OSS communities does and don't
Picked this video from Slashdot coverage. http://video.google.nl/videoplay?docid=-4216011961522818645. This was a talk done by 2 Subversion developers. They explain how to manage OSS and get rid of annoying people in your OSS project.
Having run in this situation many many times I was curious to hear what they had to say. I think they cover the "volunteer OSS" angle very well. I found myself smilling at many of the examples they gave as we found many similar examples at JBoss.org.
I thought, however, that their coverage lacked the dimension of"Professional" OSS (POSS)management. Some things that seemed to puzzle these guys were dealt in a straightforward manner in POSS.
They talk about the irrelevance of some discussions of details. When the topic was mostly a matter of taste (emacs vs vi, windows vs linux, red vs blue!) people would quickly degenerate in religious debate because there was no right answer for it. These discussions can suck a lot of time and are, by definition, without end. It is better to ignore them early on. Having a business focus in your community was always helpful in that way.
The importance of compensation: the guys in the video seem very "non-profit" in how they manage the community, they mention that the subversion community was all volunteers. Their experience is very limited when it comes to managing a professional community where money is involved. They focus more on how to "get rid" of people rather than "attracting" them. I found that to be a little arrogant given their "non-profit" nature. I know better about the dynamics of OSS than to pretend people bang on your door wanting to give YOU their work for FREE! The couple of cases where you have to rid of someone are a luxury.
At JBoss, there was always a money dimension, first you would do training and then consulting and then become part of the equity. That, I found, was the biggest thing in managing your community. It is a big part of how things get done,why they get done. We got shit done because it was our job. It is a big part of the problem too. At JBoss most the "annoying people" came from "equity discontent". Money changes everything and specifically the dynamics of a previously non-profit group. Some things are streamlined, others more complicated. The subversion guys don't seem to have hit that point yet?
The guys have hit on a holy grail however: if you want to contribute then shut up and code. You can argue all you want: he who codes wins. They mention this in passing but it really was a corner stone for me personally. People have little time to actually review, code but will spend time writing their opinions on mailing lists! In other words, if you didn''t code, we would not pay attention to you. We just couldn't afford it.
They seem to take pride in their "politeness" only to describe minutes later a very scheeming, political and hypocritical way of getting rid of a guy who would suck too much of their time by asking too many questions as he tried to help. This was a weak point in what they presented imho: with 40 or so projects at JBoss.org we were upfront about the "you are on your own" factor.
They mention in passing the problem of "annoying users". We called this the "suck my D888" moment (or SMD moment) at JBoss. Everyone in a succesful project goes through the SMD moment when a pestering user keeps asking, in the most annoying possible way, for HELP RIGHT NOW AND FOR FREE! Dealing with these guys had its own policy at JBoss.
See, OSS is free, at least the software is and lots of people generalize this to the people, the software developers themselves, as well, as in "free help". There is a lot of free help being dished out on the forums for sure but drawing the line between "getting you going" and "doing your work for you for free" is always tricky.
Even sales had their SMD moment. One of our guys in partner sales had summarized it this way "when I read the word synergy in a partner email, I know they are trying to get a steep discount on our partner programs, so when I read the words "extreme synergies" in a email this morning, I knew it was time to run"
Since they didn't cover the SMD moment, am I to understand that their project is not that widely used? In any case I recommend listening to that presentation. It is good even though they miss the whole POSS angle.
Having run in this situation many many times I was curious to hear what they had to say. I think they cover the "volunteer OSS" angle very well. I found myself smilling at many of the examples they gave as we found many similar examples at JBoss.org.
I thought, however, that their coverage lacked the dimension of"Professional" OSS (POSS)management. Some things that seemed to puzzle these guys were dealt in a straightforward manner in POSS.
They talk about the irrelevance of some discussions of details. When the topic was mostly a matter of taste (emacs vs vi, windows vs linux, red vs blue!) people would quickly degenerate in religious debate because there was no right answer for it. These discussions can suck a lot of time and are, by definition, without end. It is better to ignore them early on. Having a business focus in your community was always helpful in that way.
The importance of compensation: the guys in the video seem very "non-profit" in how they manage the community, they mention that the subversion community was all volunteers. Their experience is very limited when it comes to managing a professional community where money is involved. They focus more on how to "get rid" of people rather than "attracting" them. I found that to be a little arrogant given their "non-profit" nature. I know better about the dynamics of OSS than to pretend people bang on your door wanting to give YOU their work for FREE! The couple of cases where you have to rid of someone are a luxury.
At JBoss, there was always a money dimension, first you would do training and then consulting and then become part of the equity. That, I found, was the biggest thing in managing your community. It is a big part of how things get done,why they get done. We got shit done because it was our job. It is a big part of the problem too. At JBoss most the "annoying people" came from "equity discontent". Money changes everything and specifically the dynamics of a previously non-profit group. Some things are streamlined, others more complicated. The subversion guys don't seem to have hit that point yet?
The guys have hit on a holy grail however: if you want to contribute then shut up and code. You can argue all you want: he who codes wins. They mention this in passing but it really was a corner stone for me personally. People have little time to actually review, code but will spend time writing their opinions on mailing lists! In other words, if you didn''t code, we would not pay attention to you. We just couldn't afford it.
They seem to take pride in their "politeness" only to describe minutes later a very scheeming, political and hypocritical way of getting rid of a guy who would suck too much of their time by asking too many questions as he tried to help. This was a weak point in what they presented imho: with 40 or so projects at JBoss.org we were upfront about the "you are on your own" factor.
They mention in passing the problem of "annoying users". We called this the "suck my D888" moment (or SMD moment) at JBoss. Everyone in a succesful project goes through the SMD moment when a pestering user keeps asking, in the most annoying possible way, for HELP RIGHT NOW AND FOR FREE! Dealing with these guys had its own policy at JBoss.
See, OSS is free, at least the software is and lots of people generalize this to the people, the software developers themselves, as well, as in "free help". There is a lot of free help being dished out on the forums for sure but drawing the line between "getting you going" and "doing your work for you for free" is always tricky.
Even sales had their SMD moment. One of our guys in partner sales had summarized it this way "when I read the word synergy in a partner email, I know they are trying to get a steep discount on our partner programs, so when I read the words "extreme synergies" in a email this morning, I knew it was time to run"
Since they didn't cover the SMD moment, am I to understand that their project is not that widely used? In any case I recommend listening to that presentation. It is good even though they miss the whole POSS angle.
Comments
Great post, and you are right about how money changes everything (i.e., equity discontent). I hadn't seen this video before you posted it, but in Fogel's book, he does cover the money aspect. However, what is left out, is the real impact of money on the social dynamics of open source projects. Sure, getting paid to do open source is encouraged, but there's more to be said once you have a dramatic impact on the world--money quickly becomes very important. To be fair, JBoss fell under the Service Augmentation category Fogel mentions, but at the same time, JBoss (the corporation) was in no way augmenting so much as it was the embodiment.
The Subversion guys (while having a successful open source project) seemed to handle things differently. I have to give them a bit of credit on their politeness. They are indeed polite. Money was involved, to some degree, but not the millions of dollars to which JBoss was exposed.
I'm still looking for Money Patterns in open source projects (i.e., drop a shitload of cash on a project and see which developer is murdered first). I'm seeing conflicting outcomes, and I can't put my finger on any specific emergent behavioral pattern.
Anyway. Interesting post, and I hope you are doing well.
Cheers,
Damon
Yes JBoss the corporation was in many way a reflection of the community. The community professionalised and the equity was distributed.
The service as a separate entity model is both an avoidance of the business angle and an unfair development if succesful.
As you point out, succesful communities have to deal with that angle at some point or another.
Take care,
I miss some of the least relevant discussions on core the most ;-)
-Andy
I think you're only 25% right on why JBoss attracts less committers than SVN. What it boils down to is that recruiting takes tremendous effort. As a project lead, if you have minions that already work for you, you can basically ignore recruitment. I know the few times I focused on recruitment, I was able to attract people.
Bill
That myth is still enduring. I was over that point in 1999.
Middleware takes special talent, special talent requires pay.