XSF Discussion - 2023-04-19

https://blog.whatsapp.com/an-open-letter

Cool to see that they listed Matthews signature first :D

Open Protocols!!! \o/

Why isn't the XSF on there?

My reaction ^

Probably because this is the first any of us are hearing about it

The ietf isn't there too...

Yeah, I worded that needlessly combative. What I do find annoying is that we seemingly aren't in a place where we are a natural invitee to the draft of something like this.

Menel; the IETF would make a great addition to that list, but I believe that the content of this letter is even more specifically applicable to the XSF

perhaps coz the XSF is not a technology company ?

Matthew: do you have any insight?

> <@_xmpp_vanitasvitae=2fxsf=40muc.xmpp.org:matrix.org> Cool to see that they listed Matthews signature first :D alphabetic order. (plus we wrote the letter)

> <@_xmpp_Guus=2fxsf=40muc.xmpp.org:matrix.org> Why isn't the XSF on there? Why isn't the Matrix.org Foundation on there?

xd

That's for them to decide :)

it was put together by the Element policy team, and I guess they just had their corporate hats on - plus reckon that the government is going to listen more to "these jobs and companies will be impacted if this goes ahead" than "these open source projects will be impacted"

but given there isn't an Element-equivalent for XSF, it'd have made sense to reach out - sorry; i just didn't think of it.

(i also forgot to reach out to Status.im, who are an investor in Element, which is a bit awks too)

Understood, thanks.

i guess another aspect is that both the Matrix Foundation & XSF *won't* be directly impacted, given we'll just go and build our own FOSS matrix/xmpp clients and ignore the whole nightmare

In any case, nice job on the letter. The prospect of e2e encryption with backdoors is scary. Similar discussions are being had on this side of the channel.

it's the companies who will get punched in the face by OFCOM to add GCHQ-accredited binary scanning blobs into their apps :|

(and the reason why i'm leading the charge on this is that Element is the only UK-based one, afaik) ✏

and so will be doubly screwed by this :|

Still - standardization organizations are a knowledge holder here. Their voice matters more than what you appear to give them credit for with the above.

yup. plus it would avoid accusations of "greedy CEOs of capitalist companies try to protect their business model"

oh well

bit late now.

we live and learn.

can't update ?

let's cooperate more on stuff like this in the future.

the blog post has been and gone - it all happened yesterday, and got national press

given the biggest pushback was "don't trust greedy corps", you're right that we should have aimed for more sigs - will pass it onto the guy organising it.

(at least the fact we also didn't get the M.org Foundation to sign shows we weren't being anti-XMPP though :P)

> Matthew: > 2023-04-19 11:48 (GMT+02:00) > i guess another aspect is that both the Matrix Foundation & XSF *won't* be directly impacted, given we'll just go and build our own FOSS matrix/xmpp clients and ignore the whole nightmare Thanks anyway for the letter. I wonder if it does not affect open protocol implementations.

> Matthew: > 2023-04-19 11:54 (GMT+02:00) > given the biggest pushback was "don't trust greedy corps", you're right that we should have aimed for more sigs - will pass it onto the guy organising it. Happy to offer my contact for future collaborations

Other feel free too (@board)

Breaking News: 2023-04-xx - Thousands of signatures from all over the place agaisnt the UK's e2ee policy and numbers are still rising

if we can update the kernel, i'm sure we can update the news

A question on DNS and SRV. I'm trying to help someone that wants to migrate an XMPP service to a new server that will have a new hostname. The XMPP domain name is equal to the hostname of the current machine. To prevent data migration, I'd rather not change the XMPP domain name. Can I get away with creating a set of DNS SRV records that look like `_xmpp-client._tcp.oldserver.example.org. 86400 IN SRV 5 0 5222 newserver.example.org.` ? My concern here is that I'm inconsistently using the same value (`oldserver.example.org`) as a domain name and a server name (in the `example.org` domain).

host and service name do not need to be the same

but, how many clients check SRV records? and how many assume domain = host name?

(not all languages support checking of DNS SRV records)

100% of the clients that these people use, check SRV records (and to be honest, I assume that most in the wild also do?)

(older) .NET clients would not, unless they used a custom library, to check DNS SRV records. But as long as that is not a problem it “should”™ work. ✏

tx

(also, depends on the certificate also, btw)

we had this discussion earlier, about differences in checking domain names in certificates, depending on binding

true that, but that's under their own control. A wildcard cert should suffice there.

(and/or they can generate a new one)

it's an internal network

Yep should work, and if any clients people use don't support SRV records (or hostmeta) we should bug devs or patch