Weekly Edition
Return to the Development page
| |||||||||||||
Henrik Ingo: How to grow your open source project 10x
Henrik Ingo has posted the
results of a study on project governance concluding that the key factor
distinguishing large and successful projects is the existence of a
nonprofit governing foundation. "There appears to be a glass ceiling
for single vendor projects prohibiting their growth from the Large category
upwards. To truly reach their fullest potential, open source projects are
recommended to consider the proven governance model of a non-profit
foundation around which participants collaborate."
(Log in to post comments)
Henrik Ingo: How to grow your open source project 10x Posted Dec 3, 2010 21:53 UTC (Fri) by clugstj (subscriber, #4020) [Link]
It appears to me that they have possibly committed the far to common mistake of confusing cause with effect. A large project may, possibly by necessity, need to have multi-vendor/non-profit support. This doesn't in any way mean that creating a non-profit will make your project grow.
Henrik Ingo: How to grow your open source project 10x Posted Dec 3, 2010 22:14 UTC (Fri) by martinfick (subscriber, #4455) [Link]
True. It likely takes a very big tank to raise a big fish. But growing the tank may not grow your fish.
Have you actually read the article? Posted Dec 4, 2010 0:42 UTC (Sat) by khim (subscriber, #9252) [Link] They don't say that creation of foundation will magically grow your project. Just a recent example. But if you insist on keeping large project development "in house" then you can not go beyond certain point - and this POV is showed quite convincingly. You need a very big tank to raise a big fish - but as in real life if you'll put tiny fish in huge tank you'll do more harm then good: you must reach certain size before "one vendor model" will become a problem.
Have you actually read the article? Posted Dec 4, 2010 12:36 UTC (Sat) by clugstj (subscriber, #4020) [Link]
Yes, they did:
"To truly reach their fullest potential, open source projects are recommended to consider the proven governance model of a non-profit foundation around which participants collaborate."
It says that to me.
Have you actually read the article? Posted Dec 4, 2010 12:39 UTC (Sat) by clugstj (subscriber, #4020) [Link]
And I DID read the article. It reads much more of a conclusion looking for justification than it does a research project.
This is still better then what you are doing... Posted Dec 5, 2010 0:08 UTC (Sun) by khim (subscriber, #9252) [Link] May be. But at least it's using sane logic. Where "P is necessary for Q" means "Q => P" and not the other way around. Here Q is "project reaches extra-large size" and P is "project is using the proven governance model of a non-profit foundation". They quite succicingly show that Q => P (that is: if your project can reach "extra-large size" it must be governed by non-profit foundation). Please read at least wikipedia article before you'll go around with crazy accusians. It will explain how cause and effect work is mathematics. Hint: if we are talking about logic then cause does not always come before conclusion chronologically. If you don't like the articles where effect comes chronologically before cause then don't read research papers: it's quite common case when necessary conditions are discussed.
Huh? Posted Dec 5, 2010 5:24 UTC (Sun) by clugstj (subscriber, #4020) [Link]
I don't understand why you say I'm making "crazy accusations (I'm guessing that's the word you were trying to use)". I never said anything about chronology. I merely said I thought they had taken a correlation and jumped to the conclusion that it was causality. The correlation is between the size of projects and the existence of a non-profit foundation governing it. This does not demonstrate cause and effect.
Thank you for picking the quote - it shows your agenda quite well Posted Dec 4, 2010 23:48 UTC (Sat) by khim (subscriber, #9252) [Link] To truly reach their fullest potential, open source projects are recommended to consider the proven governance model of a non-profit foundation around which participants collaborate. You have trouble with logic. What this quite is say is exactly what you are saying: for the large project to succeed ("truly reach their fullest potential") you need multi-vendor/non-profit support ("recommended to consider the proven governance model") - but it does not guarentee success.
Vertical software Posted Dec 3, 2010 22:59 UTC (Fri) by ejr (subscriber, #51652) [Link]
I would be very interested in independent results on more vertical (yet still widely used) software. In my specific case, an analysis of numerical packages like LAPACK and ScaLAPACK would be very useful. Those underlie many other analysis, simulation, and even game packages but require a huge initial investment for becoming a "developer." You need both programming skill and mathematical (vertical) skill.
I'm sure there are many other packages that are similar. Medical ones come to mind as being more newsworthy. I'm just familiar with LAPACK, so...
Conclusions invalidated by incorrect assumptions Posted Dec 4, 2010 16:35 UTC (Sat) by jejb (subscriber, #6654) [Link]
Unfortunately, the one of the input assumptions:
"All large projects have some form of formal governance, either single vendor or non-profit foundation."
Isn't true for one of the cited projects: Linux (by which the article means the linux kernel). The kernel has a governance structure which is more like shared ownership + benevolent dictator. It certainly doesn't have a non-profit foundation governing it, nor does it have a single vendor.
Conclusions invalidated by incorrect assumptions Posted Dec 4, 2010 16:58 UTC (Sat) by JoeBuck (subscriber, #2330) [Link] You are aware that there is a nonprofit foundation that signs Linus's paycheck, so he can support his family, right? The Linux Foundation doesn't govern anything, but it allows Linus to focus on the technical work unimpeded by a day job.
Conclusions invalidated by incorrect assumptions Posted Dec 4, 2010 18:16 UTC (Sat) by BrucePerens (guest, #2510) [Link] It's not at all clear that Linus really needs that foundation. Which is probably why it works and others don't work as well.
Conclusions invalidated by incorrect assumptions Posted Dec 4, 2010 18:34 UTC (Sat) by jejb (subscriber, #6654) [Link]
> You are aware that there is a nonprofit foundation that signs Linus's paycheck, so he can support his family, right?
Amazingly enough, I have something to do with that, yes.
> The Linux Foundation doesn't govern anything, but it allows Linus to focus on the technical work unimpeded by a day job.
Which is the basic point: The kernel governance model is completely independent of the LF. Who actually pays Linus is a fairly orthogonal question.
My point is that one of the input statements in the article is that the "XtraLarge" projects are "developed as collaborative community projects governed by non-profit foundations", which you appear agree with me to be untrue in the case of the kernel.
Conclusions invalidated by incorrect assumptions Posted Dec 6, 2010 9:01 UTC (Mon) by lacostej (guest, #2760) [Link]
The thing is that the Linux kernel is a very special beast.
The benevolent dictator is driven by technical beauty and compromise, not by politics (at least very little). The project direction is derived from those decisions.
Another important aspect is that the kernel isn't an end product in itself. It's a foundation. So the direction the kernel is going isn't affecting too much the individual vendors. Especially as the ecosystem allows for feature forks (every enterprise contributor company tries stuff on their kernels, and may tune them differently).
For most other large software projects, the conditions (multiple vendors, visible projects & conflicting roadmaps, lack of a benevolent dictator) might require a nonprofit to grow properly and reduce dissensions & risks.
Conclusions invalidated by incorrect assumptions Posted Dec 4, 2010 22:33 UTC (Sat) by hingo (guest, #14792) [Link] (I am the author of the post.)1) It is not an assumption, it is a conclusion. (Ie, there are no "Large" or "XtraLarge" projects in the "just a project column.) 2) I think many people misunderstand the usage of "governance" in the article. It is not focusing on decision making structures of the actual development work, or copyright ownership, or such. Key properties of the kind of governance that is referred to are topics like "Who owns the trademark that is the project name, if registered" or "who owns the domain name that is the online home of the project"? Having said that, it is true that all of linuxmark.org, linux.org, kernel.org... predate Linux Foundation and are not owned by it. On the other hand all of these are still held by non-profits, so it is close enough for the chosen categorization. There are also other differences between the foundations, for instance Mozilla Foundation, via wholly owned Mozilla Corp, has revenues bigger than any of the single vendor products and therefore employs many/most Mozilla developers, which is not the case for any other foundation.
Conclusions invalidated by incorrect assumptions Posted Dec 4, 2010 23:41 UTC (Sat) by jejb (subscriber, #6654) [Link]
> I think many people misunderstand the usage of "governance" in the article. It is not focusing on decision making structures of the actual development work, or copyright ownership, or such. Key properties of the kind of governance that is referred to are topics like "Who owns the trademark that is the project name, if registered" or "who owns the domain name that is the online home of the project"?
Well it's understandable: The term governance comes from the corporate world and does, in fact, mean how things are controlled and run rather than who owns what.
> Having said that, it is true that all of linuxmark.org, linux.org, kernel.org... predate Linux Foundation and are not owned by it. On the other hand all of these are still held by non-profits, so it is close enough for the chosen categorization.
Actually, not all: the trademark is owned by Linus. Arguably the most important piece of infrastructure is vger.kernel.org, which runs all the mailing lists we need, is managed privately by DaveM and Matti Aarnio outside of the rest of the kernel.org infrastructure.
Conclusions invalidated by incorrect assumptions Posted Dec 4, 2010 23:57 UTC (Sat) by hingo (guest, #14792) [Link] Yes, but if there was a problem with vger.kernel.org, then ultimately who owns kernel.org could reconfigure the DNS to point to a server run by people they like better. If a private person owned kernel.org, then that person would be in control. The fact that the non-profit owning kernel.org does not care about the routine operation of vger.kernel.org is uninteresting to my study - in fact I assume in most cases the board of a foundation does not meddle with the actual development work.Regarding the trademark, while I don't know the details, I have made the assumption that the arrangement between Linus and the Linux Mark Institute is somehow perpetual, so that Linus' ownership is mostly symbolic and in practice trademark policy is and will always be set by the LMI. But this was an assumption, I admit.
Conclusions invalidated by incorrect assumptions Posted Dec 5, 2010 0:10 UTC (Sun) by dlang (✭ supporter ✭, #313) [Link]
but if there was a problem with the people running the kernel.org domain, Linus would just announce that he was going to publish all new kernel at newkernel.org and that the mailing lists were going to be moved to @newkernel.org, the question of who owned kernel.org would become irrelevant very quickly.
this is a good thing because it provides a check and balance on all actors, there is not any one entity that really controls the kernel.
even the control that Linus has is only because he is doing a better job than anyone else, if he were to stop releasing kernels, or start doing a horrible job at it, someone else's kernel would start being the one that everyone would pull from and sync with.
Conclusions invalidated by incorrect assumptions Posted Dec 5, 2010 0:19 UTC (Sun) by hingo (guest, #14792) [Link] Yes, totally agree. So like I tried to say already, Linux goes even far beyond the "foundation bringing together a community of collaborators" in that 1) there are many non-profits 2) of which Linux Foundation doesn't even control that much and 3) arguably, the non-profits don't even control what they are supposed to control since like you say, if Linus and a non-profit had a fight, Linus is likely to win.The one thing that I do consider important though is the trademark. If Linus was doing a bad job, and someone challenged him for the leadership position, would that new leader still be able to produce a kernel called Linux, or would it have to be called Newlinux? Could Red Hat still ship Enterprise Linux or would they have to change to Enterprise Newlinux? If latter, then the cost of forking would be significant. In Linux's case even the domains are not that important, since end users don't go to kernel.org or linux.org to get their Linux, but for most other projects this is not true, the owner of the domain essentially owns the project.
Conclusions invalidated by incorrect assumptions Posted Dec 5, 2010 15:35 UTC (Sun) by ewan (subscriber, #5533) [Link] if he were to stop releasing kernels, or start doing a horrible job at it, someone else's kernel would start being the one that everyone would pull from and sync withIIRC that more or less happened around the early 2.4 era when the -ac kernels were a much safer bet than the Linus ones, and were commonly used by both individuals and distributions.
Henrik Ingo: How to grow your open source project 10x Posted Dec 4, 2010 19:10 UTC (Sat) by bobsol (subscriber, #54641) [Link]
Reverse causal fallacy.
Henrik Ingo: How to grow your open source project 10x Posted Dec 4, 2010 22:40 UTC (Sat) by hingo (guest, #14792) [Link] The article does not say:Choose foundation model => Get 10x bigger project It does say: Choose single vendor model => Won't grow as big as those other guys. On the other hand, if everything else is going well with the project, then removing the key obstacle to growth often does mean it can "magically" grow to the next level, until there are new obstacles. Yes, the headline is provocative, but look beyond it. (I'm the author of the article.)
Look beyond the headline Posted Dec 5, 2010 5:35 UTC (Sun) by clugstj (subscriber, #4020) [Link]
We did, the rest of the thing looked pretty "provocative" also. Look, I am not even disagreeing with your conclusions, but what I disagree with is writing a paper that tries to appear to be unbiased research, but reads like a corporate "white paper" (where the conclusion is written first and the research just done to justify it).
Look beyond the headline Posted Dec 5, 2010 10:50 UTC (Sun) by hingo (guest, #14792) [Link] I will happily admit that I had no intention to write a proper academic paper out of this :-)
Henrik Ingo: How to grow your open source project 10x Posted Dec 6, 2010 14:36 UTC (Mon) by spaetz (subscriber, #32870) [Link]
One problem I see is that "project" can apply on many different levels:
Many of the sample "projects" are really umbrellas or "metaprojects" rather than projects.
Within KDE, and Gnome there are both plenty of projects that are neither big nor growing. So taking the Apache Foundation with its dozen of projects driven by multiple actors (and yes firms), and finding that umbrellas are ususually driven by multiple actors is not that exciting IMHO.
Now, if we found that e.g. GNU tar is thriving and growing because it belongs to a non-profit foundation, that would support the finding. :).
I believe the findings of this study, I just find the level of analysis a bit awkward.
|
|||||||||||||