I think I agree with you, I'm not sure what purpose flagging as low quality has. I am working using the following process:
- If the user didn't do the homework, how bad is it? New user? If I can, I'll edit, comment, and wait for an improvement. However, if the user has say 100+ rep or a history of asking such questions, I use my close vote. They should really know better.
- If the user asks an overly broad question, again, is the user new? Is English not perhaps their native language? If so, perhaps a comment. Perhaps a vote to close.
- If the user didn't make the question relevant to others, vote to close (too localised). This is almost always obvious. Questions like "what is wrong with my code pls fix" or "I'm doing the facebook hacker cup how do I solve this
I never vote based on not liking something. Thankfully there's those nice favourite tags so I don't have to see questions about iPhones and PHP scripts or JavaServer Faces. So I'll take not liking something as being fundamentally technically wrong, such as:
I'm trying to dereference a null pointer in a kernel module, please help.
In this case, I have downvotes.
So what purpose does Flag|Low Quality actually serve? I agree with Shog9 that flag weights should hopefully filter out noise flags as the system learns. I'm flagging carefully so that my flag weight is high. As I rarely flag anything, when I do it's because I've seen something quite seriously wrong and I want that to have some weight.
Now I'm actually going to answer my own question via an educated guess. You know those poor questions that pop up along the lines of "can ne1 help???? pls tell me how parse regex with html" etc? (yes, that's the wrong way around on purpose...) Well what usually happens is lots of people see it and think expletive! downvote-downvote-downvote. I've never understood that - just wait till enough 3k people come along. Anyway, I suspect that this low quality flag is designed to allow users with less rep a way of saying to 10k users "look, problem here".
I wonder if it isn't already in the pipeline already that rather than have low-quality go directly to a moderator (since this is unnecessary - diamond mods are for spam, stopping open/close wars, revenge downvoting. They aren't needed to do self regulation us common folk can do) that a flag of low quality direct a post to the
In that case objectively measuring flag weight gets harder, but perhaps could be measured by outcome i.e. if an edit happens and the question is answered or if the question is closed, then the flag weight increases. However, if the question is upvoted to a certain threshold (say 2 upvotes overall after 24(12?) hours) then the flag weight decreases because this clearly wan't low quality at all.