When I looked at some really old posts, I see that they are garbled and below them it says:


[Please note that this post has been rebuilt from a corrupted database.]


Why? What happened to corrupt the database so long ago?

I *think* it has to do with this:

I think you're correct. Since the backups that were kept only held on to keywords (throwing away stuff like "and," "the," "should," etc). But now our backups are more robust, if we were to have a corrupt or lost database we'd be able to restore the posts 100%.

I feel like it's worth mentioning that we would probably lose some posts, depending on how long ago the last backup was, but the posts that were restored would be 100% complete.

What? This is hilarious! How could that possibly have been considered a "backup" by any means? The result is also pretty funny - just click "View next topic" and the hilarity continues.

My wording was wrong and I should have chose a better word, this happened before my time so the details are not my first hand account. But, I think we used the keyword database as a backup, since having the posts was more important than leaving them behind.

It's also possible the backups were originally complete but the corruption was that everything that wasn't a keyword was purged. We'll have to wait for a proper first hand account should one ever surface. I recall this being talked about before but a quick search doesn't show anything. It could have just been discussed over SAX/IRC.

While I have no firsthand insights to offer, I assume they would have been rebuilt from a search index, since the site's built-in (no longer used) search builds an index of words in posts and maps each word to a list of posts that contain it.

Yeah, I was going to suggest what Tari said. The ticalc.org post databases have similar search indexes