jdev - 2024-08-18

moparisthebest: is this better then just quic for xmpp. Or only meant as disguise to look like web traffic?

Menel: it's basically quic with a header, but that might be an advantage

Is it possible to make selective items private (XEP-0223), even when node access model (XEP-0060) is open? This is an error which I have received after attempting to make an item private. ``` <iq to="blasta@jabber.i2p/blasta" id="c2e3e5568995416d92050f1c37b48d7e" type="error"><error type="cancel"><conflict xmlns="urn:ietf:params:xml:ns:xmpp-stanzas" /><text xmlns="urn:ietf:params:xml:ns:xmpp-stanzas">Field does not match: access_model</text><precondition-not-met xmlns="http://jabber.org/protocol/pubsub#errors" /></error></iq> ``` See https://codeberg.org/poezio/slixmpp/issues/3551

No

lovetox. Thank you!

Perhaps I should use two nodes. One for public access and one for private access.

> moparisthebest: is this better then just quic for xmpp. Or only meant as disguise to look like web traffic? The same modulo a defined hadshake ↺

> Perhaps I should use two nodes. One for public access and one for private access. Unless you want to do some server dev that's how it'll be yes ↺

> Unless you want to do some server dev that's how it'll be yes Thank you, singpolyma. ↺

And also protocol dev

Thank you, Zash.

Is it worthy to work on such specification or would it be of a too high risk? ✏

Is it worthy to work on such PubSub specification, which would allow mixing private and public items, or would it be of a too high risk? ✏

> Schimon, software should not use xmpp:* namespaces without a XEP MattJ, I intend to make use of the following node names. ``` # Bibliography node_id = "urn:xmpp:bibliography:0" node_title = "Blasta" node_subtitle = "Bibliography" # Private bibliography node_id_private = "urn:xmpp:bibliography:private:0" node_title = "Blasta (Private)" node_subtitle = "Private bibliography" # Reading list node_id_read = "urn:xmpp:bibliography:read:0" node_title_read = "Blasta (Read)" node_subtitle_read = "Reading list" ``` Blasta is still in testing phases and was not yet released, albeit it is designed to serve multiple people from a single instance. ✏ ↺

Schimon, you will now be reported to the protocol police for using a namespace not owned by you!

Upon your refusal, I will use `tld.blasta`, `tld.blasta.private` and `tld.blasta.read`. (`tld` was not determined yet).

Ahahaha!

I don't backwards-dns namespaces like that are very common in XML outside of OMEMO :/

Do you mean to `eu.siacs.conversations.axolotl`?

That, yes

I want to avoid this, because I am offering a new system, which is based on the Atom Over XMPP (Atompub/Atomsub) wgich is utilized by Libervia and Movim, and is also intended to integrate with Libervia, Movim and other software, so having the brand "blasta" or a specific dommain would be weird if that system would be accepted by the XSF. ✏

I mean, the form `tld.something.foo` is weird and not an URI I think?

Yes. I agree with your statement.

While it would work, it's just an opaque string to everything. But https://www.w3.org/TR/xml-names/ seems to say you should use URIs

So I want a URI, yet I intend to coordinate with you.

Zash. I am not familiar with XML Names. May you offer a URI?

https://www.w3.org/TR/xml-names/ is an URI, but also an URL

urn:xmpp:example is an URI, but also an URN

xmpp:example.com is also an URI

Yeah an https or xmpp URI is the obvious choice imho

https://en.wikipedia.org/wiki/Uniform_Resource_Identifier

I've done some Prosody things with xmlns="xmpp:prosody.im/something/or/other" :)

You may notice lots of http://jabber.org/protocol/* namespaces from the XMPP middle ages. And earlier jabber:x:various which are weird and wrong too

Then I will use `xmpp:bibliography`, `xmpp:bibliography:private`, and `xmpp:bibliography:read`.

What do you think?

Or these: `xmpp:bibliography`, `xmpp:bibliography/private`, and `xmpp:bibliography/read`.

note how xmpp: is followed by a valid JID

prosody.im is a valid (host-only) JID

so unless you own the .bibliography TLD, that'd be weird too

Zash. I understand. Thank you for the clarification. I think I will ask for an i2p address. ✏

So you may understand the suggestion of just generating a random UUIDv4 and claiming a namespace like urn:uuid:9ff2548d-1384-41e4-9c4e-d563e78c5f20:whateveryouwant

Can you do suffixes on uuid urn like that?

Pretty sure you can just continue the :-separated path all you want after the uuid component

and add #hashes and whatever

> and UUIDs are opaque, which makes the protocol a lot less accessible and debuggable > and UUIDs are opaque, which makes the protocol a lot less accessible and debuggable ↺

I suppose I will use a domain.

What do you think of this statement? > You cannot use urn:blasta - please use something else. > All URN namespaces need to be registered with the IETF / IANA. > Please see RFC 8141,. > But urn:xmpp:journal is fine because the XSF manages the urn:xmpp space.

There's no really to use a urn one I er an https one or xmpp one

I am not intending to actively continue this discussion. This is too much; and I want to publish Blasta already. I will ask WPN to use their domain, in the fashion that Zash has suggested. ✏

Naming things is one of the hard parts of computer science. We're just doing our part! :)

Yes. So I notice!