State of the project application issue queue 2013

klausi's picture

Here is a brief status update of what is going on in the project application issue queue lately.

I am quite happy with the review process, it is quite fast when people take part in the review bonus program. I have no problem keeping up with review bonus applications, it generates quite the right workload for me that I'm able to spend on the application queue. As you can see from the fixed applications the duration for getting approved varies, but it is possible to get through in a couple of days, which is awesome! The average application takes longer, but we manage to get through active review bonus applicants in just a few weeks.

Let me say thank you at this point to all applicants that helped each other and all other senior reviewers and git administrators for providing feedback and approval.

Some statistics: we have currently 76 applications that need a review, down from 100 in August 2012. 2296 project applications have been created since February 2011, or +490 applications since August 2012. 2-3 new applications get created per day, down from 3-4 in August 2012. Overall 37% of applications got approved, 9% were duplicates and 41% got rejected. The rest of 13% is somewhere active in the process.

A more worrying indicator is the current RTBC issue backlog: 28 applications are waiting for a git administrator to be approved. You might also wonder why the number of won't fixed issues (41%) is so high: I'm doing regular issue queue cleanups where I close out applications that did not get an applicant response within 10 weeks. Of course the applicants can always reopen those when they find the time to fix the remarks that have been found by the reviewers.

I have also been involved in getting new code review administrators on board: stBorchert and cubeinspire have been mentored and successfully converted! Also: jthorson is now officially listed, too, and also helps out actively.

I also tried to start some sort of mentoring program for new code review administrators where they create a list of review comments they did until we find them trustworthy. I would appreciate your help in motivating and poking those people to keep up the good work. We have mentoring pages for carwin, gisle, anwar_max, bhosmer and fr3shw3b. Let's get back to them together and encourage them to finish off their efforts.

You might also be interested in our review process automation efforts: there has not been any progress on drupal.org (except for the general Drupal 7 upgrade work done by others), but our usual tools at ventral.org for reviewing and simplytest.me for testing work quite reliably for the time being. Big thanks to patrickd for the hard work on this!

Finally there has also been some criticism about the project application review process at levelten and comm-press. That could serve as input for future plans we develop for this process.

So that's it from me, now I would like to hear your news in the comments. How are we doing?

Comments

Thank you klausi for making

Artusamak's picture

Thank you klausi for making these move forward, i think that the process is really smoother right now and we really have more people doing project application review!

Kudos to all of you guys that help make this work!

Thanks for the great summary!

patrickd's picture

Sorry for still letting you down, I really need to kick my self in the as and start reviewing again :/ me so lazy

I don't know. People help in

greggles's picture

I don't know. People help in different ways. As far as I can tell you are helping greatly providing tools to make the process better for applicants and reviewers. That's very valuable in my opinion even if it isn't "doing reviews."

simplytest.me FTW

Wim Leers's picture

I completely agree with greggles. I'd much rather see you continuing to kick ass with simplytest.me than doing project application reviews. simplytest.me helps more with project application reviews (and general patch review!) than doing singular patch reviews. You're helping to increase efficiency/velocity. :)

thanks :)

patrickd's picture

But one review per day would already help and it wouldn't be too much to ask for^^
(It's like doing sport, hard to start but once you started it's easier to keep going)

First of all, you're all are

iwhitcomb's picture

First of all, you're all are doing a great job, I realize these things eat up a lot of your time and I do appreciate the work that you do.

Klausi, being that you were the one to do the final review on my application, you did get back to me quickly after I did the review bonus, I think it was something like a day or two.

I know I flipped, but at the time I was mad, and did feel like it had been a waste of time, as should be expected. If I weren't as patient as I am then I would have given up after the first, or maybe the second application -- I'm now on my fourth. In hindsight I realize that I wasn't alone in my frustration. My beef really isn't with you or anyone else involved with running the queue, I just think there are some kinks in the process. Frank Ralf(comm-press) spoke a bit more eloquently about it than I did in my blog post.

41% rejected is a pretty substantial number, it seems that everyone should be asking what can be done to make the process more efficient(ie: more accepted, fewer rejected, and generally easier for everyone). As I look through some of the rejected applications it seems like quite a few of them were closed due to lack of activity which begs the question, are people just getting frustrated and giving up?

It seems like the review bonus is needed to get any real movement on your application, which is fine(and I appreciate), if it's required then it just needs to be made known that it is a requirement.

As for the PAReview tool, it's awesome, but as I've said(and you've documented), It shouldn't be used as the basis for doing the reviews, too many people just get lazy.

I don't know the answer, but I really don't feel like applying more and more rules around getting a project through the queue is it. Personally, I'd much rather see fewer rules, and let the community decide, maybe with some kind of rating system. Wordpress, for instance, has little to no moderation on new plugins. They tend to get a lot of junk, but the community does a pretty good job of filtering out bad plugins and letting the good ones bubble to the top.

Again, thanks to everyone for hearing me out, please let me know what I can do.

If you look at the projects

greggles's picture

41% rejected is a pretty substantial number

If you look at the projects that get rejected you might have a different conclusion ;)

Personally, I'd much rather see fewer rules, and let the community decide

Well, right now we are letting the community decide...there's a queue where more members of the community could do reviews but they aren't.

Wordpress, for instance, has little to no moderation on new plugins.

Can you point to some detail on that? In 2009 a friend's theme got stuck in the WordPress moderation queue for basically the same reasons that projects get stuck in Drupal's queue. So, my own very limited experience with WordPress is that they actually are very similar.

Well, I've taken over my fair

iwhitcomb's picture

Well, I've taken over my fair share of really crappy Drupal sites to at least know what's possible lol... I don't really think that no moderation is the answer either. Security would be an obvious issue among other things.

I can 100% agree that if you're trying to get something into core then the rules should absolutely apply, but the level of perfection that's expected of new contributors is really discouraging.

In terms of getting involved with the community, IMO getting a module/theme committed is like step one for anyone seriously considering Drupal as a full-time job. There are, of course, other ways to contribute, but getting your code out there has a kind of permanence. It's something you can show off and say "yeah, I did that," even if it's not super amazing. It's also a good way to make people feel like their contributions actually matter, and to motivate them to keep doing it.

It's been several years for me, but I was pretty heavily involved with Wordpress at one point. I've actually got a few plugins out there. I don't remember ever going through much moderation at all. It's a moot point, but the instructions on their site say to simply fill out a form with a description of your plugin(you may be asked for more info). Then they'll setup the repo and you're good to go. http://wordpress.org/extend/plugins/about

Within some vaguely defined

greggles's picture

Within some vaguely defined amount of time, your plugin will be manually reviewed. You may be emailed and asked to provide more information.

That sounds really similar to our process, right?

Let's survey the 41%

sreynen's picture

As I look through some of the rejected applications it seems like quite a few of them were closed due to lack of activity which begs the question, are people just getting frustrated and giving up?

I haven't run the numbers, but after looking at several random examples, my impression is that most of the abandoned applications have nothing to be frustrated about, as they never responded to the very first review of their project. By default there's no email for application comments, so I suspect a substantial portion of these never realized there was something for them to reply to.

This is, of course, as much theory as the idea that they're giving up. I suggest testing these theories by contacting people who abandoned applications and ask them why they never came back. Many people have said they'd like to help improve this process but don't feel they have enough time or experience to do reviews. This is something that would help and takes little time and no experience. Just contact people with their contact form, ask why they abandoned their applications, and post results here to better inform discussion.

Any volunteers?

Good job

cubeinspire's picture

Thank you for all that analysis and coordination Klausi, thumbs up!

cube inspire - web design and web development solutions !

Code Review for opt in security advisory coverage | Home

Group organizers

Group notifications

This group offers an RSS feed. Or subscribe to these personalized, sitewide feeds:

Hot content this week