Is it safe to leave a flash drive plugged in all the time?

First I'll answer your questions, in reverse order:

What wears out? You're correct, there are no moving parts, but there are electronics. The type of electronics that are being used in flash drives actually wears out a little each time something is written to it. Eventually attempts to write further can be lost because of that wear.

Perhaps the best analogy I can give you are old style CRT screens at airports and in cash machines that display essentially the same thing for months on end. Eventually, the coating underneath the glass display wears such that even when the device is turned off you can see a ghost or pattern of what had been displayed there before. This is called "burn in" and is the reason that screen savers were invented. Eventually that ghost can become so prominent as to interfere with the readability of whatever is displayed when it's turned back on.

Flash memory doesn't really "burn in", but it does wear in a conceptually similar way.

Why "Safely Remove"? In an effort to make accessing the device faster, Windows may "buffer" data before writing it to the actual device. Safely Remove forces Windows to flush everything to the device. NOT doing so, but just yanking the device out without warning, could then cause the files - or worse the file structure - to become corrupt and unreadable.

Writes to flash memory are expensive (slow) and as we've seen causes wear. In addition, writes only happen in multiples of the sector size - typically 512 bytes. So let's say you write 100 bytes of data to the device, wait a while, and then write 100 bytes more. There are two approaches:

• Take your 100 bytes of data followed by 412 bytes of anything and write that into a 512 byte sector. Then on the second request read the sector back in, add the next 100 bytes of data following the first 100 bytes, and write it back out.

• Place the 100 bytes of data into a 512 byte buffer in your computer's memory or RAM. When the second request comes along add the next 100 bytes. Later when the file is closed or more data is added to fill the sector completely, write the data you've collected.

As usual that's a gross oversimplification to make my point, but as you can see the first scenario has three operations to the device: a write, a read and a write, whereas the second has only the single write. The optimization is simply that writes to the device are delayed until they are necessary. This maximizes speed and minimizes the wear on the device.

Safely remove simply makes sure that any partial data managed by the operating system that is waiting to be written to the device actually is. In addition, it also checks for applications that have files open on the device that may be in a similar partially-complete state.

Should I leaving it plugged in? This is where "it depends" really comes into play. And the problem of course is that it depends on things that actually aren't all that easy for you to see.

What it depends on is easy: is the device being written to - even occasionally - while its left plugged in? I'm not talking about your work, I'm talking about other programs like your operating system, anti-malware software, utilities and other things that - on the surface - would have no reason in the world to write to the device. And yet some do. Maybe.

Me? I wouldn't. The only way to be sure is to remove the device to maximize its life.

And for the record, any time your computer is powered down - be it off, hibernate or stand-by - there's no issue. No one's writing to the device when the machine is turned off. If, however, like me you leave the machine on 24 hours a day, then there's a small risk.

So, how big is the risk?

You're going to hate me for this.

It depends.

It depends on the quality of the flash drive. It depends on what software you have installed on your machine. It depends on how your USB or other connecting port is configured (some of the buffering might be turned off). It depends on how long you plan to keep your flash drive. It depends on how catastrophic it would be to you if it failed.

For many - perhaps even most - there'd never be a problem. Heck, most people would probably lose the flash drive before it actually wore out.

And the quality of flash memory is constantly improving. High end (read: expensive) flash memory is being used to actually replace hard disks, so you know that while they still wear out their lifespan has been increased to the point of it no longer being a practical issue. But that inexpensive flash drive you picked up? It's not going to be that same high quality. In fact, where it'll be on the quality scale - a scale that is constantly moving - is anyone's guess.

But the risk is definitely there. And Murphy's Law being what it is, failure will happen when it's least expected, and most inconvenient.

Mitigating the Risk

Here's what I recommend you do.

Plan for failure.

Just assume that sometime out of the blue a write to your flash drive will fail. Or that a later read from it will fail.

Make sure that through some form of redundancy - periodic backups come to mind - you can recover to a convenient spot. Perhaps that means you need to save two copies every time (one to your laptop, one to the flash drive), perhaps that means you're ok losing a weeks worth of work. In any case, do something that makes sense for you.

It's always important practice to have at least two copies on different media of anything important. That's advice that's independent of, and certainly predates flash drives. Heck, back in the day floppy disks wore out too.

And let's face it - even if your flash drive doesn't wear out you could lose it, with the same potentially disastrous results.