Highlights from another week iNatting instead of working on my household chores. 
I’m adding some longhorn beetles to my identification repertoire. One species has over a thousand observations at Needs ID, although I’m trying to go slow, in case I’m making some error that feedback might correct.
Following up on the ID-a-thon, I felt that what we were telling new identifiers, that adding a broad-level ID would help bring observations to the attention of experts, is largely an empty promise. Too many end up in those big piles (e.g. plants, angiosperms, dicots) and remain there for years. So, this week, I’ve been spending some time looking at plants with IDs in the Kingdom-Class range, and sorting randomly to see what I can pull out of the pile with some judicious use of the CV suggestions. Even if I’m only picking one observation out of every random selection of 30 (I’m looking for clear photos including flowers or fruit, usually), there’s a lot that can be refined. I admit I’m being bolder than usual, identifying taxa I’ve never seen before, but I’m keeping a close watch on my notifications and will withdraw an ID that’s gone to far. (That happened once so far this week, with a helpful comment that explained my error.)
I’ll note that there’s another category of observation stuck in these huge piles. Someone identifies an observation as species A, then someone else adds a conflicting ID for species B from a different order. Neither identifier tags anyone else. Now, the CID is bumped up to the common taxon (e.g. Dicot) and there the observation sits for years. Unless they add an ident_taxon_id= URL filter to search for them, experts for species A and species B are left unaware of these observations of potential interest. It takes at least one more ID to push these into the pre-maverick project, from which more help might finally break the deadlock.
There’s one other thing I stumbled across this week. I’ve favored more use the DQA to get observations out of the Needs ID pile, but I’ve been troubled that we can’t search for observations with this DQA field set. Well, @jeanphilippeb apparently found some way to do that, since he created a project for The Community Taxon is as good as it can be three years ago! Some of the entries in that project are intriguing and warrant some further examination. 2,186 observations with an CID rank of Kingdom to Class are currently Needs ID. If any IDs really can be improved, these are surely some of the likeliest ones to check. And then there are 351 observations entered as Unknown! This suggests two possibilities to me: either observers are inadvertently clicking this entry or they have a severe misunderstanding of what that DQA means. They’re probably interpreting it as “I personally don’t have any idea what this is (and don’t even realize that I can enter something broad like ‘plant’) so, yes, Unknown is as good as can it can be.” Ugh. These entries are also something I intend to check. And thanks to reading the DQA votes disappearing thread, I know that adding any better ID will eliminate this faulty DQA vote.