Hacker News Reader

This guy is absolutely, positively, dead wrong about OSS projects and money.

I think this is a case where your credentials matter--while Atwood doesn't have a huge array of Open Source projects under his belt and just this one data point, he still has one real data point. What Open Source projects do you manage, or have you managed, that would have been dramatically positively impacted by receiving a no-strings $5000 cash infusion?

I've built two companies based on Open Source software now, and I can assure you that spending money to make things happen in Open Source can be challenging. People are often involved in Open Source for reasons other than money. Over the years I've probably managed a couple hundred thousand dollars worth of expenditure on Open Source development...and on many occasions it wasn't with the core developers on a project, because the core developers had no interest in the money, either due to their employment contracts, difficulty in dealing with foreign payments, or simply a desire to keep their Open Source work purely for fun.

When I was considering what my options were when shutting down my old business (just prior to starting Virtualmin), I spent a little time building a website specifically for throwing money at Open Source projects to get things done (called ransomware.com, and I still have the domain...maybe once I've got some free time from Virtualmin I'll spin it into something). It's still a gaping hole in the Open Source world, but it's one that's incredibly hard to fill, and you can't do it in the obvious ways.

Google SoC covers one aspect, but it's an "internship" style model, which doesn't produce "core" code very well--it's just as much for teaching as it is for producing code, and requires significant time investment on the part of the project. SourceForge donations provides "tips", but it doesn't seem to work well for getting specific project work done. There have been a few others, but they seem to fizzle out pretty quickly, as I think the profit potential for the orchestrating company is small...it's almost gotta be a non-profit to convince people to take part. Finally, a lot of folks use eLance, Rent-A-Coder, and others, to get work like this done--I frequently see projects come up related to our software in my Google Alerts--but it's not targeted to this purpose, and the code ends up being "one off" rather than finding its way into the core codebase...I know we have never heard of any of these projects coming back into our codebase (which is fine...a large percentage of our code is BSD-licensed, and so there's no obligation to let us know what's going on with it).

So, while I agree that money can certainly be used to great effect on Open Source projects, it takes time to make it useful, and sometimes you don't want to use it for the obvious "get more features", since a lot of developers on Open Source software like to do that stuff themselves even if it comes more slowly. We've begun spending money on Webmin and Usermin, now that we have some...but for ancillary stuff, like design and logos, and graphics. Jamie still writes almost all of the code (and a couple of other occasional contributors and I write a tiny remaining percentage).

You ask about how many open source projects I manage. None. I am a completely private sector guy.

I've done a lot with $5K; employing that focus I talked about.

I was not taking a shot at anyone, nor at open source. Just at resources vs resource administration. I do not know which project this was (I know the guy mentioned it on his post), nor do I particularly care. A resource is a resource and a manager is a manager. A project that is in development must move forward, open source or not.

Case in point: We are using a certain OSS project in house. We ran into a bug, it's a well-known bug, been there for a long time. We fixed it because we had a time motive, resources, and focus. Right there: done. No hesitating, we needed to move onto something else. That kind of focus would certainly help OSS. $5K or not in our pocket, it was fixed.

Anyway... That may very well explain why private projects move faster than open source ones, the private ones have a profit motive, failure means financial failure. No matter how altruistic you are, if you fail financially, you don't eat.

You ask about how many open source projects I manage. None. I am a completely private sector guy.

So you pretty much don't get Open Source. And that's OK. There's plenty of room for both commercial and Open Source development. But, it nullifies your claims on the topic, since you've established you do not share, or comprehend, the mindset of Open Source developers.

We fixed it because we had a time motive, resources, and focus. Right there: done. No hesitating, we needed to move onto something else. That kind of focus would certainly help OSS.

What makes you think OSS needs help? I think Apache just called to remind you not to teach your granny how to suck eggs.

You're making the very common mistake of assuming Open Source software is for you. Open Source software is for the developer of the software. If anybody else likes it, that's wonderful. But it's not (generally) the reason it exists. It exists because someone wanted to build something. Sometimes they also want to build a business around the software they build, in which case they have to answer all of the needs of commercial software, in addition to keeping an Open Source community happy (which makes it a doubly hard challenge, but it brings some advantages, as well).

Anyway... That may very well explain why private projects move faster than open source ones, the private ones have a profit motive, failure means financial failure.

And it may explain why OSS projects, in general, have dramatically higher quality code. Motivations are different. "Needs" are determined by hackers...not marketing guys, MBAs, or even (perhaps to the detriment of user friendliness) customers.

Though I question your unfounded assertion that proprietary projects move faster, and I think you need to be specific. Does Microsoft innovate faster than the Linux developers? How about Apple? Would Apple be where they are today without FreeBSD and Mach?

I would concur that proprietary software companies sometimes move faster on usability. Certainly Microsoft has better accessibility than Gnome and Linux. But Linux has better large systems support than Microsoft or Apple (but not as good as Sun...though Solaris is now Open Source, I think we can give this point to the proprietary side, since most of the features were developed by a single company and before Solaris was fully Open Source). Likewise small systems support, where Linux has a very solid lead over Microsoft and Apple (and Sun). Hardware support is also better in Linux than both Windows Vista and Apple OS X (and Solaris). So who is moving faster?

And, to come back to Apache...name any proprietary product that does half of what Apache can do or gets support for new protocols sooner.

Open Source developers and commercial developers have different priorities. Your priority seems to be condescension. ;-)

No matter how altruistic you are, if you fail financially, you don't eat.

Altruism isn't the point of most Open Source software.

It's not about being condescending, my point was that the manager of the OSS project was a bad manager. A bad allocator of resources. That's pretty much it.

I have no desire to get into a Technology A vs Technology B pissing match so I'll not even address those issues.

It wouldn't be the first time Atwood made a generalization based on one data point.

How is it possible to "generalize" from one data point? The normal method of generalizing is finding a similarity shared across many data points, and proposing it applies more widely.

Perhaps one could generalize badly with 2 data points, but isn't it just impossible with 1, and therefore he must actually be doing something else?

That logical impossibility is exactly what I meant when I made the comment. If generalization doesn't work for you, how about this:

"This wouldn't be the first time that Atwood came up with a theory that only had to fit one data point."

He could just as easily have said that all open source developers are Italian or that Italians can't spend money, and he would have been just as wrong.

ok i agree

I used to work with Chemical engineers (I was an elec eng) - they had a saying: "one point is a line, two points is a trend" when trying to analyse production data (always only half joking).

This guy is making a generic conclusion (open source does not need money) because of a bad project admin/lead.

I'm not sure I follow this assertion. Jeff is asking a generic question, yes, and he does make some generic suppositions, but he does not really declare any conclusions.

If anything Jeff was surprised by the exchange and is asking the question as he is because he believes (or used to believe) that money could help OSS projects and is curious how, since it is not something that he has experience in himself.

As a side note, do you really think $5K will buy you a quality full-time developer for 3-6 months? I would be very surprised, though I admit I have little knowledge of programming labor markets outside of Boston or SV.

Oh you'd be surprised what you get abroad for $5k... US-based developers with mediocre skills: Be afraid, be very afraid.

On the ScrewTurn website, I can read: "ScrewTurn Software - Developers for Passion". Whatever that means, I don't see them spending the money they got to hire someone.