What Motivates Software Developers?

TekZen asks: "LWN posted a reference to an ABC News article that claims 'Other than Linux, all the other open-source projects move along at a rate best described as glacial.' The reason? 'This probably is a function of how motivation and lack of fear work among open-source developers.' Are you motivated by fear to develop at work? What motivates you as an Open Source software developer?"

Comments

Simple.

[Mockingbird99]

Money. Just like everyone else.

Re:Simple.

[Xtifr]

Aside from the fact that you left out a few motivations, I agree. (And you didn't deserve the "troll" mod, IMO.) The motivations of software developers are much like the motivations of everyone else: money, sex, power, prestige, frustration, curiousity, comfort, etc.[*] There's no mystery here--developers are people, and, generally, have the same motivations as other people.

The quote, "[o]ther than Linux, all the other open-source projects move along at a rate best described as glacial," is both false and rather stupid, but that's a separate issue. Open source projects, like all projects, move at a variety of rates, some fast, some slow. I'm a little surprised that Dvorak said that, since he's usually not quite that clueless, but then I'm also not quite sure how his article ties into the question of the motivations of developers.

[*] "Money, sex, power and elephants," is the curious list proposed by Barrayaran ImpSec Chief Simon Illyan.

what major projects move slowly?

[spectral]

KDE releases with decent frequency. Gaim does too. There's eclipse, which has come quite a long way in a short amount of time. Mozilla/firefox.. yeah, real slow there. PHP, hasn't changed a single bit, nope nope.

Compare it to other projects in the same vein. Independent ones, there's not many developers, it moves slowly. Major ones, or ones with funding, there's motivation to work on it besides "Hey, I have a few minutes and wouldn't it be cool if...".

Re:what major projects move slowly?

[MBCook]

The big projects tend to move quite well. But they often have corporate funding or full-time coders.

But what I think the article means is that if you get away from large projects, many really do move rather slowly. Take a look at the 2 billion things on SourceForge. They often move incredibly slow or are abandoned. Also small projects often reach their goals and really don't NEED updates very often.

Fear is a good motivator

[djcapelis]

Fear is a good motivator to do a rush job. Many open-source projects pride themselves on quality work. The BSDs actually follow this philosophy fairly strongly and refuse to commit anything that's just a quick hack.

So yes, open-source projects aren't motivated to do it, they're motivated to do it right.

Fear is a horrible motivator

[Vilim]

And there is no fear to drive anyone to do anything -- no fear of getting fired or yelled at by a mean boss.

I would say that this is actually a credit to the quality of Free Software. Most open source projects won't commit quick hacks, they will wait for someone to do a proper job of it (especially the BSD's). An angry boss breathing down your neck is a very good reason to do a quick hack of the job

Sensationalist headline.

[reynhout]

OK, it was a dumb comment by someone who doesn't know what he's talking about (and yeah, he's a columnist, but that is no insurance against ignorance).

But the article was actually pretty good.

Summary:
1. The open source world is pretty confusing and seems complicated to the not-previously-exposed.
2. More and more people are growing up in this culture and that's trouble for Microsoft.
3. Wouldn't it be great if all that legacy software (that gets sold as part of the assets when a company goes under, and then disappears from view forever) was released as open source. It's not doing anyone any good where it is.

The last point is a great one. The problem is that all that code is usually encumbered by licensing agreements with companies that are also defunct, and the successors are sometimes hard to find and/or don't know how to renegotiate a contract when they don't even recognize the products and licenses being discussed. That, and of course, that there's no monetary incentive to do so, certainly not one that outweighs the risk of getting sued a la SCO by the fifth-generation successor to the ownership of the assets of XYZ corp (d.1990, RIP)..

and FreeBSD and KDE and GNOME and Mozilla and

[bluGill]

There are thousands, perhaps millions of open source projects (over 100,000 on sourceforge alone). Some move, some don't. Linux isn't the only fast mover, in fact I'd have to say that KDE moves faster. Course KDE has about ten times as much code, in a lot more different areas, and that leaves more room. Remember the Mythical man-month applies, but when there is more modular code there is more room for more people.

Check the linux game of the month project. Just a couple months to take a game from almost nothing to great. Course the scope is much smaller than linux. Just one example of what motivates developers, and a good one because they are not a single project like KDE or the linux kernel.

When you define success you can eliminate all projects except linux if you want. However there are plenty of projects other than linux that are successful if you would give a definition that isn't so narrow.

Fear and recognition

[mnmn]

Say your manager doesnt run after you with a stick at work, to motivate you. Say youre given 6 months to develop something with no required feedback during that time. Will you develop the way you'd write drivers for the linux kernel?

Its just kinda yucky to put effort into software which will make another man rich. Its like youre told to sculpt the best sculpture you can, which will be snatched by someone else, youll get paid $5 and he'll put his name on it. Even if you couldnt personally sell the sculpture at $5, its still a bad deal.

And then, there are the constraints which REALLY and I mean REALLY demotivate you. Say you absolutely love developing in ANSI C99, using lots of pointers, using the Intel compiler, QT, and making the code portable across BSD, Linux, Win32 and Solaris. Youre put on a VisualC.NET platform, where you have to use C++, not allowed to touch many libraries you'll depend on, not allowed the Intel C compiler, not allowed to port to Linux etc. Say you think the code should be distributed across many small shared libraries to make it neat, and youre told to pack it all in 2 DLLs. It just makes you want to stop working on it. You just push yourself from that point onwards to get the paycheck.

If someone asked Da Vinci to put a smile on Mona Lisa, he'd probably spend only a few hours on it, PREFER not to have his name signed on it, and he'll paint another personal copy with passion, just for the satisfaction.

Just put 12 developers together, make a list of their IDEAL development environment, OS, language, tools, algorithm etc, and you'd know how hard it is for the employers to instill motivation into the work. Many people have preferences on 'just because' rather than reason, for example I just hate Java and want to use C, perl, OpenBSD, Postgresql, QT everywhere.

Just because.

Re:Fear and recognition

[spectral]

Totally not trolling here, but I was under the impression that QT was nearly entirely C++? You imply 'not c++' as your prefered langauge, but then like qt?

I ask only for clarification, because I haven't coded any gui apps for linux. Windows (api, mfc, vb, foxpro) and generic (java/swing, wxwidgets) only.

In fact, I thought that was one of the big 'draws' of gtk, was that being written in C, it could export to everything. QT, being written in C++, was much harder to make work in other languages.

Re:Fear and recognition

[sql*kitten]

you absolutely love developing in ANSI C99, using lots of pointers, using the Intel compiler

I'm going to guess here that you're probably very inexperienced, in fact have never worked as a developer. Real developers don't give a monkey's about things like that. Languages and compilers to us are tools that are used to solve problems. A mechanic doesn't develop an emotional attachement to his hammer and then try to use it to bang in screws!

put on a VisualC.NET platform, where you have to use C++,

No-one "puts" you anywhere. Last I heard programmers weren't conscripted! You get a job, you work it, you get paid, if you don't like the work or the pay you can either leave, or get yourself promoted to a position where you're making the decisions. That's called the "real world".

Just put 12 developers together, make a list of their IDEAL development environment, OS, language, tools, algorithm

Development teams standardize for a reason - so everyone can work on everyone elses code, and it all integrates nicely. And real developers, as I say, don't even care about those things. What matters is elegant solutions to tricky problems.

I just hate Java

You hate it, but can't explain why? A real developer, who hates any language, will be able to say why, and why that is a problem, for example he might say "I hate java because it has no MI, and MI is useful for...". But it's unlikely that hate would come into it, languages and compilers are just tools after all.

You'll learn all this for yourself, just as soon as you actually start developing something non-trivial.

Lack of a girlfriend.

[Anonymous Coward]

'nuff said.

Chaos and Unification

[russellh]

A note about one point this Dvorak character makes: he seems to think that one day there will be unification among competing open source projects. He states that from an outsider's perspective, it all seems confusing. Well isn't that always the case? If you know which are the most popular, then don't you pick one of those? ie, go to the computer section at Barnes & Noble and buy one of the ones in the box. I doubt there are many non-drivers waiting till there is only one car to buy.

I also have to say that there will not be unification of competing projects in general - but quite the opposite. Projects will be forked and specialized or created anew because they can be; what unification that may happen may be in the form of common component architectures or other standards like common file, DB formats that make such specialization possible. However, to an outsider, it will still look the same even if all the projects are perfectly interchageable and interoperable, which is not unlike the linux situation. Where would project unification motivation come from? I just don't see it (generally) when these projects are for insiders, not for outsiders.

Think there are too many competing {standards|projects|languages|ideas} today? Consider retirement.

Making a difference

[HeyLaughingBoy]

I work for a medical device manufacturer. When I took the job it was really because I wanted a change of scenery and they paid to relocate me 1,200 miles away.

Now the reason I'm still here many years later, and what motivates me to do a good job is that my work has impact on people's lives. It's nice to see the occasional letter from a patient who thanks us (along with his doctor, to be sure) for saving his life.

Sure, our QA department will remind us from time to time that if we screw up, we can kill people, and that is a *huge* motivator to try to do a good job, but a much larger one is realizing that my code really does improve someone's quality of life.

And for that reason I'd really like to spend the rest of my software career in the medical/biotech field.