January 3, 2007

Dunc-Tank continues to make splash

Author: Bruce Byfield

Dunc-Tank, the unofficial organization set up to fund selected Debian activities, has made its first experimental payment to release managers, but community members are still debating whether a missed deadline and the alleged demotivation of some programmers make the experiment a failure.

As reported previously, Dunc-Tank is managed by leading Debian developers, including Anthony Towns, the Debian Leader, and Steve McIntyre. The organization has been controversial since its announcement in September, with such concerns being raised as whether paying some developers would demotivate others, and whether the composition of Dunc-Tank would cause outsiders to believe that it was an officially sanctioned project. The group was satirized by the Dunc-Bank site, and its advent even resulted in a general resolution to impeach Towns for conflict of interest.

The resolution failed, and, although the controversy continued, Dunc-Tank was able to pay release managers Steve Langasek and Andreas Barth for one month each in the hopes of Etch, the next Debian release, being ready for December 4. However, with the general freeze on Debian packages -- the last major milestone before the release -- occurring only on December 11, the release seems unlikely to occur before the end of January 2007.

A missed deadline

Although Debian has a history of missed release deadlines, opponents have been quick to claim that this instance proves that Dunc-Tank is a failure. "The experiment failed to do the release," M. J. Ray noted on the debian-project mailing list.

However, others are less certain. Barth, who blogged regularly about his month of full-time work on the release, says, "The Dunc-Tank experiment has positive and negative effects, and we shouldn't watch only one side -- whichever side that is." Barth notes that funding allowed him to focus on fixing a couple of bugs per day, instead of taking several days for a single bug fixed in his spare time, but adds that this focus "took away most of my time for other things I should also have worked on." He professes that "I am happy with my own involvement, but that doesn't necessarily apply to the full experiment."

Others suggest that conclusions are hasty. "I'm still not personally ready to draw any final conclusions yet," Towns told NewsForge. However, on debian-project, he specifically denies that releasing Etch on schedule was "even put forward as a metric for the experiment." Technically, he is right, since Dunc-Tank, while suggesting that its actions could help Debian make a more timely release, never committed to a specific date, but that does not stop Ray from describing the claim as "the politicking of a [Debian Project Leader] trying to hide the negatives of his decision."

Are developers demotivated?

The post-mortem on Dunc-Tank is further complicated by a claim in an article by Steven J. Vaughan-Nichols that attempts to meet the deadline were sabotaged by demotivated or disgruntled Debian developers doing less work than they might have because of their opposition.

This claim apparently originates in Barth's comment on his blog that "There was a large disadvantage of the whole experiment: Some people who used to do good work reduced their involvement drastically. There was nothing I could do about it and that happened way before I started full-time on the release, but on the global picture that still counts."

In an update to this entry, Barth implies that Vaughan-Nichols' claim "doesn't reflect what I said" and attributes the delay to the usual outstanding critical bugs.

While a few Debian project members, such as Barth and Wouter Verhelst, have attempted an impartial analysis of Dunc-Tank, many have been quick to take sides. Some, like Pierre Habouzit, claim, "There was no conspiracy, undermining work, or power games. People claiming otherwise are making complete fools of themselves." By contrast, others, like Joey Schultze, claim that it was "clear from before the 'experiment' was started that several people will be demotivated it. Several developers have also left the project for various reasons around [Dunc-Tank]."

What Debian members believe seems motivated largely by their position on Dunc-Tank. "People don't appreciate if one tries to summarise openly what happened and look both ways," says Barth. "Is it so wrong to look in both directions, and not only say 'that and that' were the improvements? I thought at least within Debian we don't want to fall into the usual management mistake of only speaking about how great everything is, but be honest with ourselves."

Reports of death are greatly exaggerated

Exactly how serious the controversy over Dunc-Tank has become is hard to determine. On the one hand, Debian has always conducted its business publicly, and with a heightened rhetoric that outsiders are apt to regard more seriously than those who are actually involved. On the other hand, Debian has a history of inflamed controversies, such as those surrounding the departure of Bruce Perens, so seeing Dunc-Tank as yet another such incident is hardly a major stretch.

Yet even if Dunc-Tank has caused divisions, the long-term effects are likely to be slight -- no matter who carries their point or leaves the project. With more than 2,000 members in the community, Debian has long outgrown any dependency on any individual contributor, or even any group of contributors. Whatever the outcome of the Dunc-Tank controversy, the Debian Project will likely stumble on as it has in the past, with its members contemptuous of deadlines and arguing fiercely at every step. There are many members who wouldn't have it any other way.

Bruce Byfield is a computer journalist who writes regularly for NewsForge, Linux.com, and IT Manager's Journal.


  • Linux
Click Here!