On the Topic of "Format Versus Delete"
One topic that comes up periodically is whether to delete images from your cards or format them. Usually the consensus recommendation ends up being formatting. But do you know why?
The usual line of reasoning is fairly consistent. Your compact flash or SD card contains more than just images; it essentially has a "table of contents" called a File Allocation Table (FAT) that keeps track of what all is on the card and where it lives. When you delete an image from your card, that File Allocation Table gets updated to reflect the fact that the space that image used to be is available for use again. But the FAT itself may get corrupted over time, leaving it broken. The card may still contain images, but if you can't get at them, they're effectively gone. Reformatting the card not only deletes all the images in one fell swoop, it rebuilds the File Allocation Table so everything on the card is guaranteed to be back in working order.
The FAT format used on your card is a standard technology so it's possible to format the card in either your computer or your camera. This has to be the case or else you wouldn't be able to read the card on your computer. But while it is indeed a standard way to format a storage device, there are several quite similar standards that have evolved over the years. So while it is possible to format a card in your computer, it's generally best to do so in your camera so the FAT structure ends up being the exact variety preferred by your camera. If you format it in your computer, everything should be fine, but the best way to make sure is to let your camera do the job.
All this makes reasonable sense, and if nothing else one quick format in the camera is easier than a whole bunch of individual deletions. But there's more to it than this.
A common misconception is that deleting and formatting are basically the same — that both write zeros over everything in their path, such that deleting everything leaves you with the same thing as if you had formatted the card. But this isn't true. Formatting the card resets only the File Allocation Table and does nothing to the images themselves. Only the information in each sector about what is on the card changes, not the data actually in each sector. Essentially, it tells the card there's no longer anything on it, without actually removing the images. When you next need to write an image to that card, the FAT will say there's plenty of space available and away you go. Your image will be written overtop of what was on the card just as if it wasn't there. The old zeros and ones will be ignored and all the new bits have been written the FAT will be appropriately updated.
In some cameras and on some cards, formatting will also mark any bad sectors it finds on the card as unusable and map around them. Manufacturers know that everything isn't perfect and plan for this eventuality, building cards with spare sectors specifically for this purpose.
Cards wear out over time too. Any given location can only be changed from zero to one or back to zero so many times. Manufacturers know this too and use complicated methods of relocating fixed structures like the File Allocation Table so they don't cause wear on the same locations. Each time the card is reformatted the card gets remapped. If you never reformat it, this never happens and the heaviest wear will always be in the same locations.
Cards can suffer from read errors too, and manufacturers know this as well. They are built with error checking and correction (ECC) circuitry to retry read requests without user involvement. There are limits though, and at some point retrying a read a failed read will itself fail. This is the point you want to avoid getting to.