25
\$\begingroup\$

As of 2013-09-05, Area 51 stats for CodeReview are as follows:

Days in beta: 960.

Questions per day: 19.4. Healthy at 15, need work at 5. Check.

Answered Questions: 89%. Healthy at 90%. So we're anywhere between 88.5% and 89.4%: I'd love to see the actual decimals. I'm sure we can bring this metric up in no time.

Per StackExchange, Stack Overflow has it at 78%.

Avid Users: 859. Total Users: 26,655.

  • 200+ rep: 859. Recommendation at 150. Check.
  • 2,000+ rep: 34. Recommentation at 10. Check.
  • 3,000+ rep: 18. Recommentation at 5. Check.

Answers per Question: 1.9. Healthy at 2.5. It's code review. I think this is where CR is fairly different from other QA sites: I find 1.9 is an excellent average in the context of code reviews. It means the average code review request gets at least one review, and odds are strong you'll even get two.

Visits per Day: 8,064

Healthy at 1,500, needs work at 500. CHECK!!


(user count as of 2013-09-05, stats as of end-of-beta)

  • Skeptics: 11k users.
    • Q/Day: 5.7 | A: 90% | Users: 464/5698 | Visits: 3786 | A/Q: 1.7
  • Theoretical Computer Science: 13k users.
    • Q/Day: 7.4 | A: 92% | Users: 164/2432 | Visits: 1688 | A/Q: 2.8
  • Travel: 6.8k users.
    • Q/Day: 5.4 | A: 100% | Users: 427/3848 | Visits: 4306 | A/Q: 2.1
  • Mathematica: 7.7k users.
    • Q/Day: 12 | A: 96% | Users: 230/1708 | Visits: 1307 | A/Q: 2.2
  • Science Fiction & Fantasy: 14k users.
    • Q/Day: 8.2 | A: 96% | Users: 400/3375 | Visits: 1632 | A/Q: 2.3
  • Role-Playing Games: 6.1k users.
    • Q/Day: 3.4 | A: 100% | Users: 408/2725 | Visits: 1667 | A/Q: 3.5
  • Photography: 15k users.
    • Q/Day: 4.9 | A: 99% | Users: 162/1783 | Visits: 806 | A/Q: 3.6
  • Mi Yodeya: 2.4k users.
    • Q/Day: 6.6 | A: 97% | Users: 163/1072 | Visits: 758 | A/Q: 2.2
  • Sharepoint: 14k users.
    • Q/Day: 20.9 | A: 91% | Users: 171/2931 | Visits: 5071 | A/Q: 1.8

All have graduated, and all did so with one or more metric at or below "Ok".

I think we're more in the league of:

  • WebMasters: 24k users.
    • Q/day: 9.3 | Users: 125/2297 | Visits: 658 | A/Q: 3.0
  • Game Development: 29k users.
    • Q/day: 6.6 | Users: 166/2611 | Visits: 1180 | A/Q: 3.9
  • Database Administrator: 22k users.
    • Q/day: 8.9 | Users: 170/2652 | Visits: 2060 | A/Q: 2.2
  • User Experience: 27k users.
    • Q/day: 6.7 | Users: 452/8939 | Visits: 1465 | A/Q: 3.8
  • WordPress Answers: 28k users.
    • Q/day: 27.3 | Users: 118/3115 | Visits: 3463 | A/Q: 2.0

Clearly, having all metrics at "Excellent" is not a requirement for graduation (Photography graduated with a "needs work" Q/Day of 4.9; RPG did with Q/Day at 3.4!). So what the ?&@! is going on? CR has similar A/Q ratio (1.9) to WordPress at the end of beta (2.0), and better than Sharepoint (1.8) and Skeptics (1.7). CR has more questions per day than the average beta site (about 2x?), and many more daily visits than any other. The more you dig, the more you wonder why CR is still a beta site.


I already gave my answer to meta-question "Why is Code Review still in beta?", where I hinted that we should make t-shirts to promote the site... And I was serious about it (I'd actually pay for one!)

Looks like it's time for t-shirts... I envision a set of black-on-white or white-on-black tees with the CR logo in the middle, with a small witty phrase around it and the url at the bottom (the backs of all tees would be just the logo and the url).

Here's my take on it:

think you code crap? you're not alone.

make it happen. (beta needs you)

my code is CR-approved.

And of course the obligatory: I♥CR

(of course this isn't really an appropriate topic. the idea however, is legitimate: CR is an amazing resource with huge potential, and letting it shut down would be a terrible loss, so let's do everything we can to make it bloom - including turning ourselves into walking ads! ...and then you could have mugs, keychains, usb keys, bumper stickers...)

Feel free to share your own :)

\$\endgroup\$
  • 3
    \$\begingroup\$ If all else fails, I'd say we should start with the front page design. ;-) It'll only be beneficial if and when this site graduates, of course. \$\endgroup\$ – Jamal Sep 5 '13 at 1:07
  • 1
    \$\begingroup\$ Yes, that's correct. All betas look like this, so even the site name doesn't stand out. But, I don't want to get too optimistic. We still don't know for sure what's holding back graduation (I'm assuming it's the low voting and somewhat high number of unanswered questions). \$\endgroup\$ – Jamal Sep 5 '13 at 1:13
  • \$\begingroup\$ For my part I don't think I have a huge expertise, I added favorite tags c#, vb6 and vba but reading through recent unanswered C# reviews there's not much to add (except votes!), the newest VB6 review request is over a year old and there's only 5 of them.. VBA seems more active. I really think the root of the issue is low traffic. I think t-shirts could help (and mugs/keychains for that matter), since they would make the site more visible online (given SE would sell them online). \$\endgroup\$ – Mathieu Guindon Sep 5 '13 at 1:31
  • 1
    \$\begingroup\$ Another thing that gives online publicity is more unique questions. Questions on CR do show up in search engines, which is what gives all SE sites most of their traffic. Fortunately, the questions are still coming, and Area 51's stats on this are good. \$\endgroup\$ – Jamal Sep 5 '13 at 1:43
  • 1
    \$\begingroup\$ Actually ALL Area51 stats for CR are either "Excellent" or "Ok". One would think going official is imminent! \$\endgroup\$ – Mathieu Guindon Sep 5 '13 at 3:59
  • 1
    \$\begingroup\$ Seems like they should all be excellent, if that's what allows a site to graduate. I've answered a few unanswered C++ questions before, but there's nothing else I'm capable of answering (I think). \$\endgroup\$ – Jamal Sep 5 '13 at 4:20
  • \$\begingroup\$ Then we need only 1 percent point more answered questions and the main issue to work on is the answers per question ratio! \$\endgroup\$ – Mathieu Guindon Sep 5 '13 at 12:19
  • 1
    \$\begingroup\$ Yes, that'd help a lot. But, that won't happen until more people look over the unanswered questions. I figure time is a large factor here. \$\endgroup\$ – Jamal Sep 5 '13 at 12:22
  • \$\begingroup\$ let us continue this discussion in chat \$\endgroup\$ – Mathieu Guindon Sep 6 '13 at 1:23
  • 3
    \$\begingroup\$ We're still waiting for the feedback on the community evaluation. The issue with CR is not stats at any moment in time, but the fact that the site isn't able to attract long-time users. \$\endgroup\$ – Quentin Pradet Sep 11 '13 at 9:17
  • 1
    \$\begingroup\$ I think you should update the title now. ;-) \$\endgroup\$ – Jamal Oct 21 '13 at 7:45
  • \$\begingroup\$ I removed "Approaching", but don't count on me to keep the number of days up-to-date! ;) \$\endgroup\$ – Mathieu Guindon Oct 21 '13 at 11:21
15
\$\begingroup\$

I'd like to bring your attention to a recent comment conversation I've had with SE community manager Tim Post. I've embedded the comments in case they go away (italic emphasis added by me):

Can we add Code Review to the migration list?

Me:

Although my response is late, CR is still having this problem [with voting]. Some of the Area 51 stats could be a problem. Today marks Day 1001 of beta status.

TP:

The only thing holding CR back from graduating is the lack of high rep users - there just aren't enough 1k+, 3k+ users on the site, and it can't quite yet sustain an election. Every single thing about CR is ready but that - it's really just a matter of time at this point (and, well, voting).

Me:

Assuming the voting thing won't improve sufficiently at some point, I hope CR isn't prone to being shut down.

TP:

We'd never dream of closing such an active and otherwise healthy site. Once it can sustain an election and we're confident that community moderation won't be too adversely affected by shifting to the 'graduated' privileges scale, I'm sure it'll go. Kind of perplexed as to why voting is so light there, though, but that's just me chewing on it - nothing more :)

\$\endgroup\$
  • 1
    \$\begingroup\$ Cool! So all we're missing is code review gurus :) \$\endgroup\$ – Mathieu Guindon Oct 17 '13 at 15:57
  • 1
    \$\begingroup\$ @retailcoder: Indeed. Unfortunately, I don't think a large expert base from SO will come here (a number of reasons that I can see). For instance, there's no "Fastest Gun in the West" (at least right now). But, the number of bad questions is far below the number of good questions. Still, which would an expert prefer: giving a short but badass answer in a minute or so, or writing a lengthy review (and with no guarantee of receiving a lot of upvotes)? \$\endgroup\$ – Jamal Oct 17 '13 at 16:02
  • 2
    \$\begingroup\$ Well knowing what our problem is, perhaps the best CR tee would be something like "VOTE!! It's good for you!"... I don't think there can be a Jon Skeet -grade reviewer - a "fastest gun in the west", that gets dozens of votes on every answer they give. We just need to emphasize on the importance of voting. \$\endgroup\$ – Mathieu Guindon Oct 17 '13 at 16:08
  • \$\begingroup\$ @retailcoder: Yes, this should definitely be made aware to everyone. Most importantly, questions should be voted on. They require no programming experience, unlike answers. If a question is written clearly overall shows research (if not for a general review), an upvote is good. If it lacks research when needed or... is just crap, downvote it. The same applies to answers, except voting is best determined on the accuracy of the content (upvote if useful/correct, downvote if useless/misleading). \$\endgroup\$ – Jamal Oct 17 '13 at 16:16
15
\$\begingroup\$

There's a reason why this site is still in beta:


NOBODY VOTES


Simple. Sorry for caps but it's true.

\$\endgroup\$
  • \$\begingroup\$ I too have to agree with this. :/ The number of Suffrage and Vox Populi (perhaps also Civic Duty) badges awarded on main is very low. \$\endgroup\$ – Jamal Sep 13 '13 at 23:33
  • 3
    \$\begingroup\$ I see your point, I too, until recently, was pretty much taking this site for granted, only posting when I needed to and then thank-you-bye-see-you-next-time-I-need-a-review - but it's a beta site and if we all do that it will eventually be gone. And that would be bad. \$\endgroup\$ – Mathieu Guindon Sep 14 '13 at 1:38
  • 8
    \$\begingroup\$ I'll correct you there. Not enough people vote. I can't share the specifics, but we do get thousands of votes every month. The challenge is to improve that further, along with our our answer ratio and number of answered questions. \$\endgroup\$ – Adam Sep 15 '13 at 2:36
  • \$\begingroup\$ @codesparkle: I've heard an argument before about the answer ratio being a "special case" for this site. Specifically, many questions get (good) full reviews, thereby giving little room for further answers. As for the answered questions, I've already answered a few C++ questions that are within my expertise. I could also look for tags that may need to be burninated. \$\endgroup\$ – Jamal Sep 15 '13 at 4:50
  • \$\begingroup\$ Voted this answer up..... :) \$\endgroup\$ – rolfl Jan 8 '14 at 21:39
  • 1
    \$\begingroup\$ Hm, from my personal anecdotal evidence: I got 5041 rep for 139 answers on CR = 36 rep/answer, I got 12245 rep for 548 answers on SO = 22 rep/answer. Unless for some reason all my answers on SO are of much lower quality I can say voting on CR is much better for me than on SO. \$\endgroup\$ – ChrisWue Jan 10 '14 at 9:03
  • \$\begingroup\$ Last I saw, SO had a lesser rep bonus on each action/vote and that would easily explain the 22/36 ratio. \$\endgroup\$ – Jonny Sooter Jan 11 '14 at 10:17
4
\$\begingroup\$

So here we are. In just a couple days CR will have been a beta site for a thousand days.

I think the site is healthy. If 1000 days weren't enough to prove it can't be successful, nothing will. CodeReview is definitely different and it's widely agreed that the Q&A format is more or less an optimal format. Let's just live with it.

CR is, regardless of how much of a good fit it is or isn't for SE's Q&A format, a StackExchange site. In the programming area, SE has its community, a community that's always growing, a community that loves SE's Q&A format. No other platform would be better, and if there was, it would be at StackExchange's loss. People google up programming stuff and filter their results with ":stackoverflow" (I know I do!) for a reason: that's where the best answers are, and that's all thanks to the community.

I have a few ideas:

  • Keep the beta privilege thresholds, or make them somewhere between what they are now and what they would be on a normal site. Reviewing code isn't like answering a question you know the answer to. You can't just read the title and say "here, that one I know I can answer!". Most code reviews demand a lot more effort than the average answer on any other Q&A site, and the laws of common sense say there will be fewer views, fewer answers and fewer rep awarded than on other Q&A sites. By going live with lower thresholds than SO, we will encourage more people to stick around and earn the perks, because rep is harder to earn here, and 10K is a longshot, at least for the common of mortals.

  • Stop worrying about the scope. The off-topic reasons are fine as they are. Even SO has its truckload of borderline questions and some that are clearly off-topic that get flagged as such two years later, even with the traffic they get. If a question is bad, it will get few answers if any, downvotes and low views; that's how SE works!

  • Introduce more answer badges. Current answer badges seem to focus on the votes you get out of them. Which is good to encourage higher quality answers, but not enough to encourage answering many questions. It seems badges are shared across all sites, so I'm not going to make them CR-specific, but here are some ideas:

    • bronze Resourceful: Answered 10 questions with a score of 1 or more.
    • silver Ubiquitous: Answered 80 questions with a score of 1 or more.
    • gold Encyclopedic: Answered 200 questions with a score of 1 or more.

1000 days: time has come to graduate


What I mean is, CR is StackExchange, and I wouldn't want to get my code reviewed anywhere else. The world needs a place like CR. Enough debate already, let's just go with it: asking for a code review is off-topic everywhere else on the network and that's enough of a reason to have CR around.

1000 days, 9.8K questions and 17K answers later, with 10K+ visits per day, the demand is undeniable and if a code review website has to exist, it being part of the StackExchange network is, in my humble opinion, the only way a code review website is going to be credible.

\$\endgroup\$
3
\$\begingroup\$

I just came across the perfect slogan for a T-shirt: "Does this code make my ass look fat?"

\$\endgroup\$
  • \$\begingroup\$ Haha! Credits to Jeff Atwood himself! \$\endgroup\$ – Mathieu Guindon Dec 12 '13 at 21:13
  • 1
    \$\begingroup\$ Actually, credits to @bob-the-destroyer. \$\endgroup\$ – 200_success Dec 12 '13 at 21:14
  • 1
    \$\begingroup\$ And me for bumping the question? :-) \$\endgroup\$ – Jamal Dec 12 '13 at 22:17
-5
\$\begingroup\$

I actually love the site to be beta. It somehow gives it the right flair. Just posted a "question" about it on Meta, actually.

Is there a problem being Beta, especially when it is about code-review? I don't think so. So let it be 10 thousand days; all fine from a reviewer's perspective.

\$\endgroup\$
  • 4
    \$\begingroup\$ I have my own reasons for liking and disliking beta. I like it because it allows us to keep testing the site (hence the "beta" status) and let it grow without great demands. I also don't like it because, with any other beta, it still hasn't been given a permanent place. In other words, any beta site can be shut down if it stops growing and improving. Overall, the difference between shut down and graduation starts with the community. Visitors from search engines also play a big role, and we have to do what we can to keep bringing them in. \$\endgroup\$ – Jamal Sep 7 '13 at 22:16
  • \$\begingroup\$ @Jamal: I think all of that is fair and dandy. I also didn't mean to push my personal taste too much forward here. Just was expressing my feelings with a question of it's own and finding others concerned about Beta so thought this was related. Thanks for sharing (and personally I don't think that this Beta will get killed easily, I mean it's not an ordinary Beta, isn't it?) \$\endgroup\$ – hakre Sep 7 '13 at 22:20
  • \$\begingroup\$ Not a problem! This is also what Meta is for. What do you mean by "ordinary" beta? A beta site is a beta site, regardless of the time spent in that state. One established point regarding graduation is that there's no set time that a site will be beta. \$\endgroup\$ – Jamal Sep 7 '13 at 22:23
  • \$\begingroup\$ I read on meta.stackoverflow that beta sites shouldn't be migration targets. Beta sites don't get the visibility of "graduated" ones, and can be shut down I guess if stats fall below a threshold other sites don't need to worry about. I think part of the problem is that we're still somewhat defining the scope after all this time.. and that is actually what I don't mind - let it graduate as is! I don't mind the look & feel either, perhaps once CR graduates we can keep it similar. \$\endgroup\$ – Mathieu Guindon Sep 7 '13 at 22:24
  • \$\begingroup\$ Well, I do not know all betas but I'd say that CR is record-breaking so it's not ordinary. Also there is pretty much content so far, isn't it? \$\endgroup\$ – hakre Sep 7 '13 at 22:24
  • \$\begingroup\$ @retailcoder: Well I'd say this is for a good reason and also especially for CR not such a show-stopper. E.g. I personally do direct some (probably) interested users over here from SO. Imagine (for real) one could close against CR for diverse questions. I mean technically all "this is my code (3-260 lines), it's 'not working, why?'" questions are actually concrete code-review questions but honestly, you don't want them. As you probably can imagine, this grey area is pretty large. So perhaps here Beta is a benefit actually. \$\endgroup\$ – hakre Sep 7 '13 at 22:27
  • 1
    \$\begingroup\$ @retailcoder: We, as a community, also have to make sure that we know what is and what isn't allowed. Major policy changes during graduation doesn't seem good, unless it's absolutely necessary. Do you recall Winston Ewert's featured post? That's one example of site policy reevaluation in progress. \$\endgroup\$ – Jamal Sep 7 '13 at 22:27
  • 1
    \$\begingroup\$ @hakre I find the "working code" guideline is fundamental for CR. Make it work and then ask us what we think you could improve. CR really boils down to this imho. \$\endgroup\$ – Mathieu Guindon Sep 7 '13 at 22:46
  • \$\begingroup\$ @retailcoder: Good and necessary point I'd say :D However working code gives errors, doesn't it? And failing should be an option of any code, so failing is working, doesn't it? ^^ \$\endgroup\$ – hakre Sep 7 '13 at 22:47
  • 1
    \$\begingroup\$ How about defining "working" as "fulfills the requirements"? There can be bugs, inefficiencies, design principles being shredded to pieces. That's what reviewing is for no? \$\endgroup\$ – Mathieu Guindon Sep 7 '13 at 22:57
  • \$\begingroup\$ Sure, I mean you folks here with CR are taking a lot of stuff. I wonder you could make it so long actually (my respect). \$\endgroup\$ – hakre Sep 7 '13 at 22:58
  • 2
    \$\begingroup\$ @retailcoder: If the OP knows the code isn't working, then it's off-topic. According to the FAQ, the code must be working to the best of the OP's ability. If errors are found afterwards, they may be included in an answer (along with the review) or as a comment. \$\endgroup\$ – Jamal Sep 8 '13 at 1:26
  • 2
    \$\begingroup\$ One main reason for this website to get out of beta is because it is a little different from other SE websites. It isn't an exact fit for QA format. It needs some additional tools for functioning to the best. A beta website won't get that but a website out of beta can get those features. \$\endgroup\$ – Aseem Bansal Sep 9 '13 at 16:34
  • 5
    \$\begingroup\$ I downvoted because I disagree. The whole point of creating a site is to see it graduate. Beta is temporary — either the site graduates or it dies. \$\endgroup\$ – Adam Sep 15 '13 at 2:22
  • 1
    \$\begingroup\$ @codesparkle: A site can only die if it dips below SE's expectations, regardless of the time it spent in beta, correct? It seems (to me) that we're doing enough to prolongue CR's death, while we're still not quite meeting graduation requirements. I wish I could hear an analysis of CR's performance straight from SE. \$\endgroup\$ – Jamal Sep 15 '13 at 4:58

You must log in to answer this question.

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