From my point of view, the currently recommended taxonomy is the important thing we should focus on. This is what makes new questions better, easier to find.
SE contains lots of “crap” from today’s point of view. Some questions were popular and are provided with historical lock, some have been always considered crap and by accident never met criteria for being purged from the system automatically. Some are really useful, but would not meet today’s quality standards; those can be good candidates for editing.
What about lazy burnination? You imply that burnination must be finished ASAP after the tag is decided to be burninated. Why? Why don’t we just let those old questions be and improve them only when we meet them later? Only then burninated tags should be removed/replaced and other things improved as usual edit workflow requires. Don’t these questions fall in the “useful but not meeting today’s quality standards” category?
The only problem I see is that these tags would be still offered by auto-complete when editing a (probably new) question. Orientation in the jungle of deprecated tags would become gradually harder and harder. This would need support from the system to be solved – we can issue a feature-request for adding “burninated” attribute to a tag. Tag with this attribute would not show in auto-complete and would have some unified identification in tag wiki and its excerpt. This way they would be easy to identify. Choosing different look is IMO too much as it would be confusing.
If lazy burnination is bad and fundamentally hurts the system somehow (I don’t see how, though; explain please!), why don’t we have review queue for questions with tags being burninated? The question would be shown with the tag being automatically removed greyed-out and the options could be No other action needed, Edit and Skip. This would ease the hassle with keeping track of burninate-requests. Or is there already a better way of keeping track of which tags are being burninated? I will gladly take part in burnination process if somebody tells me what work is to be done. Yeah, discussion on Meta is important part of the process, but I mean the real, editing work.
I’m against batch stripping of burninated tags from questions.
- Lots of trivial revision history records would be generated. Noise.
- Questions could lose important information, context. There may be unexpected uses of the tag.
Human exception handlers should be triggered.