We NEED your help!

[radix]

We NEED your help!
From the beginning, hashes have been using an incremental numbered format. This worked well when there were 10 or 20 types, but as support for hashtypes grows, we need YOUR help to come up with a more logical way of doing this.

The problem is as follows:

MD5 is currently grouped in increments of 10. Plain MD5 is as we know '0'. For each salted type, we add 10. This becomes an issue once we get 9 salted types because SHA occupies 100+.

The fragmentation becomes apparent when you see MD5crypt, which has a base of MD5, numbered as 500, or, SHA256/512crypt as 7400, and 1800 respectively.

We have mulled over this for the better part of 6 months with no clear solution. Previously, we had just realigned the numbers to create a standardization between hashcat and oclHashcat-*, but it is clear this method should be changed.

So we ask you, the users, for input on how you think this could be better accomplished. I'll leave this thread open for posts with your opinions, and the dev team will take your recommendations into consideration and credit the selection (if any) in the release that implements the change.

[epixoip]

maybe use a string instead of an integer. use standard names for the well-known algorithms (raw-md5, md5crypt, etc), and come up with a logical naming convention for the lesser-known algorithms (like maybe md5-salt-pass, md5-pass-salt, etc?)

[magnum]

I agree with epixoip. And after that is implemented, have someone write a bash completion script (dynamically parsing the -h output).

[epixoip]

bash completion would be awesome

[Rolf]

A hybrid of both, ordered by cracking speed (gradually coming up).
As a result, each algo may have a number assigned to it (every range frees up for each unique algo name).

-m md5-0 == md5
-m md5-1 == md5($pass.$salt)
-m md5-2 == md5($salt.$pass)
etc

-m sha1-0 == sha-1
-m sha1-1 == sha1($pass.$salt)
etc



Alternatively, the ordering system can just include the real algorithm name, as a string, which is already mentioned here as "Hash-Name".

[Algreen]

I'd like to help but I just understand 50% of your writing!
Use of standard names would be appreciated.
thanks
A.

[Dinarius]

That's an interesting problem if you don't mind such a new user chiming in. The need to organize is universal. I'd see if something like scientific name formats could help if you want to stay with numbers.
00.00.00 making room for 100 bases, 100 groups, 100 salt types or whatever the diminishing orders are. Going from -m 400 to -m 000400 or -m 30 to -m 000003 is my suggestion.

[inoo]

I think it seems to be nice representation suggested by Rolf
md5 or the other based hash numbering representation are make sense and easily understandable..
thanks.

[zokszo]

Hello All,

Just signed up so I can second this idea.
Easy, simple and descriptive, you can't go wrong and it's all there.

maybe use a string instead of an integer. use standard names for the well-known algorithms (raw-md5, md5crypt, etc), and come up with a logical naming convention for the lesser-known algorithms (like maybe md5-salt-pass, md5-pass-salt, etc?)

[2die4]

Using the actual name of the hash algorithm sounds pretty good to me too.