I don't know if I'd go as far as classify that as a security vulnerability as such... but it's an interesting approach. Don't quite have the cycles to fully think it through, but I think it merits some investigation/discussion.
I haven't used WebAssign since 2010. Apparently it hasn't been worked on since then, either. When it was still up, it asked me to upgrade Flash.
user315433
9:19 PM
On the subject of web apps: have you noticed how, whenever a range of dates is selected, Google Sheets helpfully calculates their sum in bottom right corner?
Hey everyone! We’re just about ready to start up today’s Town Hall Chat. I’m going to put this room in timeout, go over the plan and the guidelines, and then give the floor to our hosts. Here we go!
This room was placed in timeout for 3 minutes; Town Hall Chat. Today’s topic: More productive triggers for the reopen queue
Like last time, we’re keeping the format of this chat pretty informal, in the hope that we’ll be able to keep setup easy and low-pressure so we can keep having these chats more regularly.
We’ll start by letting @Shog9 and @JonEricson opine a little bit. Then, we’ll open the floor to y’all to react and respond.
After that, we’ll open the conversation up to whatever topic you folks want. If we want to keep going on that topic, that’s cool; if there are other subjects you’d like to hear from us about, that’s cool too. Second half of the chat will be the time to bring that other stuff up if you have it.
As with last time, please keep the meta-chat to a minimum; we'll welcome feedback on the format and such on MSE after the show. We're going to do our level best to maintain a reasonable signal-to-noise ratio here, so if you're going to type words and hit enter, please keep them on topic.
Attention! The numbers I posted in the first revision of this answer were very, very wrong. Please re-read!
Here's how reopen review outcomes break down by creation triggers over the past 14 days:
CreationReason ReopenTasks % Reopen % Leave Closed % Edit % Task Invalidated
-----------...
I’ve never done this before - break out the outcomes by input source - because the logic for this is, uh, really complicated and buggy, so figuring out where things came from is error-prone and slow. And… I was kinda surprised by just how bad 3rd-party edits actually fared.
That got me wondering… The 3rd-party edit trigger serves two purposes: giving folks who can’t vote to reopen a path to get something reviewed, and giving questions in topics that don’t have a lot of reopen voters more exposure to people who can help them.
Clearly, neither use is particularly effective though. What other options might be more effective? And no one say "flags"!
It is really surprising to me that editing is so ineffective at getting questions reopened. Maybe better guidance would help? Or maybe people just aren’t good at fixing other people’s questions? Or maybe third-party editors tend to edit before the question gets closed.
Mod note! OK, folks - we’re past the 15 minute mark, so now you can feel free to bring up any other topics you have in mind, or continue on this topic if you like.
I think we need to do something about top voted bad answers. Culturally we need to make it ok to put up banners on deprecated stuff, and outright delete (after some debate) answers that were never right.
@Shog9 So it may have exacerbated the issue, but is not a primary factor? Or has this been an issue the whole time? Would it make sense to not have 3rd party edits kick it into review queue for 24 hours?
This answer is completely wrong. It saddens me that it has been so upvoted. The example will error (don't believe me? instantiate a Square()). Even if the example's error is fixed it will create another error when it exceeds the recursion limits or the process runs out of memory. If you want to understand why, I explain this in my answer. Again - this answer is completely wrong. — Aaron Hall ♦Jun 12 at 16:31
The answerer has had ample opportunity to fix it, but never has.
@Shog9 by accident. Sometimes people think something works so they upvote, but it doesn't work, and very sometimes there's a lot of people upvoting such an answer.
@JonEricson Might need to take a close look at a bunch of cases to find out why there's such a difference with other sites, but that sounds like an awful lot of work.
@Shog9 we downvote at a rate of 1:10 upvotes. 11 downvotes is actually pretty controversial: people sacrifice their own reputation to downmod the answer.
I want to create a process to delete highly voted never-right answers:
1 Identify (highly voted, outsized (or numerous anonymous) downvoted answers that have allegedly never been right) 2 Escalate (on Meta, make the case in a Q) 3 Debate (with other gold badge tag users) 4 Delete (a moderator reviews and deletes, citing the meta question)
That just about does it for time today. Thanks everyone for coming! If you have feedback on the format, please leave it on the meta post. It wasn’t a huge amount of time, but we’re trying to keep these short and lightweight so we can do them more often, since it’s important that we get this time to talk to each other directly. Next one in two weeks - watch meta for details. Thanks for town-hall-chatting, y’all!
And if you do it too much, you also incite the inverse, where there will be increasing amounts of meta discussions trying to figure out how to undelete highly upvoted posts.
@JonEricson Here's the comment that you read as "he disagrees":
@AaronHall It saddens me that you took the time to run my example and print out its error, but didn't read the discussion in my post where I clearly explain that it produces an error and that the alternative structure results in infinite recursion. — AnjoManJun 14 at 13:10
Attention! The numbers I posted in the first revision of this answer were very, very wrong. Please re-read!
Here's how reopen review outcomes break down by creation triggers over the past 14 days:
CreationReason ReopenTasks % Reopen % Leave Closed % Edit % Task Invalidated
-----------...
