Users as Innovators - Why Open Source Works

eaglemoon writes "Many people still have difficulty understanding why open source software projects are successfull. The Boston Globe has an interview with Eric von Hippel, a Professor at MIT Sloan School of Management, on users as innovators. In his new book, von Hippel, discusses how open source projects draw on the creativity of ''lead users," who are often ahead of the curve on technology and marketplace trends. Von Hippel shows the trend already is more advanced than is generally known, and users often freely reveal their innovations for the common good. The social efficiency of a system in which individual innovations are developed by individual users is increased if users somehow diffuse what they have developed to others.....he also notes that the transition to user-centered innovation is hard for some companies to swallow.
The online version of the book is available under a Creative Commons license."

Comments

Open Source Presentation

[siokaos]

My writing class had an open topic presentation, and some friends of mine and I just did a presentation on the Business and Development elements of open source projects :)

Check it out
http://neuclid.com/OpenSourcePres.pdf

Problem

[elid]

One major problem with open-source is the lack of artists willing to work under such a license. For an example ot what results, see the "new" FreeCIV.

Re:Problem

[aCapitalist]

The reason is that artists don't have a culture of sharing like coders do (way before FSF or GNU), because that model doesn't really work with art - at least traditional art for obvious reasons.

Re:Problem

[NanoGator]

"One major problem with open-source is the lack of artists willing to work under such a license. For an example ot what results, see the "new" FreeCIV."

I see your point, but I'm not sure the problem is the licensing. I think it's the lack of incentive.

To get a job as an artist, you have to do some stand-out work and/or have project experience. Sadly, this creates a nasty problem: How does one get a job as an artist if they haven't had a job as an artist? The answer? Do some cool stuff on your own. Many artists do this. (Check out www.cgtalk.com to see what I mean.) However, it can be difficult for a self-starter to complete some ambitious work. I know I had that problem. If I couldn't get something done in an evening, I wouldn't do it. So what'd I do? Simple: I took on some pro-bono work.

I did some artwork for a game called Ferion. (Now some of you will understand why it's in my sig.) There was no paycheck. Instead, the agreement was that I'd do the work provided I could take the time I needed to expand on my artform. The result? The work I did for Ferion almost single-handedly got me my dream job. There's absolutely no way I would have produced anything like that without somebody needing me to do it. I'm too lazy.

So, how's this relate to OSS? I think really all it needs is the right presentation. There are LOTS of people who want to make artwork for a living. So long as they know that they're gaining valuable experience, you'll be able to find people willing to get the job done. If the project can offer some visibility, even better!

Re:Problem

[NanoGator]

"The reason is that artists don't have a culture of sharing like coders do (way before FSF or GNU), because that model doesn't really work with art - at least traditional art for obvious reasons."

You'd be surprised. Head on over to www.scifi-meshes.com. There are models of a bunch of starships there that anybody can download. Then you can post your artwork with those models. It's been around for years.

You've got a point that traditional artwork isn't so easy to share. However, communities can share in more ways than one. Technique is just as valuable as sharing code. At least with the communities I hang out with, most people are quite happy to share how they achieved a certain effect. Quite a few even take the time to post how-to's.

I realize we're not talking about precisely the same thing here, but I hope you get my point. Quite a bit of time is spent between artists helping each other out.

Re:Problem

[Benny2891]

There are artists out there who do embrace the FOSS method in the production of their work. They just don't make graphics for video games. (As far as I know). There is a population of contemporary "fine" artists (I really hate that term) who are attempting to do this. You are not going to see there work in conventional venues that often however. Simon Yuill and Chad McCail are working on such a project [http://www.spring-alpha.org]. [http://www.agile-process.com/] and [http://www.machinista.org/] are other examples of artists looking at this. There are others, I just mentioned these to make a point. In a lot of ways, artists and programmers operate in similar ways.
Actually, I am one of them. I recently started a PhD candidacy on this very topic. While my research is still very much in the early stages, I am beginning to realise that in order for the FOSS way of doing things to be more widely adapted, the methodologies need to be reworked. I don't think that actually economics are the determining factor here. Very few of the artists I know (and I know a lot) make much money from the direct production of there artwork. They need to supplement their income either by doing something like teaching or constantly chasing down grants, endowments and the like. It wouldn't be too hard to make the argument that "opening" your practise would actually help you secure more of this type of income. For the most part, Creative Commons [www.creativecommons.org] is sorting out the availability of licenses that are appropriate for artwork, so thats not that big of a deal. I think the biggest reason that artists are reluctant to embrace this kind of working is that we are still too hung up on the "original". It is deeply ingrained in the collective culture of the the arts. Originality of the idea and the art object itself is of the highest value when artists are assessing art. The FOSS way of doing things is perceived as encouraging copying and unoriginality. This is a big no-no for your average contemporary artist. If artists were just to look their history, they would realise that this obsession with the original is a relatively new thing, and that all creative works are built on top of the creative output that has already happened.
While all that i just mentioned is within the framework of art as in Tate Modern / Guggenheim kind of art, the artists that do more commercial work most likely went through some sort of art education that the "fine" artists went to. They all more or less leaned the same values, and regard their work in a similar way.I think it will be a while before you see this happening more commonly in areas like Graphic Design and CG stuff.

the user's perspective

[kebes]

Users will have a perspective on products that the programmer never will--namely the perspective of someone who *doesn't understand* how the application works! After designing and/or programming a piece of software for a long time, you can get to know it so well that every aspect of it seems obvious. Yet to a fresh user, who has no clue what is going on behind the scenes, your choice of layout may seem confusing. It has been said (many times) before, but programmers/designers need to *listen* to what users are saying. If something is hard to use, then it should be fixed! And yes, the users of a product will have tons of useful ideas for how to make a product better.

Programmers know too much about the inner workings of a system... and thus they will immediately think of all the reasons why an innovative idea (interface element or feature) won't work. But the naive user, not encumbered by such restrictive thinking, may propose powerful features and novel interaction schemes. Some of these won't work, and some really are too hard to code, but there are many gems.

Bottom-up innovation

[Spy der Mann]

I think this may have to do with companies' "top-down" corporate education. They believe they're "the best", they spend MILLIONS on hiring "the best". They spend millions on maintaining this structure, hiring even more people, buying the competition (*cough* Microsoft *cough*), etc etc.

What can a simple user teach them? What can one single guy do? He's got no budget, doesn't have the resources to get "the best of the best", and can't possibly manage dealing with copyright issues. 'You think a bunch of hobbyists can do better than us?'

Bottom-up, gentlemen, bottom-up.

Re:Bottom-up innovation

[cgenman]

They believe they're "the best", they spend MILLIONS on hiring "the best".

I know you put "the best" in quotes for a reason, but it's worth pointing out that corporations in no way get "the best" for their money.

Generally speaking, corporations are stuck in the illusion that if something costs more, it must be better. An employee that was making 105k as an Active X programmer must be better than the QNX programmer making 85k, so let's hire the 105k programmer and pay him 125k. If a fast-talking guy can come in and say all of the right things, the heads will believe that this guy is the perfect person to be lead programmer, even though he just sold you on the idea of doing your 1,000 concurrent user database app in Access. Corporations are great at throwing money at getting great salesmen, but they're not always so good at getting good leaders, programmers, designers, etc.

Re:Bottom-up innovation

[QuantumG]

You do know there are technology companies out there that let the engineers hire people themselves? When I was at VMWare an interview took all day. The recruit would be taken around to every engineer in the company and asked to help out with whatever the engineer was working on. The engineer would form an opinion of the person and their abilities and then send them on to the next engineer. This worked especially well because, at the time, every engineer at VMWare had their own office, or shared with 1 or 2 other engineers, but there were no cubicals. At the end of the day the recruit was sent away and the team leaders asked everyone what they thought of the recruit. If there was an overall good feeling the recruit got hired.

Just remember

[YrWrstNtmr]

open source projects draw on the creativity of "lead users," who are often ahead of the curve on technology and marketplace trends.

Some of those curves and trends lead to dead ends. Valid dead ends.

Don't get discouraged when they do, know when to kill it, and move on in a different direction. But do move on.

Polish

[mr. marbles]

My question is even though Open Source can create massive amounts of ground work, why is it still generally incapable of shipping fully polished products? Take a look at the Mac, they went the extra mile, they took all the innovation of the open source world and did all the work hobbiest don't do. What does open source need to make linux or something else fully polished? What makes open source projects like Firefox beat the curse?

Re:Polish

[jayloden]

I often think about this very topic...if someone or someones would come along and put the pedal to the medal on polishing those things off, Linux would become one hell of a competitor.

I think it comes down to the hacker mindset, where the interesting problem is king. No one wants to work on the "boring" tasks like cleaning up GUI buttons and interface text. They'd rather work out exotic new gut level kernel code or rough out the next great feature. We're always questing for the next feature and getting it sketched into the system, and never quite get around to inking everything in. I think eventually, this will happen, at least to some degree. In some places it IS happening, but it's going to take time.

-Jay

Re:Polish

[Bastian]

The same thing that allows the OSS community to produce so much ground work so efficiently is what keeps many (most?) projects from becoming particularly polished. The great majority of OSS programmers like to just sit down and hack stuff out, and aren't particularly interested in the overhead and extra effort involved in maintaining a high degree of internal coordination and consistency. Plus, since most OSS developers are working on a volunteer basis, you can't really force anyone to conform to anyone else's standards.

I haven't been a member of the OSS community for very long - about a decade - but I get the impression that this is largely a fairly recent cultural development that coincides rather closely with the rise of Linux. If you look further back at older projects such as BSD Unix and XFree86, you may notice that there isn't nearly as much of this explosion of forks and competing projects. BSD only has five OSS offspring that I can think of - Free, Open, Net, Darwin, and Dragonfly. Of these, all have very different goals - FreeBSD is aimed at being a high performance Unix for commodity hardware. OpenBSD is designed to be rock-solid secure and stable. NetBSD is insanely portable. Darwin has its own kernel and is largely a move by Apple to get OSS help in developing its own operating system, and Dragonfly is aimed at scalability.

Compare this with the Linux community, where there are oodles of different distributions - many with only minor differences in architecture or philosophy - in a constant state of flux. Many of the Linux distributions that I have used as my primary OS over the years have all but disappeared (Yggdrasil), and many others appear to be in a state of rapid decline (Slackware).

Again, this is both a blessing and a curse. On one hand, this culture leads to a tremendous amount of exploration and innovation - consider the plethora of package management philosophies you have to choose from in the linux world, or the huge pile of GUI toolkits available to software developers. On the other hand, this leads to a whole mess of duplicated (some would say wasted) effort - consider how many different packages of the same program many software projects have to maintain and how most major distributions roll their own packages of all the most popular software, or how you may find yourself installing several UI libraries (all of which, you must admit, mostly do the exact same thing) in order to use all the applications you want.

The projects that escape this - Mozilla, the Linux kernel, Mono, etc. - mostly do so because they get a lot of corporate backing, which provides a lot of paid developers and business discipline which can exert a degree of control over the swarm of amateur and hobbyist programmers who are constantly coming and going.

Re:Polish

[Peter La Casse]

I think it's a matter of the "90/90" rule; the first 90% is the fun part, and the second 90% is the not fun part.

Lead Users

[pyrrhonist]

In his new book, von Hippel, discusses how open source projects draw on the creativity of "lead users," who are often ahead of the curve on technology and marketplace trends.

I'm made of copper, you insensitive clod!

Re:MIT

[AvantLegion]

>> PLEASE, can we get some opinions from some other schools please. There is nothing this professor is saying that hasn't already been said a thousand times on slashdot.

Indeed. Quotes from my professors that are regularly on Slashdot:

"In our discussion on type systems last session, we noted that, in Soviet Russia, systems type YOU!"

"A lambda term is in normal form if it contains neither a redex nor hot grits."

"In Korea, only old people use the nameless lambda calculus."

"An ALU may consist of an adder, a block carry circuit, an input circuit, ?????, and profit."

Eric von Hippel's course at MIT

[patiwat]

Eric is a great teacher - I took a graduate course with him on Innovation Management (15.356) a few years ago. The course was recently renamed "How to Develop Breakthrough Products and Services" and is available via MIT's OpenCourseWare at http://ocw.mit.edu/OcwWeb/Sloan-School-of-Manageme nt/15-356Spring2004/CourseHome/index.htm. The course home page also gives a very brief overview of Eric's lead user concept. It's one of his pet ideas, and although it isn't the sole focus on the course, it certainly is one of the foundations.

The actual class was wonderful: a mix of working scientists and R&D executives, Sloanies and other MIT grad students, and a couple of undergrads sitting in. Lots of student interaction and learning from your peers. The individual project was a good experience as well - I wrote a paper analyzing why Lockheed's X-33 space plane project failed, and what could have been done so that the the technologies developed (autonomous navigation and landing, composite materials, linear aerospike engines, metallic thermal protection system) didn't die with the project. Eric gave lots of guidance and advise on the analysis.

The nature of the developers

[hellfire]

It's not limited to linux. I've seen several Mac and Windows programs that don't appear to have a lot of polish. Hell, my own company, who happens to have the largest market share of software sales within its niche isn't nearly as polished as some competitors.

So many developers are either simply not experienced in proper structure, or proper GUI design, or they are under pressure by execs or marketing departments to get the project out the door. The important part of the software is does it do what it's supposed to do? Yes? Then ship it.

The type of polish tends to differ. There are plenty of windows and mac apps with bad GUI, but it happens more in the Linux community because the body of developers don't have a deep background in coding GUI interfaces as to Mac or Windows developers. Windows programs tend to have more bugs but the windows environment provides a stable interface usually. Mac apps probably have the best of both worlds (as long as you agree the OS X interface is as good as the OS 9 interface).

Re:What does "lead users" have to do with open sou

[benjamindees]

There must be a big myth that open source end-users have more influence with open source projects.

There's more money in OSS than you think. Of course proprietary developers listen to their customers too, but to a lesser extent. The difference is, OSS providers can't hide behind lock-in file formats, obtrusive licenses, and established monopolies if they want to make money. They have to earn it, by listening to users and providing for their needs.

And it's not as easy as a "usability study" would have you believe. It means living with users day in and day out and dealing with all of their problems, not just watching them click a few buttons for a couple of hours and optimizing the menus. A vast majority of the proprietary crap software wouldn't exist if all programmers were forced to then support it from the "hell desk". Fortunately for OSS, many developers *are* supporting end users directly, and code accordingly.

Redhat should have done that before Apple..

[xtal]

I often think about this very topic...if someone or someones would come along and put the pedal to the medal on polishing those things off, Linux would become one hell of a competitor.


Redhat had enough money to do what apple did for BSD.. In a way, apple is the perfect company to put that polish on - it's what they do. I'm not sure there's a place anymore for a polished desktop UNIX the way there was a place for it in, say, 1997.

It certainly is what Linux needs to smash onto the x86 desktop. People seem to get too caught up on Holy Wars (tm) to make this one happen though. Maybe I'm wrong - but I'm writing this on a powerbook, too, and in 1997 I was one of those point-and-laugh at mac types.