1 |
W dniu śro, 04.07.2018 o godzinie 10∶01 +0200, użytkownik Kristian |
2 |
Fiskerstrand napisał: |
3 |
> On 07/04/2018 09:54 AM, Michał Górny wrote: |
4 |
> > > We also keep gnupg 1.4 in tree that does not, and will not, support ecc. |
5 |
> > |
6 |
> > Well, we have developers using ECC (Curve 25519, to be specific). |
7 |
> > I don't really know enough about this to judge but we either need to |
8 |
> > allow at least this, or convince those devs to change to RSA. |
9 |
> |
10 |
> incidentally curve25519 is the one I'm thinking of that isn't |
11 |
> standardized, although it is part of current draft version of rfc4880bis |
12 |
> (but WG is stalled so no update expected any time soon there). |
13 |
> NIST/brainpool are included in RFC6637, but we wouldn't want to accept |
14 |
> them for various reasons. |
15 |
> |
16 |
> There are good reasons these are not provided in the regular interface |
17 |
> of gnupg, but requires --expert |
18 |
> |
19 |
|
20 |
To be honest, I have mixed feelings here. |
21 |
|
22 |
While I agree interoperability is a problem in general, I'm not sure if |
23 |
it's really a problem this large. I agree that we shouldn't recommend |
24 |
ECC but should we ban it entirely? |
25 |
|
26 |
Things to note: |
27 |
|
28 |
1. I suppose the ECC/cv25519 packets won't change in incompatible manner |
29 |
at this point. |
30 |
|
31 |
2. Hardware incompatibility issues are not really relevant to us but to |
32 |
the person using the key. |
33 |
|
34 |
3. Developer keys are mostly for internal use, while the majority of |
35 |
users verify only the infra signatures, so I don't think we have to be |
36 |
that concerned about interoperability of the algos, provided that it |
37 |
works for infra purposes. |
38 |
|
39 |
-- |
40 |
Best regards, |
41 |
Michał Górny |