Re: des-die-die-die and RC4
"Henry B. Hotz" <[email protected]>
| Newsgroups | gmane.ietf.krb-wg |
|---|---|
| Message-ID | <[email protected]> |
On Feb 17, 2012, at 11:39 AM, Jeffrey Altman wrote: > On 2/17/2012 12:33 PM, Henry B. Hotz wrote: >> If it weren't for the need (for a little longer) to support a compile-time option for 1-DES compatibility with AFS, I'd be fine with an even stronger MUST NOT implement, provide, or on any other way allow kind of provision. If AFS doesn't get workable rxgk support soon that alone may kill it. > > Whether SHOULD NOT or MUST NOT is used should not be determined by the > needs of any one application that is currently deployed or might be > deployed in the future. Other people have pointed out off-list that there are other things. Also I must admit we have a lot of Solaris 9 machines around. > This specification only impacts software that is developed and deployed > after the specification is published. Organizations are always going to > have to provide back level compatibility for existing deployments. My concern is that 1) we make the deprecation as strong as possible, but 2) we not label a new implementation "not Kerberos" just because it has some optional backward-compatibility feature needed e.g. by current AFS. All of the discussed wordings are acceptable compromises IMO, however much I might want something stronger. > Jeffrey Altman ------------------------------------------------------ The opinions expressed in this message are mine, not those of Caltech, JPL, NASA, or the US Government. [email protected], or [email protected] _______________________________________________ ietf-krb-wg mailing list [email protected] https://lists.anl.gov/mailman/listinfo/ietf-krb-wg