Good morning, Brisbane! First "real" day of #IETF119. https://www.ietf.org/how/meetings/119/
We start with the new "all dispatch" session: new job which does not know where it fits and has to be dispatched somewhere in the large IETF.
Good morning, Brisbane! First "real" day of #IETF119. https://www.ietf.org/how/meetings/119/
We start with the new "all dispatch" session: new job which does not know where it fits and has to be dispatched somewhere in the large IETF.
SSH v3 = SSH over HTTP
One of the reasons is that SSH is too easily blocked by firewalls (port blocking or DPI+RST). (And the new IP is HTTP, as you know.)
Let's follow the agenda https://datatracker.ietf.org/meeting/119/materials/agenda-119-alldispatch-05
First, #SSH v3
(I hear people chat and laugh during the reminder of the Code of Conduct.)
@lanodan @shaft Là, il en est à faire tourner SSH sur HTTP/3 sur QUIC sur HTTP/2 (pour les réseaux qui bloquent UDP et donc QUIC).
@bortzmeyer @lanodan @shaft Et t'a encore des profs pour venir me dire que le modèle OSI a 7 couches et seulement 7 couches…
WTF, le nombre des trucs imbriqué les uns dans les autres…
@bortzmeyer @lanodan @shaft Après le modèle OSI, il y a le modèle TCP/IP, mais finalement on va simplifier en modèle IP/HTTP.
@eragon @shaft @lanodan Comme je dis toujours en cours, le modèle en couches n'a qu'une seule utilité : découper le cours en chapitres. Mais ce modèle n'a aucun rapport avec la réalité.
Now, happy eveyballs (updating RFC 8305) https://datatracker.ietf.org/doc/draft-pauly-v6ops-happy-eyeballs-v3/ One problem of happy eyeballs is that it hides problems (on purpose, but it does not help to solve the problems) so reporting is an issue.
Other issues: IPv6-only networks (more and more common) and choices QUIC vs. TCP (not just v4 vs v6).
So, the new algorithm prefers ECH then SVCB/HTTPS.
The proposal does not seem consensual (for instance, HTTP is hop-by-hop while SSH is end-to-end). Instead using WebTransport? Or MASQUE?
Now, "Unichars", "Unicode Character Repertoire Subsets" Defining subsets of Unicode useful for some applications (for instance, I-JSON, the subset of JSON that is used at the IETF)
"Old people in the room will remember what 'control characters' are."
For instance, "Unicode scalars" are Unicode code points without the awful surrogates of UTF-16.
Human Rights Privacy through Deterministic Hashed Based Elision
https://datatracker.ietf.org/doc/draft-appelcline-hashed-elision/
Once data is sent, you can do nothing, it will be redistributed, correlated, etc.
(He has a solution but, frankly, I did not really understand it.)
"The problem existed for the last 35 years"
Wrong-Recipient URL for Handling Misdirected Emails
https://datatracker.ietf.org/doc/draft-dweekly-wrong-recipient/
Proposal: report the wrong recipient via a HTTP POST to an URL indicated in the header of email.
@lanodan No, it's called control characters but people born after the disuse of text terminals don't know what they are about.
Modern Network Unicode (moving after RFC 5198, which defines the use of Unicode in IETF protcols when they need "plain text")
https://datatracker.ietf.org/doc/draft-bormann-dispatch-modern-network-unicode/
"Plain text is no longer enough, data has structure."
The problem is that IETF already has a framework for that, PRECIS. Why not using it? (One possible reason is that nobody knows and uses PRECIS.)
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.