Platform for Wouter Verhelst

Why am I running for DPL?

Hi. I'm Wouter Verhelst, and I'm running for DPL again. Yes, again, since I first ran back in 2007, after Anthony's term. Not much has changed for me personally since then. I moved—I now live in Mechelen—and I got a few years older, but that's about it.

The same cannot be said about Debian, however. In three years, much has changed. There has been a gradual replacement of people in role positions in the project; people like James Troup, Ryan Murray, Manoj Srivastava, and others, have stepped down from the position(s) they have held, and handed their responsibilities over to others. Sometimes this went smoothly, sometimes not so smoothly. At any rate, I feel it is safe to say that the new blood in role positions is doing a reasonable, if not good or very good, job. It has been a long time since I last heard people complain about the lack of communication from these teams; though I should also add that I personally do not perceive an important difference between communication as it happened then, and as it happens now[0].

Another old pain in the Debian community has been the time it took us between releases. That, too, has been dealt with; our previous "lenny" release took us far less time than some of the more problematic ones in the past, and it looks like squeeze, while it may be somewhat delayed from the initial planning, is not going to be delayed forever, and is getting closer and closer.

So Debian's doing well, right?

Well, not necessarily.

Now that the old problems which have plagued us so long seem to have been dealt with, it feels to me that people have lost focus somewhat. This very DPL election is a very good example of what I mean: after the record high of eight candidates in 2007's DPL election, 2008 had only three. Last year's election had only two candidates; and this election got very close to having only one candidate[1]. This, I think, is a symptom of a loss of interest and a loss of focus in the larger Debian picture, and it is well past time we did something about that.

Background

In 2007, I ran on a platform that was intentionally somewhat vague. This seems to have confused some people into thinking that I was indecisive. I was not; instead, I wished to outline what I thought the correct course of action was, without committing myself too much to specific plans which in the end might turn out to be impractical for one reason or another. Since I wish to avoid the same confusion this time around, I will provide some more details of my plans this year. However, do note that they are not set in stone; if it turns out that some specific details of my plans cannot be executed properly, I will not hesitate to change them. If you know something I don't know and feel that parts or details of my plan are impossible to execute, do not hesitate to let me know. The broad lines of what I wish to accomplish will, however, not change.

First of all, I plan the usual things: I intend to communicate as much as possible, and I intend to fund developer meetings wherever possible, etc. Pardon me for the lack of emphasis on those points, but for as long as I've been a debian developer, DPL candidates have promised to be good at communicating. Yet, many if not all have failed at doing so, and while I will do my best to be at least as good as the average DPL in past years, I hold no hopes of being the best DPL we've had in this area. As for the funding of developer meetings, it feels to me that that's about the only thing we can reasonably do with our money that has little chance of offending anyone, so I see it almost as a no-brainer to do this, where and when it makes sense[2]. Furthermore, I honestly don't think that Debian's future hinges on whether the DPL communicates, or what our money is spent on (though it does obviously have some influence).

What I do think it hinges on is whether we continue to receive outstanding technical contributions. Debian has long been[3] the best volunteer-developed distribution out there[4], which meant that most people interested in distribution development, but who were not prepared to work for a commercial distribution (for whateve reason), were attracted to Debian. This is no longer true; not only are there other significant volunteer-only distributions out there today (Gentoo springs to mind), it is also true that most significant commercial distributions now have (mostly) community-developed distributions. There's RedHat's Fedora, Novell's OpenSUSE; and Ubuntu, too, has a community (the MOTUs) who do a lot of development that goes directly into their distribution.

We could say that it is not important or relevant to us what happens with other distribution communities, and to some extent that is true. We could say that, on a technical level, Debian is still the best distribution out there, and I certainly agree with that statement, though I know plenty of people who do not.

Be that as it may, however, technical excellence does not create itself, and if we wish to continue being technically excellent, we need to continue attracting developers; I believe this is where Debian's main challenges lie in the years ahead.

How should we move forward?

Debian is already having troubles due to the fact that there is more competition out there for distribution development. Many important packaging teams, including the xorg and libc ones, are or have been badly short of manpower. There seem to be issues with the packaging of the python interpreter, too[5]. Of course these are just examples; there are more similar problems—some of which, I'm sure, I'm not even aware of.

The only way to solve these problems is to get more people to participate to Debian, and to keep motivating them so they keep working on Debian where and when they can. The best way to increase that influx of interested contributors is to get more people interested in our distribution. And about the only way to make that happen is to figure out what our problems are, as perceived by not only ourselves, but also those who are not part of our community, and to work on them. Thus, I intend to talk to many people, both within the Debian community and outside of it, to find out how we could improve. I don't expect that either I by myself, or the Debian community as a whole, will be able to find a resolution for each and every one of the problem areas that we will identify in this manner; but I do expect that, by asking people to critically look at our distribution, we will be able to identify blind spots of things that we do not have (and that we are not aware of missing), or generic areas that a lot of people are struggling with, and that we will be able to prioritize what to work on—all on a voluntary basis, of course.

So that's how to get people interested into Debian. As for how to keep contributors: one of the more known pain points that still exist within the larger Debian community, is that Debian isn't as often as nice a place to be in as we could make it. While our mailing lists aren't the same flaming hell that they used to be some years ago, we are still a long way from what we could be[6]. This is something we need to work on as a community, not only by stiving to flames whereever and whenever possible, but also by politely challenging people who do flame; and if elected, I intend to do just that[7].

I believe these two steps will be important steps forward to improve Debian as a whole, and I invite all of you to work with me towards that goal.

Thank you for listening reading.

Rebuttals

Stefano Zacchiroli

Stefano's platform is, er, long. It looks like he's been thinking a lot about running for DPL, and has done a lot of preparation. He has set himself some rather ambitious goals; if he manages to achieve them all, that would be quite an achievement.

Which is good, of course.

But I'm not convinced that will be easy; the DPL job is a demanding one. Of course, I wish him all the luck.

Wouter Verhelst

Looks like this guy was a bit tired when he wrote his platform. It's full of grammar and spelling errors.

Hang on, that'd be me.

Why, yes, I was.

Charles Plessy

Charles seems to identify some of the same problems in his platform that I do in mine. But what he proposes to do in an effort to work on these problems is vastly different.

Since I have recently (as in, less than a week ago) resigned as a frontdesk member, I have a good understanding of how our current NM process works. While the DM process seems to be working well, it is certainly not true that all DDs are equal, and/or that all DDs can be trusted to make a good judgement with regards to other people's qualities.

This does not necessarily imply malice or incompetence; it simply is a fact of life that some people are better judges of character than others, and that whenever you have people who've been part of an organization longer than others, there will be people with more experience than others.

As such, I'm not convinced that handing DD status to someone based on the statement of a single other DD is the best way forward. If the average time it takes to go through the NM process is long, that is mainly because there are so many applicants who apply (and are advocated!) before they are ready to become a DD. Since unready applicants slow down the process for everyone else, this is a bad thing, and it is therefore that the NM process has become such a complicated procedure. But frontdesk and the DAM have been happy on several occasions now to fast-track applicants who so obviously were ready to become DDs, waiving the requirements that are made in most cases, and there is no reason to think that they're about to stop doing that any time soon—on the contrary. In short, I don't think the NM process is the problem.

On the release bits, I'll be short: No, the m68k port, to which I was an active contributor, did not die because we had to build more packages than we could; instead, it died because there was insufficient upstream work being done on the port, and because we did not have the manpower to do this work ourselves. For instance, there is still no libc for m68k capable of thread-local storage, a requirement since glibc 2.3 or thereabouts. As a result, almost no modern software can be compiled on m68k anymore.

Building high numbers of packages is just a scalability issue; a problem that, worst case, can be solved by throwing money (for hardware) at it. In contrast, getting and maintaining a working architecture involves far more than just building packages; it requires people to look at bugs, and fix them. This is why attracting developers is far more important than redefining how we decide which architectures to include.

Margarita Manterola

First of all, a confession: I am deeply happy that for the very first time in Debian's history, a woman is running for DPL. Male dominance is a big problem in the Free Software world; if Marga were to earn that DPLship, that could be a very important step towards ending that dominance, and I would be quite happy, if only for that fact.

But the key word in the above phrase is earn. While Marga talks about a few things in her platform that all individually could be good ideas, I'm not convinced that these are the things Debian needs most at this point in time; what I lack in her platform is a sense of vision. Therefore, I'm not convinced that she's the best candidate to be next DPL—not yet, anyway.

Footnotes

0 Before anyone misunderstands: that wasn't meant as either criticism or praise; it's merely an observation.

1 For most of the candidacy week, Stefano Zacchiroli was the only candidate. I submitted my candidacy to avoid him running alone, which I think would reflect bad on the project as a whole; Charles' candidacy was a direct result of that. Marga submitted her candidacy rather shortly before the submission deadline.

2 Where 'makes sense' is probably defined something like "I'll not be funding every possible and impossible beer-drinking-excuse until we run out of money, but will not say no if I feel that allowing a grant would help the project by helping someone reach a meeting who would otherwise not be able to make it"

3 Not that I think it isn't anymore, mind, but the subject has become far more, er, subjective recently.

4 mostly by virtue of being the only significant volunteer-developed distribution, I guess

5 While it is no secret that I have an irrational dislike for python and want as little to do with it as possible, I do of course understand that python is an important language in the free software world, and that without it we cannot have a proper distribution. Also, while I understand there are problems with python packaging, I do not currently pretend to understand the details. The latter is a result of the former; clearly, however, if I do get elected as DPL, I will not let any of that stand in the way of helping resolve the issues, should that be necessary.

6 I do realize that we can't be a perfect happy family. We aren't a family to begin with, and there is some resentment within the community. However, I'm quite sure there is still room for improvement.

7 The alert reader will notice that I have not done this all that often in the past, though I've been known to try. This is because I feel that challenging someone on what is accepted in a community as common-if-not-entirely-wanted behaviour, is not often a very good strategy. However, if I were elected DPL with the explicit statement (as above) that I will do this kind of thing (with my prospective leader hat on), that would be something else.