[interesting] Database Encryption

[Robert Brenstein]

>Agree. In ideal it must be:
>
>IF you have db with existed data,
>And you say encrypt it,
>Valentina must encrypt existed data,
>And ZERO ONCE (!) all empty space.

So we agree (my email suggesting exactly the above hasn't showed up 
on the list yet, although it was sent a while ago).

>Problem like in Microsoft was fixed on first month of born of Valentina.
>Allocated disk space is zeroed.

I did not mean that so literally, Ruslan. If a DB exposed private 
content because of deletion side-effect, the scandal would be LIKE 
(as in similar to in implications not technical aspect) what 
Microsoft faced with their Word file problem.

>Robert, in fact in above situation mistake is of developer/user.

Yes, but it may be quite a common situation, which Valentina can 
prevent it without much ado as you outlined. If Valentina protects us 
from making a mistake, the better for it as a product.

>BEFORE he have decide encrypt db, that data was NOT considered as secret.
>They was even deleted before encryption. So why Valentina must consider them
>as secret ?

Well, an easy example is that a db may be intended for encrypted 
distribution but kept un-encrypted during development. Anyway, what 
you suggest above is the right way to prevent any mishaps. Just do it 
:)

Robert