[OTR-dev] OTR meeting notes - Tails hackfest in Paris 2014

David Goulet dgoulet at ev0ke.net
Sun Jul 6 09:47:29 EDT 2014


Hi everyone!

Yesterday at the Tails hackfest in Paris, there is some OTR contributors
so we gathered together and talked about what improvement could done to
OTR and the community.

Here are the meeting notes which contains some ideas that we would like
to see happen in the short term thus needing more eyes on this
especially the part about "Modern Crypto".

Thanks to the attendees that helped a lot!

Cheers!
David
-------------- next part --------------
OTR meeting notes
-------------------

Date: 06/07/2014
Location: Tails Hackfest, Paris

Attendees:
	- jvoisin
	- infinity0
	- dgoulet
	- drwhax
	- vmon

1) Modern cryptography

Migrating current OTR protocol to use modern cryptography. We would like to
replace the DSA signature to "ed25519". The DH exchange should be replaced by
"curve25519". libgcrypt supports ed25519 since version 1.6 (package
libgcrypt20). The curve25519 is unclear if it's merged or a work in progress.

The new key(s) should be derived from the old one so users can keep their
current fingerprint.

For that, we discussed the need to cross sign keys for the transition.

Also, should chacha20 and/or poly1305 should be considered as well? No one had
a strong opinion on that.

A proposal of these changes should be written first before any code starts and
Acked-by maintainers/developers/contributors off the community.

2) Tests suite

You can find here a branch of the test suite started by dgoulet which contains
some basic unit tests now integrated with libtap.

	git://git.otr.im/libotr-next.git
	(branch: test-suites)

We agree that an "OTR fuzzer" would be great also to basically hunt bug and
also be able to add this to a continious integration system.

There is a bunch of open bugs/features on
https://bugs.otr.im/projects/libotr/issues that we need to tackle but we all
agree that we should *first* make the test suite with a descent code coverage
so we can actually confirm that what we are fixing/implementing is not breaking
anything.

Once we have that, there is some kniffing to do especially on some part of the
internal ABI (for instance, second comments of this
https://bugs.otr.im/issues/23). Memory allocation used without checks, stuff
like that. Mostly this kniffing would be simply to improve the code to make it
more easily maintainable and robust.

3) OTR.im

We have a twitter account now to tweet about some stuff that's going on in the
OTR community I guess and news/update... So send anything you think might be
worth tweeting :).

	https://twitter.com/otr_im

Also, we discussed having more action on the blog (https://otr.im/blog)
especially maybe putting a "Call to action" for testing.

A new git repository containing the specifications would be a good idea to
create so we can have people looking at the progress of the modern crypto spec.
for instance.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 603 bytes
Desc: Digital signature
URL: <http://lists.cypherpunks.ca/pipermail/otr-dev/attachments/20140706/1a6f1f39/attachment.pgp>


More information about the OTR-dev mailing list