[HFTetramorph]

If you need several databases for a single project, then there's something wrong there. For an average sized project the most you probably will need is 2. (1 is usually enough...but I'll be lenient and say 2)

My bad, mis-phrased that:
Meant to say "Several databases each for their own project"
I believe the most seperate databases I have used for a single project was three.. In which one was for session management, another for standard record and user storage, and another used as a keyring for encryption means.

[Retribute]

Most people tend to forget (or simply do not know) that priority should be given to length over complexity over the password.

Let us say that we have two passwords:

1. C0mpl3x!
   
2. ThisIsMyPasswrd
   

The latter would be harder to crack.
But of course, long and complex passwords are the best.

[Crypter_]

My bad, mis-phrased that:
Meant to say "Several databases each for their own project"
I believe the most seperate databases I have used for a single project was three.. In which one was for session management, another for standard record and user storage, and another used as a keyring for encryption means.

This would be why we have something called tables. You use one database and store different types of data in tables... Jesus. Depending on the database software you use you're also probably killing performance by doing this.

[Pillows]

Most people tend to forget (or simply do not know) that priority should be given to length over complexity over the password.

Let us say that we have two passwords:

1. C0mpl3x!
   
2. ThisIsMyPasswrd
   

The latter would be harder to crack.
But of course, long and complex passwords are the best.

The first one would actually be harder to crack.

[Retribute]

The first one would actually be harder to crack.

No, it wouldn't. It would be very easy to crack that.
Run it through http://passfault.com/

[Pillows]

No, it wouldn't.
Run it through http://passfault.com/

http://www.yetanotherpasswordmeter.com/
http://www.passwordmeter.com/

Say other wise.

But of course every one will give different results.

[Crypter_]

Going back to the actual post. You would need to "hack" every .com domain six times to actually get these many databases (not that you would get unique databases each time).

The number is obviously bogus in that context. The site leakedsource[]com claims to have that many records, though, which makes much more sense (but is definitely inflated).

[Retribute]

http://www.yetanotherpasswordmeter.com/
http://www.passwordmeter.com/

Say other wise.

But of course every one will give different results.

This is a pretty accurate visual.
[img]http://imgs.xkcd.com/comics/password_strength.png[/img]

I also know Stanford recommends using Pass Phrases over complex passwords because the increased length provides so many possible permutations that a standard password-cracking program will not be effective.
https://uit.stanford.edu/service/accounts/passwords

[Ciphered]

Looks good, personal security is always key.

[HFTetramorph]

This would be why we have something called tables. You use one database and store different types of data in tables... Jesus. Depending on the database software you use you're also probably killing performance by doing this.

Not at all.. More databases != lower performance. While over time the more and more data/sections for data can start to result in slower performance, adding different databases will not "kill" performance, as if it did then most would stop using MySQL for things like shared hosting.

The problem is you are preaching the idea of a perfect database being a thousand tables each allocated different users? That doesn't even remotely make sense and if separate projects are to be using the same database server, then they should likely have different databases to improve optimization over the different projects.
Why in any world would I want cross site session data to be in some random project's database? If I am going to create something that has more than one use for a specific project then it needs to be its own database, not only for neatness but it can also come down to a security factor as well. Example: My keyring I mentioned earlier does not allow any interaction with the files and is completely self contained. You provide it a UUID and the file, it does the rest and returns the decrypted raw data. The idea is to keep the Keyring on a separate self contained system in the event anything is vulnerable, the key ring is still safe. It makes sense, not only organizationally, but generally the point you are arguing over makes no sense to defend..

However, to continue this discussion I recommend creating a new topic or PMing me, as I do enjoy debating over this kind of thing for fun. Just thought if I were to make a reply, I would make it count :P

Going back to the actual post. You would need to "hack" every .com domain six times to actually get these many databases (not that you would get unique databases each time).

The number is obviously bogus in that context. The site leakedsource[]com claims to have that many records, though, which makes much more sense (but is definitely inflated).

But not all databases are meant 1:1 per domain, I have a domain using 10 separate databases each for a separate project, If I do it, how many others do as well? Domains are as related to database usage as pizzas are. To have that many databases you need to have 500 million ate pizzas.. While domains and databases are both used in web services, they are not similar nor linked in any way other than DNS.

[Crypter_]

Well, I can PM you about the first bit, but the exact number in the OP is literally cited as records on an external site, not databases. The OP just wrote the wrong units.

[NurSpec]

Very informative thread, it's a shock that people don't know this stuff already however.

[Vitani]

This is a great thread, especially because of recent events regarding the possible leak. Thanks for making this thread man.

[Power8]

Thank you to all who find this thread helpful!

I worked really hard to make it informative.