Posts from February 2011.

The Kerfuffle

Today Jono, as well as Gabriel, announced that the new deal for the Banshee developers regarding the referral money would be a 75/25 split. To show good faith, Canonical will also be donating 25 percent of the proceeds from the UbuntuOne Music Store. As could be predicted, there was a lot of commentary about the relative merits of this outcome. Here are my thoughts on those comments.

A choice withdrawn
First, Gabriel’s blog post. As you can read from the language Gabriel uses, this was a one-sided decision where Canonical withdrew their previous offer. Reminder: that was where they gave the Banshee maintainers a choice between on by default and a 75/25 split and a not on by default and 100% GNOME Foundation. The maintainers choose the second. Some may say “but Canonical is now offering 25% of U1 Music Store revenue, which is more than they offered before.” Yes, that is true, but the Banshee maintainers specifically did not choose a revenue sharing deal. To then respond with a statement that their previous offer was a “mistake” and that this new plan devised by Canonical is the way they are going to go, nevermind the decision previously made, is bad form.

Money supports FLOSS development
This next part surprised me. And I have to preface this with that I respect [openly ;)] the work that Jono does for the Ubuntu community. The Ubuntu community is lucky it has such a devoted person working on their behalf. I’ve always been impressed with the quality of team leadership/management I have seen from Jono. But, his comments on Twitter following this announcement did not reflect the Jono I know.

Jef Spaleta asked if Jono could see the difference between legal and ethical (we all know it is permitted by Banshee’s license to change the source code, and thus the referral code, but ethics are separate from legalness). Jono responded with “my view is that it is all going to FLOSS, so it is fine.” Sandy Armstrong correctly pointed out that Canonical does not only do FLOSS development. They also produce Landscape and UbuntuOne, where the bulk of the code (my guess), the server/web interface portion, is proprietary. Jono ignored his previous statement and responded with “that’s why I said ‘most.’” Sandy correctly pointed out that he indeed said “all.”

Yes, that whole previous paragraph of “he said/she said” was only to point out, in exruciating detail, that Jono slipped up on his wording in a tweet – something I’m sure none of us have ever done. But this is a very delicate matter and should be handle accordingly.

So, it was especially disappointing to see that slip-up followed by another sloppy post: If Canonical put 100% of the money into non-Free software then he’d see the problem, suggesting that anything less that 100% is just fine.

Then his last comment (on Twitter) about the issue for the day suggested that if the Banshee maintainers didn’t want their referral code to be changed they should have licensed their app that way. This is either a suggestion to create a non-Free software license or a statement that “as long as it is legal it should be OK with everyone.” The first is just ridiculous. And the second exemplifies the world where ethics do not enter a decision making process. Something can be both legal and unethical.

Transparency
(I’m making this a separate section because I want to bring attention to the idea, and I hope it happens.) One of the suggestions by Jef was to open up the financials of Canonical to prove where the money is going. I’m a huge fan in open book financials for companies I support. The Wikimedia Foundation does this, along with Creative Commons (PDF). They are non-profits in the US and make their 990 tax fillings public. Bradley Kuhn has even started a gitorious project collection those fillings for various FLOSS non-profits.

Aaron Bockover, another Banshee maintainer, wrote an open letter to Canonical suggesting another route. Basically, let the banshee.fm server handle the revenue sharing/affiliate code managing. This not only allows a more transparent audit trail but also has many technical benefits. I’ll let you read them on his post. They are very sound and well thought out.

The decision process
This kerfuffle brings one very important part of the Ubuntu project to the fore: the governance and decision making process. Reminder, the decision to switch the default media player from Rhythmbox to Banshee was a community decision at an Ubuntu Developer Summit. This was a community-led decision. This is in stark contrast to some other decisions in the Ubuntu project that are not made by the community, an example of which is, of course, this revenue sharing decision, and another is the changes made by the design team, Ayatana.

Because of today’s developments, I decided to relisten to this interview with Phillip Schmidt (of P2PU) and Mako on my walk home from work. The interview was about, generally, governance in Free and Open projects. Obviously, much of the interivew focuses on Debian and Ubuntu governance. I HIGHLY recommend listening to that interview if you are at all interested in the governance structue of Ubuntu, Debian, Wikipedia, etc. Mako’s insights into those communities, especially Ubuntu’s, is greatly worthwhile. One thing of note is the way decisions are made regarding Ubuntu. Mako brings up the point about recent changes made by the Ayatana team and how they are presented to both the community (including the Community Council) as done deals with recommendations to “keep people on message” if they disagree with the changes.

I wanted to bring this governance aspect up because I think it really gets to the core of the issue for me. This decision was made by people not accountable to the community. The relative merits of an organization running a FLOSS project that has people and decisions that are not accountable to the community is up for debate, but not here. I’m just noting it and saying that it is something to be extremely aware of. The conclusion you make about its morality is your own. I might write more on it later, but this post is already getting to be too long.

My Opinion
My opinion is simple: This decision was made in a very bad way. No community involvement without any course for a real dialog. It disappointed me.

It reminds me, all too clearly, of the decision to name the file sync cloud service UbuntuOne instead of CanonicalOne (or similar). Some of you might remember the intense debate that happened on the topic. I watched the Community Council meeting where Mako and others talked with Mark about not only the name but also the issue that UbuntuOne was, and is, not a Free Network Service. I remember the conversation not being a true discussion with open minds. The decision had been made by Canonical people (including sabdfl) and no amount of discourse from the Council could change that. It disappointed me.

In addition, the communication of this decision was made poorly. Things could have been said better on Twitter. Things could have been more openly discussed on mailing lists. But none of those open channels of communication were used effectilvely (if at all). That disappointed me.

So I guess that’s my opinion of the matter: I’m disappointed.

This post has agonized me the entire time I wrote it. I don’t like these posts, but I feel they need to be written from time to time. I don’t mean to offend anyone, I only mean to share my thoughts on an important issue. I could have written more, because this kerfuffle brought up some problems that will need to be solved to prevent future kerfuffles. I didn’t write more because those issues deserve their own post and it is getting late.

Searching your email

As many of you may know, I’m really quite facinated by my email system. By system I mean the suite of tools that bring email sent to me to my eyes. I’ve written before on my setup (note: that is out of date now) and its migration across systems. As you can see, I’m a fan of Mutt.

I do have a confession. For the last few months (since mid-December ’10) I was using Thunderbird 3.3a2. Why? Why throw away all that time and effort setting up Mutt/Offlineimap/Imapfilter to be as effecient as possible and just go with TBird? Search across accounts.

Yes, search.

Ever since I started using Mutt I found myself sometimes popping into my gmail account to search for a specific message (if it was sent to that account, of course. My personal account is not hosted by gmail, just my work account). I felt bad about it each time. “Why can’t I just do this in Mutt?” I would ask myself. I tried various email searching systems but none of the quite worked exactly how it should/I wanted.

Then came Notmuch.

Notmuch promised to be exactly what I wanted. A text-based, threaded, and search-oriented mail reader. Perfect! Except, due to some limitations at the beginning (basically, you can’t move messages without making Notmuch angry) I never could use it full time.

But the Notmuch developers fixed that! I didn’t know that though. I missed that announcement. Where was it? Why wasn’t it being shouted on the roof tops? NOTMUCH CAN NOW SYNC WITH IMAP CHANGES!

Oh well.

This whole time I’ve been using Thunderbird (mostly) without issue and loving the ability I had to search across all 3 of my email accounts which something that even GMail, the master of search, could never do for me.

So what brought me back to Mutt from Thunderbird 3.3? Search. You see, just like Lucas I saw Zack post about how a long flight gave him the time to integrate Notmuch and Mutt.

Now, thanks to notmuch, I can do a quick search across all 3 of my mail accounts and even recreate threads from just a single message. Simply amazing.

Do you like Mutt and wish you had top-notch search? Give mutt+notmuch a try!