So, I'm making this thread for two reasons. 1) Everyone likes to tell their story. But 2), this topic can give people some insight into just what they can do to prevent data loss.

The general idea here is to give:

Lost:
   What you lost.

Cause:
   The mistake you made that caused you to lose your data.

Solution:
   What you did to prevent this from happening again.

Now, if you didn't actually create a solution for your loss, then you can't post what you lost . So, if this happens to be the case for you, and you really want to share your story, then you'll just have to figure out a solution so that you can't delete your stuff again.

I think one of the great changes that this thread can create is when someone goes, "Crap, that could totally happen to me," and then they see a solution that can prevent it.

So mine:

Lost:
    I lost the entire source to my Impossible Game.

Cause:
    I was testing out a new batch (.bat) file to assemble my assembly sources. In this test file, I accidentally moved the source code rather than copying it. Since the recycle bin doesn't catch stuff deleted in batch files, the source was gone forever.

Solution:
    I added an extra folder on my C drive (C:\backups) and whenever I assemble an assembly program, I have my batch file add a copy of it to the backups folder. This won't protect against a hard drive failure, but it does protect against accidentally deleting a file or folder.

I was testing out a new batch (.bat) file to assemble my assembly sources. In this test file, I accidentally moved the source code rather than copying it. Since the recycle bin doesn't catch stuff deleted in batch files, the source was gone forever.

This is also very true on Linux... rm something means that it's gone. Forever.
However... there *might* have been a solution for both you and me (assuming I deleted a file in Linux) - using TestDisk to recover the file. It's too late though since by now all the cumulative disk writes probably deleted it forever.

Anyway, here's my story:

Lost:
A PRIZM toolchain building shell script that worked really, really well. (Linux)

Cause:
Using the nano text editor, I didn't realize that when I was saving, I didn't change the file name to something else, since I did some heavy modifications/deletions to make the script serve another, almost completely different function. I tried recovery via TestDisk... unfortunately, this is one of those cases where you can't recover the data. It's gone - forever.

Solution:
NEVER do nano important_shell_script.
Do cp important_shell_script some_script; nano some_script.

(In English: don't edit the file directly, especially if you are planning to save it to a different filename. Instead, COPY it to the filename you want, and edit that.)

Also, BACKUP BACKUP BACKUP. I usually would stick important stuff in my Dropbox* folder, but this was my server, and I didn't want Dropbox syncing there (especially with personal info). I'm planning on setting up a rsync backup functionality on my server... but in the meantime, LOCAL backups are very effective as well.

Lesson Learned:
(new heading I decided to put in, although you don't have to!)
Even if you think you've backed up, BACKUP again and AGAIN. (I thought the script was saved somewhere, and that nano would make a backup copy if I saved over. Both, unfortunately, are totally WRONG.)
It just might save you.

* Dropbox = amazing file hosting service that is so dead simple - copy the files you want backed up into a special Dropbox folder, and it will sync automatically! It also has (limited) file revisions, which is great if you messed up on something. This plus regular source code versioning (SVN, Mercurial, Git, etc.) is a pretty awesome backup solution!

Hard drive crashes. It happens without any reason. That's all. Be sure to have all your important files at alternate places.

Quote from: Juju on April 20, 2012, 09:57:34 pm

Hard drive crashes. It happens without any reason. That's all. Be sure to have all your important files at alternate places.

Places as in not only different storage mediums, but also different cities/buildings, such as a server online somewhere in the world.

Exactly.

Lost:
   The Game
Cause:
   Jimbauwens
Solution:
   Slap him with a large trout on IRC.

I would never do something like that!

I never really had a bad disaster, since I always keep copies.
I think all calculator games/programs should be open soource, so they can never die.
My programs all are, and because of that I share pre-release code with my friends (Adriweb, Levak, Critor, etc) making them a backup medium too
And github is a good help too

Quote from: DJ_O on April 20, 2012, 08:07:19 pm

Lost:
   The Game
Cause:
   Jimbauwens
Solution:
   Slap him with a large trout on IRC.

I would never do something like that!

I never really had a bad disaster, since I always keep copies.
I think all calculator games/programs should be open source, so they can never die.
My programs all are, and because of that I share pre-release code with my friends (Adriweb, Levak, Critor, etc) making them a backup medium too
And github is a good help too

The problem with open source is that I already saw people adding their name and/or remove names in the credits (not people from the community, just guys thinking they were smart in my school).

---

Lost
A bit of progress

Cause
I don't know. It may be zStart or not. I just remember that I unarchived the source to group it, then I rearchived it and everything I tried with the source threw "ERR:Memory", except archiving/unarchiving it again. Then, sending the group to the computer didn't work either (invalid format or some other stupid reason).
I keep having bugs like this. Once I tried to send the source to another calc and the beginning of the program was missing . Some other time, I used Rcl on an archived prog and it copied hieroglyphs among my code but compiling worked well

Solution
Backup a lot. I have a lot of copies of my source on two computers (more of them on my sister's computer because of the touchpad drag and drop thing )

Yeah this is why, if I remember, KermMartian decided at one point to no longer release the source of his programs, unless someone on calc forums that has a decent amount of activity and can somewhat be trusted asks for it privately. He often was victim of program theft in the past. The worst case of all gotta be Contra 83, though: Every few month I saw it appear in someone else's author profile on ticalc.org with a different author name in the readme, and this lasted about 3 years.

Btw NEVER release the source of one of your program before you have announced it or released the executable of it on a popular calculator forum!!!!!  In the past I saw people releasing source anonymously somewhere or on a forum they were not very active on and then someone released the program, compiled as his own work, then it became hard for ticalc.org to figure out who to trust as the real author, as the theft released his copy before the original author.

I've released code to several people in the community (such as Adriweb) before release. I know I can trust them, so I'm not worrying.