#215 - Newlines without Unicode - yarn - Mills – We discussed the idea of changing the way we handle Newlines in Twts once upon a time, but nobody adopted it, we didn’t amend the spec and really we just kept the existing spec as-is using the \2028 and simpler search/replace in place. Did we want to revisit this again in a potential Twtxt v2 spec rewrite, or are we happy with how this works? 🤔 cc @movq@www.uninformativ.de @lyse@lyse.isobeef.org @tkanos@twtxt.net (please add other client authors to this thread) #Twtxt #Spec #Multiline
@prologic@twtxt.net I don’t think the solution proposed there is a good one, and here are my reasons.
The specification says quite clearly, “The file must be encoded with UTF-8”. If an old piece of software can’t handle UTF-8, it can’t produce a valid twtxt feed at all.
I believe the intention behind this solution is to make it render in an acceptable fashion in clients that don’t support the convention, but I think it’s the opposite in reality. Separating posts like that could make it very frustrating to read in a feed. I would much rather have nothing or a replacement character separating logical lines.
I think it interferes quite heavily with human readability for the same reason. When reading a twtxt feed, it’s helpful to know that each line with a timestamp represents one post.
It also says “A specific ordering of the statuses is not mandatory,” implying that the order of the lines in the file is irrelevant. If newlines are separate posts with the same timestamp, the original line order becomes very relevant. I can see how a client (that doesn’t support this newline syntax) might display posts with the same timestamp in the wrong order because of this.
@mckinley@twtxt.net Agree on all points đź‘Ś