Zulip currently allows messages to be starred, and starred messages to be narrowed to. It would be much more useful to have actual tags — containing words that could be read, sorted, searched for, edited.
Use case 1: I can't remember why I marked a particular message, so finding it again via the list of starred messages, doesn't do anything for me. A tag, since it is a note to myself, tells me why I marked it.
Use case 2: Sometimes I may remember that some question has been bothering me for some time, but not that I have a starred message about it somewhere in my long Zulip history. Viewing my tags may jog my memory.
I would suggest making a tag no less than 30 characters, so that it might contain as much as several words (at least three or four five-letter words) and a date (eight characters).
This is a good idea, and similar ideas have been suggested before.
One way to solve the tagging problem is to have somebody implement a bot that watches for tags, so that the bot can show you all the tags you have created when you ask it for them. The advantage of a bot approach is that there's less friction in terms of getting it deployed, whereas changing the Zulip product itself raises the stakes.
Currently on Zulip you can do tagging by convention and say things like "tag:todo-list", and this allows you to search for your todo-list, but of course the problem is remembering the tag.
We recently added a bot to our repo that can basically allow you to tag messages using file folder syntax. It's not specifically designed for tagging, but it's in the same vein. The code is here:
https://github.com/zulip/zulip/blob/master/contrib_bots/lib/virtual_fs.py#L79
Is there a list of bots somewhere?
sent afk
We don't really maintain many bots yet, but I'm hoping to improve that shortly, and once we some momentum on that front, I'll add some documentation. Until then you can see the three bots here:
what is the status on this issue -- have we decided on on how go about this?
I still think we should do a bot for this. I'll mention it on the integration stream.
Most helpful comment
I still think we should do a bot for this. I'll mention it on the integration stream.