Improvements to Smokey's reports in chat

There is some room for improvement of Smokey’s chat messages. For instance, the quite valuable MetaSmoke link is sometimes dropped because of the length of the message. To make optimal use of the ‘real estate’, we should consider that the messages are seen both in general chatrooms by users who have hardly heard of SmokeDetector, and by hardcore spam fighters in Charcoal HQ.

Some discussion about this can be found here.

IIRC, feedback via chat depends on correctly parsing the report message, so one can’t just change the message format without changing that logic too. Also, I’ve seen some chat messages in the test part of the repository, which also need to be adjusted.

Here is a list of possible improvements. Some of them should be ‘permanent’ changes, others should only be changed if there’s not enough space.

• The first link ([SmokeDetector]) isn’t useful. In Charcoal HQ, you can find the link to GitHub in the starred Project Links message, or in the chatroom’s description. It should be dropped there; for other chatrooms, it should probably link to http://charcoal-se.org/#whats-smokey because that’s much more informative for a newcomer than the GitHub repository itself.
• We can remove the spaces, so that the first part looks like [SmokeDetector|MS] if necessary.
• Some of the reasons can be made shorter (at least, the way they’re displayed in chat), e.g. “website” -> “site”, “repeated URL at end of long post” -> “repeated URL”, if necessary. See here for more examples.
• Reasons can be grouped, e.g. “bad URL, bad keyword (x2), blacklisted/pattern-matching website (x4), repeated URL” if necessary.
• List can be shortened (e.g. Reason 1, Reason 2, Reason 3, Reason 4, Reason 5, +2 more)
• The site, usually written out in full like graphicdesign.stackexchange.com, can be shortened to graphicdesign.SE if necessary.
• Add autoflagged information (if that’s even possible - maybe something for the long term) to the messages (in Charcoal HQ only). This is also something a userscript could do.

Any thoughts?

[Undo1]

Is there any reason we couldn’t just start truncating the reason list to the first few, or at least when necessary to save space? If a post is caught by a gazillion reasons, no one really cares exactly which ones. If they do, we have metasmoke.

[ArtOfCode-]

Mostly agree. Autoflagging is truly async, so including info in the initial report is not possible without significant changes somewhere. Probably a userscript job. As for the rest, they’re all worth considering, though we should start with whatever saves us the most space and loses us the least information.