Naming of GnuPG

[Robert J. Hansen]

Christoph Anton Mitterer wrote:
> Well I did not ask to mark it deprecated... it's also ok to maintain it
> for some time (probably one or two years?).

You said to phase it out.  The engineering term for that is 
"deprecation".  When something is marked deprecated, that means it works 
now but there are active plans to move on to something else.

> But in the end we'll either have two different gpg's (which could lead
> to a lot of problems, even security related) or one of the two will be
> phased out.

The latter will not happen, judging from our experiences with PGP 2.6.

>> The question is not whether any OpenPGP changes from 2.0 will be 
>> backported to 1.4.  They will.
 >
> Ok,.. but to backport nearly everything would make little sense,... in
> that case we could simply add the CMS stuff to gpg 1.4.x and drop 2.x
> completely ;)

Please read what I wrote.  I did not say everything from 2.0 would be 
backported.  I said OpenPGP changes would be.

> What if ECC or V5 keys will finally come? Should they be backported?

They would be backported.  Look at how many hacks people have come up 
with to the 2.6 codebase to support new ciphers, new hash algorithms, 
new... etc.  The only question is who backports them.

Keep in mind that I'm not saying wk, David, etc., are forced to do the 
backporting.  I'm just saying that it will happen even if they wash 
their hands of it.  1.4 is not going anywhere, for two very big reasons:

	1.  Werner has said it's not going anywhere
	2.  The community won't let it go anywhere

Any proposal of "well, we should phase out GnuPG 1.4" needs to address 
both of those reasons why phasing it out is impractical.

> Uhm,.. the only problem that I could see here are possible build
> problems with 2.x (are there any?).

Tons.

Building on POSIX is pretty easy.  Building on Windows is a torment of 
the damned -- I think the way the developers do it is with a 
cross-compiler hosted on a UNIX system.  I don't even want to think 
about building it on OS/2 or VMS.