It's definitely no longer plain text, if you interpret it as a code snippet/preformatted text.

But as hodlbod pointed out that line has kind of already been crossed with link previews, event and npub references, so *shrug*.

Might as well go full markdown imho. Would be easy to have a content type flag to say "this was MEANT to be markdown". Since markdown is perfectly readable as plain-text too, I think it's reasonably backwards compatible.

Now seems like a particularly bad time to release the #nostr software I was working on this past week, it'd just add to the fAr rIgHt fire. On the other hand, no such thing as bad publicity?

If you still have your twitter account...

You don't hate "journalists" enough.

The bot problem on twitter is incredibly funny when you're having a technical discussion.

Presumably relays can set limits on number and complexity of filters and subscriptions, do any implementations do this currently? How do they handle this?

I gave the nips repo a brief look and didn't see anything, but I wasn't very thorough.

In particular, I want to subscribe to a set of pubkeys of interest, is there a practical limit to that set size? What happens if I hit that limit?

#asknostr

# Test

Also a test

==========

```

this is a

test

```

*this is also a test* plus **this is a test**

_yet another test_

[definitely a test](https://coracle.social)

> finally, this is a test as well

Whoa, backticks work on kind 1s in coracle? I thought any kind of rich text in kind 1 was heavily discouraged, or has everyone just decided to do it anyway?

Hrm, I guess it could also be

['L', '-#t'], ['l', 'sfw-photography', '-#t']

?

That avoids the problem you mentioned and generalizes to other labels and tags.

nostr:nprofile1qyd8wumn8ghj7urewfsk66ty9enxjct5dfskvtnrdakj7qguwaehxw309a5x7ervvfhkgtnrdaexzcmvv5h8gmm0d3ej7qgwwaehxw309ahx7uewd3hkctcpz4mhxue69uhhyetvv9ujuerpd46hxtnfduhsqgyhcu9ygdn2v56uz3dnx0uh865xmlwz675emfsccsxxguz6mx8rygv9tg2c did you ever consider inverse labels for NIP-32?

Something like

['L', '#t'], ['l', '-sfw-photography', '#t']

to remove an event from a hypothetical #sfw-photography hashtag

I have a couple of use cases but moderation is probably the primary one.

I'm only saying that validation depends on validating the event id AND the signature. Only doing half validation is exactly as secure as doing neither.

But regardless, I think validation is mandatory at any performance penalty.

I was going through the verification flow in my head before I wrote the OP, trying to come up with a justification, maybe DoS related, but couldn't come up with anything.

I don't think there's *any* value in validating the signature without also validating the signature is actually signing the contents, otherwise you could just steal a valid signature/event id pair from any other event and fool these partially verifying clients.

You're either trusting the relay to have verified everything, or not, partial validation doesn't seem useful at all.

That's fair I guess. nostr *does* seems oddly byte conscious in some places though (many single digit tags, integer kinds instead of namespaced string identifiers)

oh right #asknostr

Why do nostr events carry their own id? It's derived from the rest of the event, isn't it?

That's 72 wasted bytes on every event as far as I can tell, which isn't *that* much, but I don't understand why it's there at all.

What clickbait, that's not a picture of a vegan steak

I'll happily die for my children, sorry journos, can't dissuade me.

lol, I went and checked my profile immediately prior to posting that, just to be sure.

... now on this post (and on my profile) it's showing 23...

Same, my coracle is showing a WoT score of 0 for myself right now, but earlier it was 21, and before that (past week?) it was ~26.

At the very least WoT scores seem unstable lately.

It's worth considering whether someone with +3 -3 is really the same as someone with flat 0. (I tend to think not)