SuperMemo horror. Data lost!

problem

marius wrote:

I keep very important data in SuperMemo. I assumed the program has been debugged to perfection after 20 years of development. However, yesterday I found a page that has been replaced with garbage! This is very important text that I wrote myself and that cannot be re-stored from the web or other sources. My trust in SuperMemo plummeted! Please tell me how I can trust SuperMemo if I can randomly lose data. What is worse, I might have not discovered the fact for years. By that time, all my backup copies might contain garbage. This would kill the information for ever!

I am sending the damaged collection and the correct backup for you to tell me what happened! I do not see alternatives to SuperMemo. However, I also do not have an idea how such a case could be prevented! Collection repair does not even see the problem!!! Interestingly, I could find the lost text using Ctrl+Fand the correct search term, but what I found was the garbage text.

answer

You can trust SuperMemo as much as you can trust your hard disk to keep your data. Some corruption is always possible and the risk will never be zero.

In your case, one HTML file was corrupted. Most likely, some sectors on your hard disk got swapped and your file was replaced with binary data of some executable/application.

Your text was still kept undamaged in the text registry and this is why you could search for it (text registry keeps the unformatted version of the same text that your HTML file kept in the formatted variant).

Repair collection does not help because it is legal to keep different versions of the same text in the text registry and in formatted HTML file. In your case, the formatting was replaced with meaningless binary file though.

The only way to minimize the risk is to keep regular backup archive (e.g. on DVDs). The more often you back up, the lesser the chance you will lose information. The text would need to be added and then lost in between two back-up copies.