Year end: 2010

Wednesday January 5, 2011, Will Kahn-Greene | share this (mastodon)

Note: This is an old post in a blog with a lot of posts over a long span of time. The world has changed, technologies have changed, and I've changed. It's likely this is out of date, the code doesn't work, the ideas haven't aged well, or the ideas were terrible to begin with. Let me know if you think this is something that needs updating.

I'm really proud of the work we did this year. Miro went through two major version releases that added support for conversions and subtitles, fixed a lot of issues, and fixed a lot of performance issues. Behind the scenes, we reworked the Windows build environment making building Miro on Windows sooooo much easier. We also switched from our home-grown httpclient to libcurl for faster, better HTTP downloads. We drastically improved our QA and testing. We migrated from Trac to Mediawiki for documentation.

Then in August, Nick stepped up to home plate, hefted his bat a few times, then pointed his finger clear across the field into the stands. Thus started the very aggressive and ambitious Miro 4.0 development cycle. The earth-shaking thunder of changes hitting git master has been growing into a deafening roar.

But the fun doesn't stop there! Miro Community has come a long way in the last year. There are dozens of really great Miro Community communities now and I regularly see urls to Miro Community site pages in Stack Overflow, forum conversations, blog posts, ....

Additionally, Universal Subtitles has gone from a gleam in someone's eye to a system for collaborating on subtitles and translating subtitles for videos across the Internet. Translating video and making it accessible to people is so important. This system is making that magic happen. It's awesome.

It's been a huge year for PCF and the projects we work on.

I want to bring it back down to the project I work on: Miro. Last year I wrote a year-end post with some stats I culled from Bugzilla and git along with some commentary. I do so again this year.

Bugzilla stats:

I took a look at last year's overall stats and I have no idea where I got those numbers. So I tweaked my script and produced new stats.

Last Year's Overall statistics
------------------------------
                                  2007  2008  2009
Opened reports at end of year:     453   705  1102
Opened:                           4052  1625  1593
Closed:                           4368  2032  1654
Users created:                     644  1083   771
Comments created:                13564  7529  8329


Overall statistics (recalculated)
---------------------------------
                                  2007  2008  2009  2010
Opened reports at end of year:     453   705  1102   720
Created:                          3805  1518  1162   968
Resolved:                         1038  1842  1157  1197
Comments created:                12588  6749  6191  6843
Users created:                     644  1083   771   611

"Opened reports at end of year" is the total number of opened bugs at the end of that year. So at the end of 2007, we had 453 open bugs. At the end of 2008, we had 705 open bugs, ...

As a reminder, we switched from Trac to Bugzilla half-way through 2007 which drastically reduced the amount of bug and comment spam we were getting and dealing with.

Looking at the overall statistics and the recomputed set, the new numbers are lower. I think the old numbers involved all bugs in our Bugzilla instance--Miro and non-Miro bugs. Prior to 2009, that was pretty much the only project we were working on, so the numbers made more sense. Now that Miro Community, Miro Video Converter and Universal Subtitles are using Bugzilla, we need to explicitly focus on Miro bugs.

We have the same problem with the Bugs closed by activity data I did for last year.

The drop in total opened bugs for 2010 happened because a bunch of us did a massive triage campaign at the end of December and resolved a lot of old, stale bugs.

Bugs closed by activity (old set)
---------------------------------

             2007  2008  2009
fixed         736   932   969
invalid       170   133    85
wontfix        35   142    71
duplicate     139   313   190
worksforme    169   344   151
incomplete      0    57    84


Bugs closed by activity (recalculated)
--------------------------------------

             2007  2008  2009  2010
fixed         545   887   648   570
invalid       167   126    72    83
wontfix        34   135    49   160
duplicate     132   305   173   141
worksforme    160   335   133    87
incomplete      0    54    82   156

For most of 2010, there were only two and a half developers working on Miro. We've picked up Geoffrey and Kaz and Paul has come back from Miro Community, so we're churning through fixes faster now. But it's definitely the case that the lower number of fixed bugs is due to reduced staff.

Top 10 bug reporters:
--------------------

     252 - Janet [pcf QA]
     100 - Will Kahn-Greene [pcf dev]
      51 - Geoffrey Lee
      47 - Ben [pcf_dev]
      26 - Nicholas Reville
      23 - Kaz Wesley
      12 - Paul Swartz [PCF dev]
      11 - m.shamraeva (qa-team)
      10 - Dean Jansen
       6 - Nicuta Nicolae

Of the 968 bugs created in 2010, 532 were reported by Miro developers and Janet's QA team.

I'm skipping the top 10 closers this year since it's not clear that number is very meaningful.

Top 10 bug commenters:
---------------------

    3417 - Will Kahn-Greene [pcf dev]
    1111 - Janet [pcf QA]
     525 - Ben [pcf_dev]
     388 - Geoffrey Lee
     192 - Paul Swartz [PCF dev]
     137 - Nicholas Reville
     113 - Luc Heinrich
      64 - Kaz Wesley
      24 - David Stoll
      16 - Dean Jansen

Most of these people are PCF staff. David Stoll is not--he helped a ton in fixing httpauth and http proxy issues for Miro 3.5.

Git stats

In 2010, we did 6 releases (2 major, 4 bugfix). We did 14 releases if you include release candidates. Version 3.0 added subtitles and a lot of performance fixes. Version 3.5 added http proxy and auth support, switch to libcurl, conversions and a lot of polish and some performance fixes.

Between Miro 2.5 and 3.0:
1017 files changed, 190346 insertions(+), 280726 deletions(-)
667 checkins

Between Miro 3.0 and 3.5:
1021 files changed, 307395 insertions(+), 253431 deletions(-)
606 checkins

Paul, Geoffrey and Kaz hopped on the Miro team post 3.5, so I think it's interesting to measure changes between 3.5 and HEAD:

Between Miro 3.5 and git master HEAD:
552 files changed, 31627 insertions(+), 22898 deletions(-)
521 checkins

You'll notice the number of insertions and deletions is an order of magnitude smaller--a lot of that churn happens before a release when I sync translations. I haven't synced translations in git master for 4.0, yet, so the insertion/deletion numbers are much lower.

In 2009, we did 1,382 commits. In 2010, we did 1,341 commits.

Contributor stats

In 2009, we had 19 contributed fixes. I doubt I checked in that many patches this year, but I don't have good stats for that this year.

Having said that, with Miro 3.5 we introduced a CREDITS file which is a much more comprehensive list of contributors since it covers people who report bugs, comment on bugs, fix bugs, translate, help with QA, donate money, ... This CREDITS file will be a better measure of how we're doing contributor-wise than looking at Bugzilla stats alone was. I still need to figure out a better method for figuring out who has helped with translations, testing nightlies, and supporting other Miro users -- those areas are important, but more difficult to quantify.

We continued to work on reducing the barriers to entry for contributing to Miro:

We switched from Trac to Mediawiki which will make it much easier to write, edit, and curate documentation on the project.
We hired Asheesh from OpenHatch to help us figure out how to build a community of contributors.
We overhauled the Windows build environment making it possible for people to build Miro on Windows. Prior to this work, we were using Python 2.5 and Visual Studio 2003 which wasn't available.
We continued triaging bugs and adding the "bitesized" keyword to bugs we think would be easier for new contributors to work on.

There's always more work to do.

Summary

The Miro 2.5 release in 2009 sucked. We've made a lot of improvements to our infrastructure, process, and code quality since then. 2010 was a good year and we got a lot accomplished despite having a painfully small number of staff for most of the year.

I'd love to see more contributions from other people. If you have some free time or some passion and want to help out, let me know. If you don't have free time, but have some spare change floating around, please donate--this helps PCF pay for staff to work on Miro. Having more staff and more contributors absolutely affect Miro's speed of growth.

Also, contributing to Miro gets your name in the Credits! I got my name in the Firefox 3.0 credits for work I did and it was one of the highlights of my year.

Want to comment? Send an email to willkg at bluesock dot org. Include the url for the blog entry in your comment so I have some context as to what you're talking about.