix1C (ITX board 500MHz AMD Geode processor, 256MB RAM, Alix1D is currently in production). board has two serial ports, parallel port, several USB ports, LAN, PCI, mPCI, 4 audio ports, keyboard/mice port, 44pin ATa port, and CF card slot, and other...

Power supply: 12V, 1.25A, idle consumption 5W plus 2-3W for HDD (Hitachi 30Gb 2.5" ATA HDD), pluged directly into board. See http://www.pcengines.ch/ for more info. I tested board from -10 to +50 degrees Celsius, works like a charm for two years non-stop now.

One PCI SATA controler, with two 1TB HDD, and 12V 2A + 5V 2A power supply, fanless. HDD's take about 30-35W together (haven't measured it).
I placed all that in one old ITX case, and right now I have replaced original power suppply with two fanless "brickl" PSUs. Case is ventilated with one 120mm fan, running on 5V.

Ubuntu LTS is installed, running services: Teamspeak, ftp, www, mysql (for WIKI and other things), DC hub (for local network), DC client, Open2300 meteo station software (see http://www.pljusak.hr/ and http://www.lavrsen.dk/twiki/bin/view/Open2300/WebHome ), samba, and some others.
I was just reassembling it yesterday... ( http://www.vallisaurea.net/staro/webcam/S5004006.JPG and http://www.vallisaurea.net/staro/webcam/S5004008.JPG )

Well, you got the point.

Unfortunately, humans do come with no service manual or test points or other service/manual documentation.

It stands that You make considerably smaller amount of false engineering decisions. When did You have default value range 1-100 out of possible 0-300 units? It is common thing in medicine.
If You put voltmeter at test point number 321, you measure exact that voltage, while in medicine, blood sample can literally be different because the room walls were of different color or because nurse said something or it was not taken in the morning but after the lunch.

It seems to me that considerable number of problems comes from the fact that engineers are used to work with models, while medicine is done in the real conditions. I agree that science part of medicine makes difference, but the ground is still shaky.

Just remember, if something is done in one hospital/county/state one way, there is no way that all of it will be the same in next hospital/county/state.

I don't have a theory I have experience.

I work for hospital that went digital (for patient recodrs) in 2006. All (billing) administration was internally digital (using different, obsolete system working on DOS and floppy disks) from 1997 and to outside world also, depending to health insurance company involved.
After 6 months of education, switch was made in one day. It was horrible, but after two weeks things were looking just as before. After several months, 75% of administration was more efficient than before, and now, 3 years later, we still print outgoing documentation, but doctors rarely look at papers. Nevertheless, printing expenses went 30% down this year.

Last year all waiting lists were computerized, and made available (with no patient data) at the web pages. That saved us so many work hours at all departments, but two people switched to that department. This year we are looking into making all internal administrative procedures digital. Hospital restaurant was really happy after we made their menu available online at intranet.

Several months before introducing the system, all work places received computers with unified user interface, and demo program installed. It was made really clear that someone should consider finding another job if they refused to work with system. People near the retirement (2-3 years) were exempt from this rule.

The problem very often lies in wanting too much (all). Process should be step-by-step. Billing first, patient records second, intra-hospital administration third or any other way. Every step should be planned, because people will suffer at it, and don't rush it. It takes months, sometimes years for one (new) work flow to settle in.

Radiology department is still not filmless, probably because it costs as much as putting all patient records in computer. Volume of data that our radiology department produces in one day is equal to 1-2 years of data from whole hospital. On the other hand, introducing PACS and RIS is so much more widespread, but the volume of data makes project harder in the long run. After testing almost a dozen of PACS/RIS demonstrations, one free PACS amazed us with results, holding test data (0.5T of images) and working better than some very expensive solutions.

I am just wondering... What would happen to classcal music if they started to use Auto-tune. the whole point of music and excellence would simply disappear on the first occasion of live performance.
What has already happened in case of "popular music". Decades ago.

Just imagine a opera singer going out of sync with others... but wait... that is what live performance is all about, to make avery performance a bit different but not wrong.

It has been proved that holding an beat perfectly makes a music boring, while artists that have tempo correct on average do sound good.

A PC user has to buy new hardware to switch to Windows7, also.

Instead of bitching about Linux and piracy, they should take care of their product(s).

One more thing about cats, mice and rodents. As they are all mammals, they can be infected with Toxoplasma gondii. Definitive host is a cat, but that protozoa can also infect humans.

Toxoplasma invades the brain of the mammal or bird and changes their behaviour so that e.g mice are no longer afraid of the cat. Problem can be bad, since Toxoplasma is carried over by cat urine, and people can become infected too. It is noted that Toxoplasma changes behavior of the people, but more important is that it invades brain of the baby if the infection is acquired during pregnancy.

For start, read (as usual).

I second this. I have several routers and wifi routers with Geode, running pfsense, monowall and plain linux.
They run from -15 deg. Celsius to 50 deg. Celsius, day and night for years without restart (I hooked several Li-ion batteries with small circuit for backup power). Even my home server (with SATA card plus several disks on separate PSU) runs on Geode.

Switching to ARM in this class is not practical, just try getting support in kernal for sata controler or wifi cards or for some less widespread device... Are there any boards with integrated USB, printer port, two serial ports, IDE/CF connector, VGA and LCD port,128/256MB RAM, single rail power supply (12-18V) and that idles at 3-5W and costs about $100-110?

And do not mention atom. Current boards consume too much power and need ATX PSU, and that is huge problem. Same goes for many VIA and other small factor boards having celeron or soething similar on board. Come to think of it, maybe board manufacturer will switch to VIA, but last 5 years, we are in love with Geode AMD.

Geode will be missed since platform like alix draws power from 12V 1.2A source and not a shabby ATX connector and power supply that draws more power than whole board!...or costs more than the board.