To Fix Or Not To Fix, That Is Often The Question.

After successfully extracting the three badly leaking batteries from the D-cell Maglite we have, I can say with authority that every YouTube video and web forum advice article I found concerning how to do this is pretty much limited to a situation in which the batteries have leaked relatively little. Mine had leaked *lots*. I ended up making a tool for the most critical part of what had to be done.

First, let me comment on the various suggestions I found which did not work:

--Use acid (phosphoric, acetic and citric all are suggested in one place or another) to eat the leakage, ostensibly freeing up the stuck batteries.  It doesn't work if there's more than a tiny amount of leakage and/or the leakage has been present for some time. (The potassium hydroxide inside the battery will aggressively attack aluminum, forming salts that are not readily dissolved by acids unless the acids are strong enough that they will also damage the aluminum further.  Anything that will eat the hardened crud that's holding the batteries in place probably will also eat the flashlight.) (Note:  Phosphoric acid is the ingredient in colas that allows them to be somewhat useful in removing very minor corrosion.  Neither cola nor any of the various commercial phosphoric-acid-based corrosion-removal products like C-L-R are effective in attacking bad buildups inside of a Maglite when the batteries have leaked significantly.  Citric acid is the active agent in lemon juice; same problem.  Acetic acid is what's in vinegar, ditto.  Don't even think about using stronger stuff like the jug of high-strength muriatic acid that you bought back when you had that house with a pool in the back.)

--Heat the barrel of the flashlight with a heat gun or by dipping the flashlight in hot water briefly to expand the barrel away from the batteries, and then either knock the open end on a block of wood or use a screw and a pulling device to extract the battery.  This simply doesn't work if the corrosion has been present for too long.

--Use a "dent puller" (slide hammer with a large sheet metal screw on the end) to extract the batteries; as above, doesn't work if the leakage has been present for too long.

--Use a long, large screw to act as a puller directly; see notes about dent puller, same problem.

What *did* work:

One battery at a time, I punched a hole in the center of the back end of the battery, and used a hole saw slightly smaller than the diameter of the battery to clear the metal end out of the bottom of the battery. Using a long punch, I upset the plastic washer just inside the end of the battery, grabbed the washer with a long needlenose plier and pulled it out. I then used a pressure washer to flush out most of the guts of the battery. This left lots of stuff still inside the battery casing, so I used a very small dead-blow hammer to whack the side of the flashlight barrel and break up the sleeve of black stuff lining the outer shell of the battery, and then flushed again with the pressure washer. On two of the three batteries, the shell of black stuff didn't easily break up, so I went back in with the hole saw, with the pilot bit drawn all the way up inside the saw shell, and demolished the black material to allow it to be broken up and flushed out. With the battery's shell empty, I needed a way to get it loose from the inside of the flashlight barrel, and that's where I concluded that I needed to make a tool.  I ground a shallow angle on the end of a long "alignment punch" to make a peeling chisel that could be driven in between the battery shell and the flashlight barrel's inside surface, deforming the battery shell toward the inside of the flashlight barrel and breaking it loose from the barrel. I then grabbed the deformed shell with extra-long-reach needlenose pliers and pulled it out.  I used the peeling chisel again to tap on the remaining metal disc that was glued to the battery ahead of the one just removed, and extracted it with the extra-long needle nose pliers when it came loose. The peeling chisel was deployed once more to break the leakage cruft away around the edge of the next battery. From there, it was a matter of repeating the steps above until all batteries were out.  There was still an irregular layer of cruft on the inside of the barrel, composed of the hard and insoluble compound that results from the reaction of potassium hydroxide with aluminum.  Scraping it out with the peeling chisel was only partially successful, and I ended up scouring out the last of it with a brake caliper hone, which is an automotive specialty tool that you probably won't have. 

Be aware that the guts of the battery can contain irritating or hazardous chemistry.  (In particular, potassium hydroxide, which is truly nasty stuff.)  Wear safety goggles and appropriate gloves and gear to avoid skin contact.  Most of the time, batteries that have leaked will have already reacted the caustic substances out to where they're pretty much inert, but it's not a good idea to assume that this will be the case.  Keep the dog, the cat, the kids, and that idiot from down the street who thinks he knows everything well away from where you're working until you can get it properly cleaned up and the debris bagged and into the trash barrel.

Important note: You have to be really careful with the drill that's driving the hole saw, or you can tear up the threads for the end cap inside the barrel. And when using the hole saw on the third battery, you'll need to get clever with inventive solutions and/or use a special extra-long arbor in order to get the hole saw to reach that depth inside the barrel.

Another important note: NEVER bang on either end of the barrel with a hammer, or bang it on a hard surface; the aluminum tube is soft, and will mushroom, ruining it. You can do a *limited* amount of whacking with a *small dead blow plastic mallet*, and a *limited* amount of banging the battery-cap end of the tube on the floor or a hard surface WITH THE BATTERY CAP INSTALLED, but even this needs to be done cautiously.

If you had a good vertical mill or drill press, and an appropriate vise and bit, it would be possible to core all three batteries in one step, and then peel them out with the peeling chisel. If you had that kind of tooling, this option would probably be obvious, and you'd likely know exactly how to go about it.

Once the leaking batteries are removed, it becomes possible to access the switch assembly, which *may* respond to attempts to clean it out, but most likely will have to be replaced instead.  I was able to source it on Amazon.

That's the whole of it.  If, at this point, you're looking at the level of complexity involved and saying "that's way too much work, I'll just buy a new flashlight instead", I can't say that I blame you.  It really was way too much work, and had this not been a special edition Maglite that I'd be unable to replace, I would have given up and chucked it myself - and most likely bought something cheaper that produced more light.  It's your call, but at least now you know what you're up against.

Three Ways To Use The Galaxy S4 USB Port For More Than Charging The Battery - Maybe.

A number of times recently, I have found myself with the pressing need to send someone a file via email in places where the only nearby Internet access was in my pocket; my cell phone.  In theory I should be able to tether the phone to my laptop long enough to make that happen, but in practice that isn't always the case.  For one thing, my phone isn't rooted, and most of the tethering utilities only work on a rooted unit.  For another, having seen the way that using a typical smartphone as a hotspot drains the battery at warp speed, there's no guarantee that I'd be able to keep the connection alive long enough to get the job done.  So I went looking for a way to transfer the file from the laptop to the phone.

And I found several.

The first, which sufficed in a pinch but wasn't particularly practical, was to transfer the file to the micro SD card that I'd installed in my phone soon after acquiring it.  With my Samsung Galaxy S4, this involved removing the protective case, the battery cover, and the battery.  I pulled the SD card out of the socket, slid it into the USB reader I carry around with my camera, and plugged that into the laptop that had the needed file.  It took the usual 45 seconds for that 13-year-old laptop to remember that it knew it had a USB port, gripe about kids on its lawn, and recognize the card, and then I was able to copy the file over and tell it to eject the USB reader.  After retrieving the card from the reader and slipping it into the phone along with the battery, I was able to restart the phone, use my file manager app to find the item I needed, attach it to an email and send it on its way.  It took longer to reassemble the outer case than to find and send the file.

And that was what made me decide that I needed to find a better way.

So, when the modern user needs a tech gadget that the folks at the electrotoy stores haven't already been pestering you mercilessly to buy, where's the first place to look?  Hong Kong!  But flying there is too much of a pain in the neck with all the TSA crap (not to mention costing even more than a cab ride from Penn Station to La Guardia), so the logical thing to do is - hit eBay.  Every Hong Kong merchant with a son-in-law who needs a little extra income has a dozen seller IDs on eBay these days, so there are lots of sources selling what looks like it amounts to the same thing.  And sometimes they are.

And sometimes they are... crap.

The first thing I decided to try was the OTG 3 in 1 Mobile Phone Connection, for less than $3 including postage to the USA.  (OTG is an acronym for On The Go, bestowed by the Android folks for accessories that plug into the SB socket.  It's not a brand name.)  It had a Type A USB socket, a regular SD slot, and a Micro SD slot.  Such a deal!  I ordered it.  Actually, at that price, I ordered two of them, so that we'd have one for the other phone as well.  A couple of weeks went by, and they arrived in my mailbox.

Eagerly, I plugged one into my phone, inserted the full-size SD card from my camera, and went looking for photos.  Yup, it accessed them.  Sort of.  About half of them, anyway.  Some of the photos produced error messages; the phone complained that it didn't know what to do with them.  So I tried the next step; I attempted to copy one of the not-compatible photos to the internal SD card, and then send it to myself via email so that I could see if the file was really munged or if the phone or the reader was the problem.  To make a long story short, by the time the file got to my laptop, it was so trashed that it was useless.  But worse, so was the directory on the SD card.  (I was later able to recover everything, but the mess that was made of that SD card's file structure was impressive.)  However, since I had ordered two of these units, I had a spare!  Maybe it would work.  Ummm, no, it didn't.  At all.

Well, not to be deterred, I figured that I'd just erred by buying these from the wrong seller; it's not unusual for ultra-cheap stuff from Hong Kong to be cheap because it's the rejects from honorable father-in-law's factory, of which some are OK and some are not so OK, but we can afford to make customer happy by replacing because we pay so little.  And that works acceptably for both parties if the minimum return shipping isn't $11, which is more than I'd paid for the units.  So the smart move was to ditch 'em and move on after griping to the seller and getting the expected "Happy to replace, here is address to return.  So sorry." message.

So I found a different seller, with recent feedback on the identical item that looked promising, and nearly the same price.  (Actually, slightly cheaper, but what the hey, the feedback looked good.)  Ten days later, I had that version.  And they did, in fact, work better, but they were still limited to 16Gb and smaller (no 32Gb) and they didn't work well with class 10 cards at all.

Hmmm.  Not so good.  Not utterly useless, but still not what was wanted.

So I went looking again, and found the 5 in 1 OTG Smart Card Reader, and the much more brute-simple OTG USB connector/adaptor cable.

The 5 in 1 Smart Card Reader shared a lot of the apparent characteristics of the 3 in 1, in that it had a Type A USB socket, a Micro SD/TF socket, and a regular SD socket.  But this time, all three of the inexpensive units that I ordered at least functioned identically, albeit still with issues relating to 32Gb cards.  No SD cards were harmed in the testing of this device, however, but the build quality still left something to be desired.  The Micro SD socket inside the plastic shell didn't precisely align with the slot for it, so it felt like it was scraping the card when it was inserted; not the kind of tactile result that builds confidence.  At least they weren't the complete fiasco of the first two batches.

But that still left the utterly simple and basic just-a-Type-A-USB-socket  version to try, and therein was a better result found.

Although in fact it was of no greater eventual utility than the 5 in 1 Reader, since it appears that the 16Gb limit is at least in part an Android 4.4 limitation, the device recognition was MUCH faster when something was plugged in.  I'm assuming that this was because the phone's own USB hub was looking directly at the connected device instead of having to rely on the handoff through the external hub of the multi-connector Reader.  With my trusty but somewhat venerable IOGear SD reader plugged into the USB cable, the recognition of the SD card was also much faster - and when I subbed in a regular USB flash drive, the file structure came up essentially right away.  WINNER!

Anyway, here's a shot of the business ends of all three units for comparison:

And since inevitably I got curious about the utter-failure units of the first batch, I popped their cases off, and inside one of them I found an obvious attempt to wire around a problem; there was a jumper added from the Type A socket to one of the cable connection points.  That unit was the least functional of the pair, and having seen evidence that they knew it had problems before they shipped it, my suspicion that those were factory seconds or outright rejects got supporting evidence.  

Anyway, the most important information to take away from this, in my opinion, is that the Android 4.4 OTG feature itself isn't quite ready for prime time, but if you're sufficiently careful not to hand it the only copy of some data that you need transferred to the phone or attached to an email, you have some potentially useful options available to accomplish that.  And the simplest works best, unsurprisingly.

On the making of ice cream at home

A bit over a month ago, a friend tantalized me by reporting the acquisition of a flavor of frozen dessert that sounded particularly enticing; dark chocolate chunk black raspberry coconut-based ice cream.  (It was nondairy, so it's technically not "ice cream", but that's a more readily recognized description than the FDA terminology.)   I looked everywhere in an ultimately futile effort to find some - and couldn't get in touch with the friend to find out where it came from.  Then, last night, I was picking up some stuff at The Store I Rarely Patronize when I noticed that they had an economy-grade 4-quart ice cream churn for $17 on clearance, just a bit over half price.  I grabbed it.  Tonight, I picked up a set of first-approximation ingredients for an attempt at making something that would hopefully be similar.

As could be expected, it didn't quite get there, for a variety of reasons.  And it turned out that there are a lot of things that aren't in the instructions that come with the churn.

So, here are some general tips about the process which I learned in the making of Batch Zero Point Five:

 -  If you will be including fruit in the recipe, and you're working from bagged frozen fruit from the grocery store, THAW IT FIRST. No matter where it came from, run most of it through the blender or you won't get much of the fruit flavor in the final product.  And use LOTS of fruit.  If you want two quarts of ice cream, figure on using anywhere from two to four cups of fruit puree. (Note:  The first version of this said "quarts" instead of "cups" for the amount of fruit.  That would still work, but you'd get something closer to sherbet.)  You can also add chopped pieces of fruit for the sake of appearance, but it's the puree that you mix into the main slush that will deliver most of the flavor.  Don't stint on it.

 -  If you are using whipping cream as one of the ingredients, the mixture you place in the churn needs to be fairly cold going in or you'll end up with butter on the paddle. But if you're making nondairy stuff (such as coconut-based), adjust the temperature (usually upward) to ensure that the starting mix is LIQUID.  Under no circumstances should you start with something that's the consistency of whipped margarine or room-temp shortening; it just will not mix. 

 -  Most of the recipes you will see call for sugar in one form or another.  Yes, you can substitute a no-calorie sweetener instead; I used sucralose in Batch 0.5 to acceptable effect.  The results will not have quite the same mouth feel, but it'll still be good.

 -  Chocolate chunks may be better held until the end of the churning, and then mixed in with a spatula or large spoon while the ice cream is still at soft-serve consistency.  (That's where you stop the churning process; it hardens to the final consistency either in your freezer or sitting in the churn with the drive stopped.)  If you put the chocolate chunks in at the beginning, they may all end up near the bottom of the mix.  Ditto for whole blueberries.

 -  Be prepared for problems; have a bucket handy that you can dump that ice and salt and water into from the churn's outer bucket.  If you stop the unit for some reason, a layer of ice can form very rapidly in the mixture on the inside of the churn, freezing the paddle to the container.  If that happens, you will need to remove the inner container, run some warm water over the outside of it to free everything up, dump the outer bucket's ice/salt/water mix into another container, and then reassemble the churn/paddle/drive and get it back into the bucket before putting the ice back in.  If you try to just slip the inner container back into the ice and water, you'll almost certainly end up with ice jammed under the inner container that will keep you from getting the drive latched back down. 

 - A lot of these units have a drain hole in the side of the outer bucket up near the top.  It's there so that if you get overenthusiastic with adding ice, the excess water will run out onto the table, countertop, your lap, the floor, or whatever, instead of getting into the ice cream.  Salt in the ice cream would be a not-good-thing, so don't block that drain hole.  Unless you're making a batch of ice cream that's close to the capacity of the unit, you really only need to keep the water and ice level about two inches above the initial level of the ingredients inside the canister.  For a small batch, that may result in the bucket being just half full; you don't need to worry about overflow at that point, but you do have to be concerned about the possibility that all of the ice will melt.  It's best if there is at least some ice down at the lower levels, and plenty at the top for the salt to work on, but I don't recommend stopping the churn to pour out water; just add ice and salt, and keep going.  If the water level is close to the hole and you've got ice piled up above the rim of the bucket, there will be some overflow shortly.

 -  The instructions I have found (both with the unit and elsewhere) call for crushed ice.  Don't sweat it; if all you can find is cubed, it will do, but you will have to do a lot of scooping of more cubes into the bucket as the cubes melt. Those same instructions call for rock salt.  If you live in an area where rock salt is sold in huge bags dirt cheap because of snow and ice (or other really hard water), using rock salt makes sense.  If, however, you live where this is not something that you need to keep handy, then any kind of salt will do.  Buy whatever is cheapest per pound without having to lug home a huge bag.  Here, that's the local store brand non-iodized table salt.  It worked just fine, at less than a quarter of the price of rock salt that was sold for use in making ice cream.

 -  Unless you like frozen desserts A LOT, or you're having a party for a dorm full of teenagers who generally treat the tableware as the fourth course of the meal, a bigger batch just means that you're going to have to finish a whole lot more of that flavor before you can move on to the next one.  Experiment with small batches - but bear in mind that a batch much less than a quart of finished volume (figure that's what you'll get from 20 to 24 ounces of initial mixture) may not churn very well.  And the cleanup task is identical regardless of how little you made.  (If you make way too much, it can get larger than normal, of course...)

 -  With great ingredients come great frozen desserts - sometimes.  Taste the mixture before committing it to the churn; if it's not right, fiddle with adding more of whatever's needed until it comes out the way you want it.  It's much easier to fix a batch that hasn't gone in yet than to rework one that came out wrong.

 -  If you decide to use high-fat coconut milk as your base, eat the final product somewhat sparingly.  Coconut oil has an annoying laxative effect when consumed in excessive amounts - and that amount varies from person to person.

I haven't tried it yet, but I suspect that this thing might be able to make frozen daiquiris.   That is an experiment which will have to wait for the days immediately before a party...

Airtex E3372M Fuel Pump Failure

In mid-2010, the fuel level gauge sending unit for our 2000 Pontiac Montana became completely unreliable; the gas gauge would swing from full to empty and back, randomly, at any time.  Money is never all that freely available around here, so it was obvious that this would be yet another do-it-yourself repair, and after some careful shopping online I found an Airtex E3372M fuel pump from a supplier on eBay for about half of the cost that local parts stores were asking.  Since all of the aftermarket brands were priced about the same nearby, I didn't think there would be a tremendous difference in the quality - and we certainly couldn't justify paying well over $300 for an AC Delco or Delphi unit.  (The price at the dealer was even more exorbitant, as expected.)  The pump arrived, and with the exception of the fact that the supplied gasket was much too small (I had to obtain one from the dealer in order to be able to install the pump) the repair went smoothly.  The Airtex pump's pressure was within spec, and the fuel level gauge was working properly again.  The only thing I noticed right away was that the Airtex pump was significantly noisier than the original one, but as it wasn't at a level that was audible when in motion, I didn't really care.

It operated well for about two and a half years.

On Friday, January 11, 2013, while we were on the road from Houston to Atlanta, it began displaying failure symptoms.  The first sign of trouble was stalling immediately after starting in Montgomery, Alabama.  After checking several things under the hood, the engine started and ran near-normally, so we continued the trip.  There were intermittent hiccups at times throughout the weekend, but we managed to make it back home on Monday.  And then, on Tuesday, the failure went past the point at which the vehicle could continue to operate; I barely managed to get it back into the driveway.  A test of the fuel pressure found it was sitting at under 20 psi when the pump was running, and it fell to 0 immediately when the pump shut off.  That pressure is supposed to stay at over 20psi when the pump stops, as an aid in quickly restarting the engine. 

Once it had been confirmed that the problem was in the pump, a replacement was obtained (NOT an Airtex this time). The tank was drained, and the failed pump was removed.  As soon as it was out of the tank, the exact failure became obvious.  The flexible hose that connects the actual pump to the mounting flange had slipped off of the nipple on the pump.  The reason why this was possible was disturbingly obvious; the nipple had no barb of any kind to help retain the hose in place.  This is a clear error in the pump's design; that pressurized connection is pretty much guaranteed to come undone without a barb to provide a physical barrier to hose slippage.  It was not a question of whether the unit would fail, but of when.  Had I noticed this flaw when I was installing the Airtex unit, I might very well have rejected it and sent it back - but unless you're aware of the mistake, and know where to look, it's not obvious. 

Most of the parts stores that sell this pump provide a one year warranty on it.  As the failure involved is a slow one, that means they will get very few of them back under warranty.  Frankly, in my opinion, these pumps should be recalled and replaced with ones that have the defect corrected.  A failure in this part in traffic could easily result in an extremely dangerous sudden stall in traffic, possibly placing the driver and other occupants of the vehicle in considerable danger. 

It is possible that Airtex may have corrected the flaw in later production, but I have not attempted to determine if that is the case.  Certainly, they are still using the same part number for their current units, so I would not expect it. 

I will note that a careful inspection of the replacement - obtained from NAPA, and showing the Carter brand name - revealed that Carter does, in fact, include an adequate barb on their pump output nipple.

Here's a photo of the failed Airtex unit showing the hose that had disconnected itself, and the nipple onto which it had originally been clamped:

Note that the hose nipple (which is just above the end of the hose in the photo) is completely smooth; there's no barb or ridge of any kind there, which there very definitely should be.

I would strongly recommend close inspection of any Airtex brand pumps for this design fault, and unless there is some other mechanism provided by which the hose is affirmatively retained in place, I cannot recommend the use of these units.  The pinch clamp, in the absence of a barb, clearly was not enough to keep the hose from slowly slipping off of the nipple under pressure.  Although there might, theoretically, be some kind of adhesive that could also help to retain the hose, I would hesitate to trust that in the presence of the very active solvent that is modern gasoline.

Is eBay taking a Facebook approach to changes now?

Facebook has a long and inglorious history of making changes that people hate, and implementing them in a way that makes it hard to avoid being subjected to them.  But since Facebook is a free service which (with only very limited exceptions) provides no real opportunity for costly misuse, such changes primarily just elicit grumbles and little more.  It's not like you spend money there, after all.  The same is not true for Ebay and PayPal, however, and there's a fresh new misfeature over at eBay that's a humdinger of an example of a designed-in security breach.

Until recently, when you won an auction or selected a Buy It Now item, the process of paying for the purchase required logging into your PayPal account for each and every transaction. That's as it should be; it guards against a number of things that one should rightly be paranoid about, the most important being an accidental security breach that allows someone to get your eBay password and use your eBay account to get access to your money. Now, however, what pops up in place of the PayPal login page is a "helpful" box extolling the virtues of linking your eBay and PayPal accounts together so that you don't have to enter your PayPal password to complete the transaction. This dialog box has two selections: "Yes, link them together now" and "Not now." There is no selection for "Not just no, but HELL NO!"

Whatever passes for a brain in their marketing department must have been on vacation in Bermuda - and had the entire security division off getting drunk in a bar somewhere while it was gone. This is perhaps the most egregious example of blatant idiocy that I have seen from these folks so far, bar none.

Did it occur to any of them that this allows a one-password breach to permit emptying a bank account entirely? Apparently not, but here's just one way that it could work: Assume, for a moment, that by whatever means, your eBay password has been phished (or grabbed by a virus, or some such) and is now in the possession of someone who wants your funds. Anticipating that they would have this opportunity, they've set up half a dozen eBay accounts and PayPal accounts as well. With your eBay password in hand, they set up several listings in their eBay account for things that have a Buy-It-Now price of three or four hundred dollars each, and then using your password from a system that's on a different connection, they proceed to buy the spurious item and pay for it from your account. Yes, you'll get an email congratulating you on winning the auction - but if they're smart, the item they will list is going to be *identical* to the last one you actually bought (except for the price) so that you'll look at the notice and think "That's odd" instead of "Holy shit, my account's been hacked!" Some people probably wouldn't notice at all - and if you're an active eBay user, they can time it so that it superficially just looks like the same notice accidentally got sent twice. As long as the payment can't happen without also having your PayPal password, this kind of approach gets them nothing unless they have both logins. But with the accounts linked, they need only the eBay login, and they can swiftly empty your bank account. If you're not paying close attention, that money can be cleanly gone and unrecoverable in a very short time.  I should point out that if I can see the potential for this kind of exploit, you can bet that the phishers can, too - and they've probably figured out a dozen more ways to use it, probably with less work and faster results.

eBay has had problems with account security breaches in the past, with varying exploits involved. Some compromised accounts were used to list hundreds of nonexistent items, and the account information was changed so that the email went to the invader instead of the account owner. Various other breaches have had other goals. With the eBay and PayPal accounts linked, however, the possibilities for ill-gotten gain by phishers have just been multiplied. And although I do not know if they have provided a way to undo the linking, it would not surprise me if this change was next to impossible to undo. At the very least, if they had that option available, they ought to say so - being able to temporarily link the accounts might be a useful choice in certain limited circumstances. It would still not be a hazard-free choice, but it would be at least potentially safe in careful hands.

Save your data - from bells and whistles.

For several years now, lots of computer systems (and motherboards from which to build them) have been supplied with a feature known as RAID support. RAID is (currently) the acronym for Redundant Array of Independent Disks. (The "I" used to be "inexpensive", but that's been changed to reflect the fact that some technobabble purveyors have found that peddling very pricey RAID setups is profitable indeed.)

Losing your computer's data is a bad thing.  Lots of people rely on regular backups to minimize the amount that vanishes when a hard drive (or whole system) gets toasted, but sometimes that's not considered good enough.  A different approach is the multiple-hard-drive solution known as a RAID.  When it works properly, it reduces downtime for recovery from a single drive failure to near zero, and permits you to simply keep working uninterrupted.  (For a full system failure, it's still vulnerable to problems; that's why even a RAID doesn't make real backups unnecessary.)

On the surface, some of the implementations of RAID look like a really good idea, particularly the simple RAID1 version; it appears to just create and maintain a full-time backup of one disk on another identical disk. And if that were the case, it would be laudable indeed.

Unfortunately, that's not the case. The hardware used to create each of the RAID versions causes the data format of the drive(s) to be changed in a manner that makes it difficult or impossible for the drive to be accessed with just an ordinary controller. For RAID version 0, and for versions 2 through 6, the advisability of using the scheme is even worse; they not only change the format, but they don't even store all of the data together; it gets split up and scattered, rendering the process of recovery far more difficult if something serious goes wrong.

An extensive description of the seven principal types of RAID is available over at Wikipedia, but the bottom line is pretty simple; using RAID increases the complexity of your data storage, and almost always makes recovery harder if you have certain types of hardware failure. For some types of RAID, recovering data after a *single* disk failure becomes easier - perhaps even trivially simple - but as you increase complexity, the number of things that can fail goes up, so the chances that there will be a failure of some sort increase with it - and no RAID can save you from the effects of the operating system going nuts due to a virus or a badly written program, and essentially all hardware-based RAIDs will make your data inaccessible if the controller fails.

All things considered, unless you really, really need either full-time continuous backup or the small increase in performance of a "striped set", even RAID1 is probably not a good idea. The possible exception is in using the RAID capabilities that were built into certain versions of Windows itself; a RAID1 setup done that way is generally as safe as a single hard disk, and provides added protection against data loss due to sudden failure of one of the two drives.

The main reason why I'm posting about this is that apparently, some of the system builders have been shipping machines that have their drives set up using RAID0 from the factory, without providing a Windows reinstallation CD to allow the buyer to convert the machine back to conventional drive formatting. Sometimes this is done to allow the system builder to advertise a unit with what appears to be a huge single drive when in fact it is equipped with two smaller ones that just show up in Windows as a single unit because of the RAID setup. When shopping for a new or refurbished system, it appears that it's wise to ask about these issues. If the machine has only one actual physical hard disk, it's not an issue; any RAID requires at least two drives. But if the box has multiple drives, or an unusually large one, it's wise to make sure to ask about how that's been done - and whether there's a Windows reinstallation CD included.

A tale of two take-outs; how not to eat fast food.

It's possible to eat low-carb and still have fast-food meals, but you really have to watch what you eat. Here's one example.

Boston Market, a fast-food chain owned by McDonalds, features rotisserie-grilled chicken and a couple of other meat dishes plus a selection of sides. They have some very low-carb items available - as well as many that aren't. Their chicken is so close to zero carbs as makes no difference, while their meatloaf evidently has considerable added starch. So I did a breakdown of the carb content of two meals using their own website's numbers for the content. (I've subtracted the fiber from the raw carb amounts since fiber can be ignored when watching carbs.)

Meal one: Half a rotisserie chicken with green beans and steamed vegetables, no cornbread.
Chicken: 2 grams carb
Green beans: 4 grams carb
Steamed vegetables: 5 grams carb
Total: 11 grams carb, 760 calories, 38.5 grams fat, 88 grams of protein.

Meal two: Large portion of meatloaf, mashed potatoes with beef gravy, sweet corn, no cornbread.
Meatloaf: 29 grams carb
Mashed potatoes with gravy: 36 grams carb
Sweet corn: 35 grams carb
Total: 100 grams carb, 1195 calories, 61.5 grams fat, 54 grams of protein.

Adding the cornbread (which is included with the meal automatically unless you tell them to leave it off) would boost the carb level on either of these by another 29 unneeded grams. I've left it out of both to be fair.

The meatloaf dinner has nine times the carbs of the chicken dinner in this example - in part because of the choice of side dishes. (It also has more than half again the calories, largely thanks to those very same carbs - and at 760 calories, that chicken dinner has plenty of calories already!) Okay, so what happens if you use the beans and steamed vegetables with the meatloaf? You've still got more than triple the carb that's in the chicken dinner; that's because the meatloaf is loaded with starchy fillers. And if you go the other way, and have the mashed potatoes and corn with the chicken, you get 73 grams of carbs in the meal. Any way you look at it, everything in that meatloaf dinner is bad news, and these aren't even the worst possible selections; get the meatloaf with sweet potato casserole and cinnamon apples, and throw in the cornbread, and you'll have 176 grams of carbs - enough to keep your insulin level elevated for many hours; maybe all night. That's why choosing the right food is essential.

What you eat is more important than how much.

When eating out, if you avoid the stuff that has the starches and sugars, you really can still go low-carb in many places. However, if you choose the wrong components to go with your low-carb main dish, you can lose the advantage rapidly. Having a salad? Ditch the croutons and the thickened-with-starch dressing; use oil and vinegar instead. Stare those breadsticks back under the napkin. Ask for something else instead of grits or hash browns or toast with breakfast. Disdain the beguiling bran muffin or whole-grain bagel with its 50 grams of non-fiber carbs (many are that high or higher). It won't be long before you won't miss the sweet, and you certainly won't miss the pounds that you'll shed as a result.