some numbers I drummed up while building Ubuntu packages....

Note: This is an old post in a blog with a lot of posts. The world has changed, technologies have changed, and I've changed. It's likely this is out of date and not representative. Let me know if you think this is something that needs updating.

After that lunch on Wednesday where I talked about how much I really love the numbers and pretty graphs that are on regularly, I wanted to do stats on Miro.


There are two things I'm interested in measuring. The first is measurements related to release cycles and development process. The second is measurements related to contributions.

Anyhow, here are some rough tables:

           tag      tv/    released    cycle
           ------   -----  ----------  -------
Miro 1.0   151 MB   53 MB  11/13/2007  N/A
Miro 1.1   169 MB   58 MB  1/10/2008   58 days
Miro 1.2   253 MB   63 MB  3/20/2008   70 days
  • "tag" - size in MB of the codebase which includes binary kits and other stuff

  • "tv/" - size in MB of just the tv/ directory

  • "released" - release date

  • "cycle" - the length in days of the release cycle

We're still doing tight release cycles. I'm hoping we'll trend towards longer release cycles. Something in the 3 month range would be easier on the devs and probably other people, too.

           bugs fixed all gtk mac win    bugs created all gtk mac win
           ---------- --- --- --- ---    ------------ --- --- --- ---
Miro 1.0   65         18  17  15  15     85           20  17  17  31
Miro 1.1   40         16  6   10  8      106          44  21  20  21
Miro 1.2   82         26  14  13  29     --
  • bugs fixed - number of bugs fixed and then broken down by platform

  • bugs created - number of bugs created against this version and then broken down by platform

I'll let you interpret the data as you like. I think the "bugs fixed" column is indicative of our priorities between the releases: 1.0 was a stability-focused release, 1.1 put out libtorrent, and 1.2 involved a code overhaul which caused a lot of regressions.

Miro 1.0  63
Miro 1.1  66
Miro 1.2  70

I'd like to figure out how to get a rough measure of quality of translations, but I'm not really sure how to go about doing that. I threw together a script to count the number of instances where msgid differs from msgstr, but the results don't seem very indicative of a correctness or completeness figure. Launchpad has statistics, but there's no way to look "back in time" at previous releases that I can find. Are there any ideas for how to do that by looking at the .po files?

          patches from contributors applied
Miro 1.0  4
Miro 1.1  2
Miro 1.2  1

What this table shows is that almost all development is being done by PCF. This table troubles me the most--more about that at the end. On to stats from Bugzilla.... First off, our Bugzilla data before October is probably mediocre, so I'm not really even looking at that. After that, the data has been getting better as more people are helping to triage and annotate bugs. Also, some bugs never make it to Bugzilla. I know that sedatg and some other people mention issues to us on IRC semi-regularly which get fixed, but aren't tied back to Bugzilla bugs. It's probably fair to say these stats are indicative of things but aren't 100% accurate.

Miro 1.2 stats
length of cycle:      70 days
bugs fixed:           82 total
  By Operating System:
     all:             26
     gtkx11:          14
     osx:             13
     win:             29

  By Severity:
     blocker:          1
     critical:        12
     major:            5
     normal:          58
     minor:            2
     enhancement:      4

  By Component:
     Channels         11
     Download          4
     Feeds             1
     Guides            3
     Install           5
     Library - New     3
     Menu - Shortcut   3
     Min - Max         1
     Playback         14
     Playlists         2
     Search            6
     Startup          10
     Storage           1
     System settings   2
     User interface    5
     main             11

bug reporters:        24 total
     pcf people:       7
     community:       17

Miro is benefiting greatly from the community with testing and translations--that's really great and it's helping a ton! However, Miro is not getting much help from the community with code and PCF is pretty much funding all development. This is troubling. Miro is getting bigger over time and the complexity is growing, too. There are a lot of moving pieces in the stack of external components that Miro relies upon. There are two ways for Miro development to scale well:

  1. more contributors

  2. additional funding for PCF so that they can fund developers

If you can contribute code, please let me know if there's something blocking your path. If you can't contribute code and/or you're interested in Miro getting better, then install iHeartMiro (there are versions for Firefox and IE) and/or donate money and help PCF fund developers.

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.