Hmm, @prologic@twtxt.net / @lyse@lyse.isobeef.org: Should we remove the section “Traditional Human-Readable Topics” from the spec? Or mark is as deprecated? I haven’t seen this being used in the wild for years. 🤔
@movq@www.uninformativ.de I think we should because:
- it doesn’t scale
- it doesn’t make a terrible lot of sense anymore
It started out to be useful, but generally is probably considered “small scale” and “legacy” IMO.
@prologic@twtxt.net if it to be removed, then yarn should not hide the (tags)
in these posts as well as collapsing them into treads as is the case now:
@movq@www.uninformativ.de I need to think about it a bit. But I tend to deprecation rather than removal.
@movq@www.uninformativ.de @prologic@twtxt.net I just reread the spec and it seems to be even a bit outdated regarding machine-parsable conversation grouping. We long dropped the need to specify a whole hash tag with URL (#<hash url>
), the simplified version without the URL (#hash
) is enough.
The hash tag extension specification is kind of missing the same. However, I’m not sure if that short form is considered supported in general (as opposed to be a special case for subjects only) by the majority of the twtxt/yarn community.
Now the question arises, in order to keep things simple, should we even only allow the simplified twt hash tag for subjects and forbid the long version? This would also save quite a bit of space. The URL is probably not shown anyways in most clients. And if so, clients might rewrite URLs to their own instances. On the other hand, there’s technically nothing wrong with the long version in current parser implementations. And deprecating stuff without very good reason isn’t cool.
@lyse@lyse.isobeef.org I definitely think we should drop the longer form that involved a full url.
Also given that there’s a growing number of clients that support this, I’d consider it pretty accepted nowadays.
Agree. we should parse the form but not output to feeds.