42
\$\begingroup\$

Uh question, why have you folks not graduated yet? - jul 13 at 21:18 by Undo

It's been quite a while since the 100 days later update, which was posted around the 100th day after our last site review, which brings us back in November 2013.

If a baby was conceived back then, that baby is probably born today. I think it's time for another update.


Zombies

Time to answer

Questions per day has been stable for quite a while, constantly revolving around 30. Time-to-answer has improved:

answer delay graph

  • Weekly "Quick Answers" (answered within 24 hours) has doubled, and went from 63 in mid-November, to 130 as of last SEDE update (with a peak at 162 in April).
  • Quick answers were answered in about 5 hours in mid-November; now 4 hours average - combined with the previous metric, this means our average time to answer has greatly improved since then!
  • Zombie population has had an all-time low of around 550 in April, and has been on a slightly increasing curve since then. Combined with the increased answering activiy, this has kept the site at 95% answered since the 100 days update.

Unanswered

Looking at the top 15 unanswered tags, we can see which tags are pushing the time-to-answer metric upwards:

tagged zombies

  • seems to need more reviewers
  • and zombies are slowly increasing
  • zombies have taken a serious beating in April, and have been a relatively-slowly recovering endangered species ever since.
  • Most of the other top-tagged zombies are also on a slowly increasing trend.

Bottom line

We'll never have enough reviewers. We're keeping up with incoming questions, but with the increased question-asking activity, in order to get the site's Answered% up closer to 100%, we'll need more answering activity overall.


New Question Badges

Stack Exchange recently introduced new question badges; I find they make a useful measurement of askers' :

  • Curious: Asked a well-received question on 5 separate days, and maintained a positive question record. 292 users have earned it so far.
  • Inquisitive: Asked a well-received question on 30 separate days, and maintained a positive question record. 4 users have earned it so far:
  • Socratic: Asked a well-received question on 100 separate days, and maintained a positive question record. Only one user has asked that many questions:

Looks like we do have returning customers after all!


Avid Users Reputation

As of last SEDE update, there's a total of 1,694,927 reputation points on the entire Code Review site, 60.8% of which (1,030,769) being held by "avid users" (>150pts) that were active in the last 60 days.

  • 10,954 users active in the last 60 days, 1,604 of which are "avid users" (14.6%).
  • An active user had, on average, 154.3 reputation points.
  • Active "avid users" had an average of 642.2 reputation points.

When the race with PCG started in mid-March, we had:

  • 1,312,342 reputation points on the site, 54% of which (703,208) being held by "avid users" active in the previous 60 days.
  • 9,786 user active in the previous 60 days, 1,319 of which were "avid users" (13%).
  • An active user had, on average, 134.1 reputation points.
  • Active "avid users" had an average of 533.1 reputation points.

Still as of last SEDE update, we have 5 users above 20K, 14 above 10K, 54 above 3K and 78 above 2K. And we've gained another 20K user today! (and I'm not too far behind either!)


The 2nd Monitor

The site's main chatroom has seen a number of new faces - the Talkative badge has been awarded 141 times (Outspoken x 42), and a lot of these users can often be found in The 2nd Monitor. This is great news, because since mid-November our chatroom has been a cradle for a lot of good things; seeing its activity continuously increase can only be a good thing all by itself.


Are we due for another site review? Can we expect a graduation announcement anywhere in 2014? If not, what do we need to do (besides "keep doing what we do") to get there?

\$\endgroup\$
  • 3
    \$\begingroup\$ "our chatroom has been a cradle for a lot of good things" - It's a FIM++ program incubator! \$\endgroup\$ – Alex L Jul 24 '14 at 20:25
  • 1
    \$\begingroup\$ I have no idea what you're talking about. \$\endgroup\$ – Mathieu Guindon Jul 25 '14 at 3:54
  • 7
    \$\begingroup\$ @AlexL I'm going to write a Pony Code compiler just so I can post a letter for review here. I swear it. \$\endgroup\$ – RubberDuck Jul 25 '14 at 4:25
  • 15
    \$\begingroup\$ I'm starting to feel that we're kids in the backseat of a car going "Are we there yet?" over and over again. I hope the Community Managers does not think of us like that. \$\endgroup\$ – Simon Forsberg Jul 25 '14 at 9:22
  • 1
    \$\begingroup\$ @SimonAndréForsberg Mat said something similar a while back: chat.stackexchange.com/transcript/8595?m=14021612#14021612 \$\endgroup\$ – Vogel612 Jul 25 '14 at 11:47
  • 3
    \$\begingroup\$ There is one thing that everyone should know, I'm so proud of all we are doing as a community! We're working a lot, and I think it's showing up in the increase activity of CR! \$\endgroup\$ – Marc-Andre Jul 25 '14 at 15:35
  • 4
    \$\begingroup\$ I posted an answer to another question about the status of Code Review earlier today. It's not exactly an answer to this question, but it's fairly relevant. While I was writing it, I thought to myself that this site looked good a year ago, and it's only gotten better since then. I'll try to find time for a full review sometime this week. \$\endgroup\$ – Pops Aug 5 '14 at 20:07
18
\$\begingroup\$

I have read most of the unanswered Python questions on Code Review. It appears that there are two major kinds of questions that are not answered:

  • The ones asking for a performance optimization of a piece of code.
  • The ones whose code rely on non-standard libraries.

The problem with optimization is that you have to write benchmarks in order to provide a meaningful answer and people generally don't bother to write benchmarks, and the results may still depend on the platform and not be relevant. The only way to provide meaningful performance advice without benchmarks is to know your algorithms. Some algorithms are pretty tricky and it is not trivial to reduce the complexity of an operation. Therefore, answering such questions is rather hard and many are left unanswered.

We might know Python, know the differences between Python 2 and 3 and which idioms should be used, but non-standard libraries knowledge is no common knowledge shared by all reviewers. Even if we want to learn about the library, there will be some time before we are confident enough to provide a meaningful answer about library-specific questions. One eight of the unanswered Python questions is about and there are also many questions about (those ones can be even trickier because they often deal with algorithms). I think that I have also seen some questions about sockets and networking.

Therefore, I don't think that Code Review needs more Python generalists, but more people who know about algorithms, networking and about some specific libraries (it would be good to have some Django guru around).

EDIT - 19 August 2014: Also, compared to and , it seems that there aren't that many votes on the questions on answers. That has often been said, but we need you to vote more, people.

\$\endgroup\$
12
\$\begingroup\$

From reading the Post, it looks like 3 things need to happen (at the least)

  1. < joking > @konijn Bot needs to be reactivated to keep the JavaScript Zombies from overtaking Code Review again < /joking >

  2. Java People We should take some time to learn a couple of other languages

    • C#
    • Python
    • JavaScript / jQuery

    so that we can occasionally review posts with these tags, regardless of what are favorites are, this will also bring up the number of Generalist badges as well.

  3. < joking > find a way to hide and keep that junk off our site < /joking >

Seriously though, it looks like we should all take some time to look through the unanswered questions and see if we can make out some things from other languages, even if it is just to activate users that asked a single question and then lost faith in Code Review in a time before the Great Revival.

\$\endgroup\$
  • 11
    \$\begingroup\$ In other words, let's break the language barriers and review all the codez. We already have a Java monkey with a C# badge, I guess we can have a C# mug with a Java badge... \$\endgroup\$ – Mathieu Guindon Jul 28 '14 at 14:20
  • 1
    \$\begingroup\$ According to Stinking Badges!, you don't actually need to review other languages to become Generalist. I got it mostly for tags like optimization, performance, agorithm, beginner, etc. (the only language I have there is c#). Also, if that site is correct, it's not possible to get Generalist at the moment at all, since we have only 38 tags with 200+ questions (we need 40). \$\endgroup\$ – svick Aug 1 '14 at 16:02
  • \$\begingroup\$ @svick that has changed, and there are several Generalist badges floating around CR already, I have one. \$\endgroup\$ – Malachi Aug 1 '14 at 16:04
  • \$\begingroup\$ @Svick, where did you find that cool app!!! \$\endgroup\$ – Malachi Aug 1 '14 at 16:05
  • \$\begingroup\$ @Malachi From Stack Apps, though I don't remember how I found it there. \$\endgroup\$ – svick Aug 1 '14 at 16:08
  • \$\begingroup\$ @svick, that app needs to be updated, I think they used some magic numbers in there or something. \$\endgroup\$ – Malachi Aug 1 '14 at 16:12
  • \$\begingroup\$ @Svick I posted on the post stackapps.com/a/4829/14123 \$\endgroup\$ – Malachi Aug 1 '14 at 16:26
  • 1
    \$\begingroup\$ My apologies, my available time has been sharply reduced ;) I think if you were to provide a nice query for @Flambino, then he could hunt/destroy zombies equally well. \$\endgroup\$ – konijn Aug 5 '14 at 19:31
  • \$\begingroup\$ +1 for 'Java People'! \$\endgroup\$ – h.j.k. Aug 8 '14 at 7:44

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .