Password encryption

Since the topic of crypt() likes to surface from time to time, I am wondering if breaking dependence on it would be a good thing. Using something portable so that it will compile and produce the same results on any system, no matter the OS.

In my own code, I'm using an MD5 routine that Khalan of DoT pointed out to me. It seems to be working great. I have thus far had no problems with it. It's also currently used as the encryption method for IMC2 connections in the AntiFreeze code which is part of the FUSS packages.

I suppose what I'm getting at mostly is this - would this be something that would be appropriate to install in the FUSS packages? I realize that it would make pfiles from crypt() using muds incompatible with pfiles from MD5 muds. But how often does a mud copy pfiles from another mud anyway?

The only real problem I can see is an existing mud wanting to use the new system, like us. There would be a need to maintain the crypt() version for awhile to allow pfiles to update with time. Once the pfiles have been updated then the crypt() support can be removed. This would require the use of file versioning, which isn't all that hard.

So does anyone have any thoughts about this? Good idea? Bad idea?

I personally think its a great idea. As I understand it, most ( or all? ) versions of *nix, including cygwin I think, use the md5 routine anyways. SMAUG, not normal and FUSS use file version anyways for the pfiles, as well as the derivitives, like SWR, so its really not a hard matter to change over.

I'd say anything is better than crypt. Even no encryption is better, since crypt causes some problems when entering passwords. (The whole password does not need to be entered to log in)

But yeah, its a great idea. I'm the only one who has shell access, or ever will, on my MUD, so I may just get rid of encryption.