The Whiteboard

Matt D

4:16 AM

I am astounded at how difficult it is to kick off an async task from a webserver using php

gnat

7:28 AM

31

Q: Why are comments now disabled on “first posts” reviews?

Today, on C.SE, I reviewed some 'Late Answers' and First Posts. I was not even near my daily vote limit. And yet, I was given a message that said comments disabled on deleted / locked posts / reviews It seems I was not the only one either: Why are comments now disabled on "first posts" r...

this looks like (a pretty lame imNSho) attempt to salvage "known good" audits in this queue, which fail miserably if one can comment. Next (lame) steps will likely be disabling flagging and edits, thus making whole review a victim of wrongly picked audit strategy. Why doesn't SE team consider keeping only "known bad" audits in FP / LA queues, simply escapes my mind — gnat 1 min ago

7

A: Review Audit/Commenting on good questions discouraged?

As far as I can tell, audits in First Posts are designed based on assumption that algorithm for known good items automatically selects posts that are so perfect that no comments are needed at all. Above assumption is absurd, plain and simple. Such posts are impossible to pick even manually becau...

overall, "known good" audits in First Posts and Late Answers are as if designed specifically to punish thorough reviewers - those who can spot things worth commenting, editing, flagging. Robo types would likely hop over these "audits" with ease, by thoughtlessly clicking No Action Needed — gnat Sep 22 at 7:12

gnat

7:56 AM

@GlenH7 found an obscure post on some tweaks in hotness formula...

9

A: How do the "arbitrary hotness points" work on the new Stack Exchange home page?

Basically what's documented here: http://meta.stackoverflow.com/questions/11602/what-formula-should-be-used-to-determine-hot-questions We have a few tweaks: Succeeding questions from the same site are penalized by increasing amounts. So, the first question from SO in the list gets multip...

Succeeding questions from the same site are penalized by increasing amounts. So, the first question from SO in the list gets multiplied by 1.0, the second by 0.98, the third by 0.96, etc)

Community wiki questions are penalized, to keep the entire home page from being Poll-type questions

The benefit of many answers is capped at 10, and we only look at the score of the top 3 answers

We only degrade based on question age, and not the last update date on a question, so questions don't pop back up to the top every time they're edited

user41796

12:12 PM

@MattD async is over-hyped. It's great in some cases, but not every case.

user41796

@gnat at first glance, that seems a bit asinine. It presumes that subsequent questions aren't as good or better than the current hot question. I think it's good to cap how many hot questions come from a single site, but that's not how I would have handled it.

user41796

12:31 PM

@gnat if I have some time today, I'll re-run the analysis based upon that answer.

gnat

1:07 PM

@GlenH7 well, per-site penalty seems rather "soft"; at 2% for every next question, it would take like 25 (!) questions from the same site to decrease score by half, go figure. As for adjustments to analysis, I wouldn't be surprised if answers-quantity issue will become even more prominent. Look, they totally ignore all but 3 top answers, plus capping on 10 questions doesn't help much

> Even with tweaked formula, stuffing 8 useless, zero-score answers into +50 question would have the same effect as giving 80 upvotes to answers. At +200 question, this would be like giving 320 (over three hundreds!) upvotes to answers...

and there's even a real, recent example for that:

9 mediocre answers stuffed into it in 20 minutes after posting the question likely made it artificially high on collider, attracting views and votes. It's neither the asker's nor voter's fault that it has gotten more attention than it probably deserves; it's just a typical effect of broken "hotness formula" — gnat Sep 16 at 9:24

9 quick mediocre answers bumped the mediocre question to the top with ease

Yeah "benefit of many answers is capped at 10", how cute.

The core of the formula (without the site-based degrading or traffic scaling) is:

(MIN(AnswerCount, 10) * QScore) / 5 + AnswerScore
-------------------------------------------------
MAX(QAgeInHours + 1, 6) ^ 1.4

gnat

1:42 PM

I think it's also worth mentioning that even cap at 10 answers doesn't really address the issue of using "bad data points" in the formula - "Even with tweaked formula, stuffing 8 useless, zero-score answers into +50 question would have the same effect as giving 80 upvotes to answers. At +200 question, this would be like giving 320 (over three hundreds!) upvotes to answers... and there's even a real, recent example for that..." — gnat 40 secs ago

user55340

2:00 PM

@enderland glad it helped. It really redoes how you think about databases.

enderland

2:32 PM

@MichaelT now if you can only solve the problem of constantly changing project requirements you'll be forever my hero! :)

user55340

2:51 PM

@enderland After you get the requirements, do something to the business owners so that they are unable to give you more requirements. I'd suggest cutting off hands and tongue though some may consider this a bit extreme (or even an anti-pattern). I call this methodology MDD - Maiming Driven Development.

Jimmy Hoffa

@MichaelT I suspect that GCD question, the OP is just never going to come back to even see so it doesn't matter. He doesn't even know it was closed.

he was last seen an hour after posting his Q

user55340

@JimmyHoffa It was probably due Monday morning and I didn't give him the answer soon enough.

user55340

However he was on SO yesterday, so he did get a notification at some point of a new answer to his question.

Jimmy Hoffa

@MichaelT You suggest he knows how to operate SE well enough to pick up on that

enderland

@MichaelT see the problem is the requirements umm are not really defined

I realized this this morning

they "think" they know what they "want"

user55340

3:08 PM

@enderland One must not remove the brains before getting the requirements. I do realize this can be a challenge - in many organizations lobotomies accompany the move up the org chart (this is simplified by a 'hire from within' policy).

enderland

heh. I'm not superman - too many hats to wear in thos job

user55340

Heh...

user55340

12

A: Is Chris Jaeger really a final fantasy cactuar?

Yes

user55340

Look at the source for that answer...

Jimmy Hoffa

@MichaelT I wholeheartedly disagree, after 13 years of gathering and implementing requirements, if requirements were well formed and useful I don't think I'd know what to do with them, I'd have to rewrite them in broken english, and paste arbitrary unrelated information from news papers randomly throughout to make them look like something I'm used to

enderland

3:14 PM

@JimmyHoffa dont forget they have to change at least once a day

user55340

Hmmm... you need a /review queue for requirements. "This requirement is VLQ, recommend deletion." "This requirement change vandalizes or significantly alters the meaning of the requirement."

user55340

As to the source of the answer I linked above - meta.stackoverflow.com/revisions/…

enderland

6

A: In hotness formula, discard answers when voting evidence indicates that these are not good data points

Sorry, marking this as status-declined. We love seeing people contributing in concrete ways (especially backed up by research!) but in this case, not knowing the implementation details makes it almost impossible for you guys to solve this one. The hotness formula that's used for the network ho...

Jimmy Hoffa

@enderland I've taken to giving BA's paper with those little spinners on them like a decoder ring that has the slat for a message to allow them to just turn it a little to change the requirement each day

enderland

@JimmyHoffa at least you have BAs :P

this is a huge drawback of being a solo dev in a compnay

Jimmy Hoffa

3:23 PM

@enderland I was making a joke; actually I have none. And it's much better this way compared to the last place I was with them...

enderland

@JimmyHoffa yeah I kinda like being able to directly talk with the stakeholders

but since they have zero appreciation or udnerstanding of software dev it's hard

"what do you mean it's hard to code an entire project in a month? !!!"

Jimmy Hoffa

BAs can be an improvement or a terrible nightmare, just depends. Here what we really need is a PM, way more than a BA. We have a marketing department in the distance that throws random requirements over, but due to low surface contact between them and us, there's very little churn on the reqs, actually works kind of well

enderland

yeah I'm basically PM/dev on this project

Jimmy Hoffa

We end up going off their initial gut feel, and have something working before we end up getting feedback from them because we don't talk enough, feedback obviously asks fro changes but it's not a change every week so we can finish something for them to see before they ask for changes which results in more reasonable changes because they can actually put their hands on it

More than anything we just need someone tracking all the conversations and changes that do float around keeping shit from getting lost/forgotten, a good PM. Oh well, very small dev teams here removes tons of the problems I'm used to dealing with in 30+ man dev teams

gnat

@MichaelT revisions history, side by side markdown, tells the story better: meta.stackoverflow.com/posts/137665/revisions

2

user55340

3:36 PM

Previous employeer, PMs entire duties consisted of creating the schedule at the start of the project based on estimations and sending out emails when people completed a task that was more than 10% over or under the estimated amount (if you're 10% over the estimated hours logged, you get emails, if you are late on the completion date you get emails... and of course, it takes a bit to fill out the email forms you get making you run even later...).

Transcript for