Retiring as a Debian developer

This is a repost and update of my retirement letter sent privately to Debian last month, July 10, 2016. At that time I received many notes of appreciation and good wishes which I treasure. Now, I’d like to say goodbye to the broader Debian community and, as well, indicate which of the cleanup items have since been addressed in strikethrough style and with annotations. Also, I’d like to stay in touch with many of you, so I have added some comments oriented towards those of you who are interested in doing that after the letter.

When in 1995, on a tip from a friend, I installed Debian on my 386 at work and was enthralled with the results, I could not have foreseen that two years later, friends I had made on channel #debian would nudge me to become a Debian developer. Nor when that happened did I have any idea that twenty years later, I’d consider Debian to be like family, the greatest free software community in the world, and would still be promoting it and helping people with it whenever I could. Debian quietly, unexpectedly became a part of what defines me.

My priorities in life have changed over that time, though. I have shifted my attention to things that are more important to me in life, such as my family, my health and well-being physically and spiritually, and bringing all I can to bear on the task of preserving our local wilderness areas and trails. In the latter area, I’m now bringing all of what Debian has helped shaped me to be to the table, launching some ambitious projects I hope will bear fruit in the coming years, and make a measurable contribution to help us hang onto our precious natural preserves where I live.

Unfortunately, as I’ve poured more time and energy into these things, I’ve increasingly not been giving my packages the care they need. Nor do I have any roles or goals now for any of the Debian projects I was previously involved in. So, after much careful deliberation, and as much as it pains me to say it, it’s time to retire as a Debian developer. It has been a great privilege to work with you, and to meet many of you in New York at Debconf 10. I plan to be around online, and will continue to take an interest in Debian, lending a hand when I can. Thanks for all of the fun times, for all that I’ve learned, and for the privilege to make awesome things with you. I’ll treasure this forever.

So much for the soppy bits. 🙂 Now, business. These things remain to clean up upon my departure, and I’d appreciate help from QA, and anyone else who can lend a hand. My packages are effectively orphaned, but I haven’t the time to do any of the cleanup myself, so please speak up if you can help.

  1. Debian Jr.
    • O: junior-doc. The junior-doc package has been awaiting an overhaul by whoever revives the project since I gave it up years ago. I’m still listed as maintainer and that should be changed to Debian Junior Maintainers <> if they want it. Otherwise, it is orphaned.
    • I should also be dropped from Uploaders from debian-junior, the metapackages source. Fixed in git.
  2. Tux Paint. This is a very special package that deserves to go to someone who will love it and care for it well. There are three source packages in all:
    • O: tuxpaint
    • O: tuxpaint-config
    • O: tuxpaint-stamps
  3. O: xletters. This is a cute little typing practice game and needs a new maintainer.
  4. XPilot is co-maintained by Phil Brooke <>, so he should replace me as Maintainer. Phil said he’ll pick up xpilot-ng and will also look at xpilot-extra.
    • xpilot-ng
    • O: xpilot-extra (recently removed from testing due to my neglect, and not co-maintained by Phil; it’s unclear if anyone really uses this anymore)
  5. GTypist is co-maintained by Daniel Leidert <> and should replace me as Maintainer.
  6. My ruby packages. A group of packages that I brought into Debian as dependencies of taskwarrior-web, which I never completed. Maybe they’ll be useful in and of themselves, and maybe not. In any case, they are maintained by pkg-ruby-extras-maintainers, but I’m the sole developer in Uploaders and should be removed: Fixed in git.
    • ruby-blockenspiel
    • ruby-parseconfig
    • ruby-rack-flash3
    • ruby-simple-navigation
    • ruby-sinatra-simple-navigation
    • ruby-term-ansicolor
    • ruby-versionomy
  7. Debian Live stuff: I am listed in Uploaders for live-manual (fixed in git) and debian-installer-launcher (fixed in git) and need to be removed.
  8. O: eeepc-acpi-scripts. The defunct Debian EeePC project has just this one package. Recently, the mailing list was asked about its status, and it was recently NMU’d. To my knowledge, nobody from the original team remains to take care of it, so it needs a new maintainer. I should be removed from Uploaders, and since the Debian Eee PC Team no longer exists, it should be removed as maintainer. It is effectively orphaned unless someone speaks up.

There are also some Alioth projects / lists that are defunct that I’ll need to talk to the Alioth admins about cleaning up in the coming days. One of these is <> and since it is still listed as the maintainer of eeepc-acpi-scripts, that needs to be sorted out before the list can be closed.

Thanks again, and see you around!

Stay in touch

For those of you who would like to stay in touch, here are some ways to do that:

  • Follow my blog:
    If you already do that, great! If not, welcome to my blog! For the past couple of years you may have noticed a decrease in technical content and increase in local trails and conservation oriented posts. You can expect more of the latter.
  • Say hi to me on irc: SynrG (also SynrGy) on ( or
    I still intend to hang out and offer support when I can, just no longer as a developer. Channel #debian-offtopic on either network is a good place to catch up with me socially.
  • Follow me on Facebook:
    For better or worse, a lot of the trails and conservation folks hang out here, and many of you in the Debian community are already my Facebook friends.
  • Look for my Bluff Trail posts on their site:
    Providing tech support to this organization is where much of my time and energy is going these days. I post here once in a while, but do most of my work behind the scenes as a volunteer and, newly this year, as a board member.

Taskwarrior new blog, getting involved

The Taskwarrior Team has just started a blog, kicking it off with a series of articles about development of Taskwarrior itself. Go, team!

Almost from the moment I started using Taskwarrior (thanks to Jakub WIlk for an excellent job maintaining this) I knew I had finally found the todo system that I could love. Right away, I started hanging out with the Taskwarrior community at #taskwarrior @ and found out what an awesome bunch of people they are, both developers and users alike. I have plunged in with bug reports and feature requests, and am helping get more Taskwarrior-related things into Debian. In NEW right now I have uploaded several of the dependencies needed by my ITP of taskwarrior-web. It’s looking like I’ll have that finished later this month or early next month. Also, I have uploaded a wheezy backport of Taskwarrior itself (aka ‘task’) which, if all goes well, enters the archive next week.


Collaborative editing, the missing Vim feature (pentadactyl + etherpad?)

Do you wish, like I do, you could edit collaboratively in Vim? This feature is number 10 on the Vim voting page, so it seems I’m not alone. How about Pentadactyl coupled with any of the existing web-based collaborative editors, such as Etherpad? OK, so it’s not quite Vim, and there are some rough edges to this particular pairing, but I’m finding it’s good enough for my needs. It even gives me a Vim-like editing experience while other participants use the default Etherpad editor.

Yes, I know about, but for the past three years I have been using a single instance of Etherpad with my family to maintain a shopping list to which we all make contributions. First of all, that’s not a Debian activity, so to make the switch, I’d need to make a personal clone of the service for our personal use. But more importantly, we find Etherpad features such as colours for different participants and the timeline are just too useful to give up on. On the other hand, the less the web editor interferes with your web browser’s default textarea behaviour, the easier time Pentadactyl is going to have. Indeed, I asked on #pentadactyl @ about some problems I was having and I was told flat out that Pentadactyl does not work with graphical web editors. So, you may wish to use another web-based collaborative editor for this reason. That being said, I did learn a few things about helping Pentadactyl get along better with Etherpad, so if you would like to try it yourself, read on.

The key to getting started was to enter ‘text edit mode’ within the textarea with <C-t>. For the most part, this behaves like Vim ‘normal mode’. I am still learning, but many basic motion and editing keys behave just as they would in Vim. Fantastic!

However, the moment I tried to :undo I hit my first problem. Using the latest release version of Pentadactyl (1.0rc1 at time of writing), pressing "u" to :undo produced no visible result. I tried the latest daily build as well, and only saw a marginally more helpful "Node not found" error message displayed in the status area. But it turns out you can use ‘passthrough mode’ to use the textarea’s own undo. Just :tmap u <C-v><C-z> and we’re back in business again.

I’m still experimenting with this setup, so the jury’s still out on whether I’ll stick with it, or whether the remaining incompatibilities between Pentadactyl and Etherpad will drive me nuts. But it looks promising. Clearly, judicious use of :autocmd to always start in ‘text edit mode’ and bind that undo key whenever I enter the site will help make the experience even better. If you try it out yourself, I’d love to hear how things went for you. Or if you have found an even better solution that works for you, do share.


Using for project management

From the start of the DebianEeePC project, we have used a combination of the irc channel and our wiki site to our great advantage. The DebianEeePC/HowTo/Install document has been actively maintained as we go, serving not only as a good reference for new users, but also for existing users and developers to see where improvements are being made and what is left to work out.

Recently, we’ve started using the wiki for task management as well. Patterned after DebianInstaller/Bugs, we now have DebianEeePC/Bugs where we link to reports on the BTS for packages we maintain and our own usertags on packages we don’t maintain. As well, we are keeping a DebianEeePC/Todo to manage our tasks.

I see from a search through the wiki that lots of other Debian subprojects keep Todo lists, but most of them have little more structure than point-form lists. A few, though, have structured these pages a bit more carefully. For instance, some of these use tables listing who each task is assigned to, when it was assigned, and when it is completed. Our point-form list is a bit more free-form regarding task assignments. Participants can register their interest in a given task and what they are doing as a subpoint of the task. As for state changes over time, we rely on the revision history on “info” page for people to be able to see state changes for particular Todo items. When something is done, it just gets dropped from the list.

I haven’t exhaustively reviewed everyone’s Todo page, but from my brief survey I haven’t seen a couple of things I’ve added to try to help people choose what to do next. The first is to tag items with six difficulty/importance pairs, rating each actionable item from Easy/Urgent to Hard/Wishlist. The second is to state dependencies between items using anchors. It would be nice to see any of you are doing something similar and hear about whether it has worked out well for you.

I’m happy with how the irc/wiki combination has helped with project management. In the past with a different project I’ve made an abortive attempt to use the task management system at Alioth, but it just seemed too cumbersome and nobody ended up using it. With the wiki, it appears barriers to participation in task management are sufficiently lowered that everyone is willing to help keep it up. And with irc, we can keep on top of defining and assigning (or more usually volunteering to do) various tasks, and making sure the wiki is kept completely up to date. It takes a bit of discipline to do, but after the habits are established, it gets easier, and the rewards are well worth it.