Please don't throw this away

The “Save” command in desktop software is a consequence of hardware design. Because primary storage (RAM) is fast, but expensive and volatile, and secondary storage (hard disk) is slow, but cheap and permanent, there is a trade-off between the choice of primary vs. secondary storage. Hence the “Save” command, which instructs the computer to move your data from primary to secondary storage.

But as Cooper argues, this is idiocy. Why should you as a user be bothered with this? When you enter stuff into a computer, it’s usually because you want the computer to keep it for you, not because you want it to turn around and throw it away.

Thankfully, lots of modern software doesn’t in fact bother you with this. There’s no save command, the software automatically saves your work for you. But this bargain only works if you can be reasonably certain that it will in fact save it for you.

Most software sans Save will only transfer from primary to secondary storage on exit, which means if the software crashes, either because it has a bug, or because some other software has a bug and brings this software down with it, you’re out of luck. You might lose hours and hours worth of work. This is even worse than the Save command, because this forces users to routinely quit and relaunch theirs apps, just to safeguard our data—a Save button, while still idiotic, is certainly less so.

So here’s what I want to ask of software makers: Save our stuff on a regular basis, not just on exit.

Transfer it from primary to secondary storage every minute or so, when there are changes to data, in the background. It should be fairly trivial and computationally cheap to have a dirty flag, so you don’t need to do this when there are no changes, and to check that dirty flag every minute, to write the latest version to disk, so you’re secure in the case of a sudden crash. In the trade-offs involved in software design, keeping your user’s data safe should always take priority. If people lose their data, your app is useless.