Posted on behalf of jmac.
The idea to test automating review of answers that fail to meet site-specific guidelines has been proposed by SE Community manager here:
let's test this first, see where it works and where it falls apart, and then implement the system that emerges. Pick a site you'r...
If anyone has experience with the data explorer, it would be interesting to get some results on the post notice experiment to see how much of an impact it has and whether the amount of input put into this machine is worth the output.
> Some sites suffer the double whammy of (a) being more subjective (answers are hard to verify) and (b) covering topics where everybody thinks he's an expert. Most of us work, so we're obviously qualified to answer any question on The Workplace. Any parent is obviously qualified to answer questions on Parenting. Anybody with some religious background is obviously qualified to answer questions about the bible. Or so the theory goes -- but it's wrong.
@jmort253 thanks. We actually don't have the worst case of this on the network (IMO); that'd be hermeneutics. But it's a big problem for us and for others, and Shog -- who mostly pays attention to SO et al -- needs some help in seeing it. I don't appear to have convinced him, but others can help too.
@jmort253 agreed. And we need baseline data before we run the experiment. So the question we need to answer before we can get started is: what exactly do we want to measure?
Good question. I see two different but related issues. (1) Getting more regular users involved in answer moderation independent of the up/down voting, and (2) allowing users with lower quality posts more time to edit and improve before they're ultimately deleted, which involves creating another feedback loop to hold them accountable.
#2 is where I think Shog has raised the most questions.... if adding post notices to things doesn't make all that much difference, then we should scrap that idea completely.
But #1, I feel like separating answer moderation from answer voting is key on sites like this.
Unless we raise the upvote threshold to 1000 lol....
So one set of data revolves around what happens after the post notice goes up: does the author edit? Does somebody else edit? Does the post receive no edits and get deleted? Receive no edits and not get deleted? And how long does all this take?
Another set of data revolves around community participation: who (and how many) votes to add post notices, who helps fix posts? (Votes are secret so we'll never get downvote data.)
2
I don't know if flagger identity is available in SEDE; I suspect not, since it's not public on the site.
@jmort253 agreed -- if the notices don't result in enough posts getting fixed, then we should bag that and go straight to deletion -- but recognize that we'll need mod-deletions for a while yet even though it's supposed to come from the community (all, what, 4 of them right now?).
@MonicaCellio - Today's Thursday, and there are only 2 editors for the week. (Although I think this doesn't show editors below a certain number... not exactly sure how this page works...)
Everyone here has been incredibly patient, but I can tell you that any mods that get elected would be in danger of being burned out. This has been an intense few weeks.... :)
@jmort253 the biggest challenge there is coming back to it. We delete bad posts now because if we decide to wait we may never find them again. That's not ideal; better to give people a chance to fix. But without auto-delete-if-not-fixed, or an alarm, or something like that, it's not scalable. For the experiment we can do something like maintain a date-ordered list on meta, but that's a workaround.
@jmort253 I can imagine. Just the subset that leaks through to the rest of us has been kind of intense. (Though as a fellow mod I also see non-public stuff in TL.)
I was just looking at the help for the DE earlier today, actually. They don't have "teach SQL" as a goal, of course; the assumption is that you can do that elsewhere. So there's the schema (if you know what to do with it), and some help for some magic columns and variables, but if you don't know SQL you pretty much need to go look at existing queries. Which is fine, but oh so many -- that's where I blocked, for now.
Exactly... the ANSWERS table may have a "foreign key" containing that UserId on posts I've answered, so there might be a userId column with "98" for all my answers
so
SELECT * FROM ANSWERS inner join ANSWERS.userId on USERS.userId WHERE userId=98
that might give you all my answers....
I don't know the schema though, so that's pseudocode :)
For simplicity, assume all relevant fields are NOT NULL.
You can do:
SELECT
table1.this, table2.that, table2.somethingelse
FROM
table1, table2
WHERE
table1.foreignkey = table2.primarykey
AND (some other conditions)
Or else:
SELECT
table1.this, table2.that, table2.somethin...
This discussion comes at a good time, actually -- it gives me some specific goals (queries we want) to experiment on, and it would be professionally helpful for me to not be a complete SQL noob. (I've just applied for a tech-writing job at a company in the DB space. I can learn on the job, just as I've done for every new technology that's been needed, but having a bit of a head start seems useful.)
@jmort253 I've heard of both of those (of course), but I don't know much about the differences.
NoSQL is where objects are mapped directly to a datastore. Querying is pretty much limited to a smaller subset, like looking for keys only... I'm not convinced yet that it's better.... but that may just be because I'm still learning it all :)
@MonicaCellio Google App Engine uses NoSQL methodologies, and I suspect because it's harder for developers to write really crazy SQL JOINS that consume all the resources. By boxing developers in, they've effectively made their platform scale better.
Here is an example of an answer that was flagged for the post notice, but it was actually edited and improved significantly, so it wouldn't be counted in any of our queries unless the DE can see flags....
Often people in senior positions feel that they need a certain professional distance. It's likely not personal, they just don't want to start getting pally with everyone.
Edit added: e.g. they want to be able to take business decisions without friendship getting in the way of those decisions. By...
And the first revision is definitely something I would have added a post notice to, and it's subsequent form does provide a satisfactory explanation and us up vote worthy IMHO.
2
I will say that it seems more posts are not improved than those that are improved....
For simplicity, assume all relevant fields are NOT NULL.
You can do:
SELECT
table1.this, table2.that, table2.somethingelse
FROM
table1, table2
WHERE
table1.foreignkey = table2.primarykey
AND (some other conditions)
Or else:
SELECT
table1.this, table2.that, table2.somethin...
The Water Cooler
The Water Cooler