Slashdot Mirror
What Makes an Open Source Project Successful?
crowston asks: "There have been a number of discussions on Slashdot and elsewhere about how good projects work (e.g., Talk To a Successful Free Software Project Leader), but less about how to tell if things are going well in the first place. While this may seem obvious, most traditional definitions of software project success seem inapplicable (e.g., profit) or nearly impossible to measure for most projects (e.g., market share, user satisfaction, organizational impact). In an organizational setting, developers can get feedback from their customers, the marketplace, managers, etc.; if you're Apache, you can look at Netcraft's survey of server usage; but what can the rest do? Is it enough that you're happy with the code? I suspect that the release-early-and-often philosophy plays an important role here. I'm asking not to pick winners and losers (i.e., NOT a ranking of projects), but to understand what developers look at to know when things are going well and when they're not."
the "..." part before the "Profit!"
taken! (by Davidleeroth) Thanks Bingo Foo!
My suggestion, read "The Cathedral and the Bazaar" by Eric Steven Raymond....not just that paper, but the actual book of papers he put together. Very good read, and he takes a lot of the ideas of open source projects and converts them into real world applications.
What makes Open Source (or ANY project) successful is ambition and drive.
You have to be realistic in your goals, and have the drive to see everything through. Open source projects that are abandoned or failed is simply because the developers gave up for one reason or another.
You know how you got together with your buddies to make a game, but never got very far? That is a classic example of a project failing due to lack of ambition and drive.
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
and it does what it is supposed to do, cleanly and efficiently, then by definition it is successful. Popular and successful are definitely not the same thing, even if you gauge a projects success by its popularity
It's easy.
Are there more people using the project than developers? If so, it's successful.
Do you enjoy working on it? Then it's successful.
Most open source projects are essentially hobby projects. Whether or not they are 'successful' on a large scale is usually irrelevant.
I don't have a sig...Do you??
Success being measured on how many hits you get on your download page and how many downloads of your project actually occur.
It's one thing to be satisfied with your own code, but to see others satisfied with it, well that's what I'd want at least.
The Pigloo
The same thing that makes any software project successful:
a win32 port.
Next question please.
I don't need no instructions to know how to rock!!!!
If a piece of software serves your needs -- whether you built it yourself, modified something someone else made, or just downloaded a pirated copy of something commercial -- it is "successful software."
"Success" is not really a concept that can be accurately applied to "software in general."
If you are an OSS designer you will have your own standards of what is "successful" and what is not for your baby. These are not necessarily standards held by anyone else, nor should they be.
Does it really matter?
Really, the only reliable measure of a software project's success is if its useful to you and meets your needs. If your satisfaction with a project is dependent on other people's useage/opinions of the software, you will probably never be happy. Remember, open source software development is for 1) fun and 2) to scratch an itch. Anything more is chasing after the wind...
What I see as successful are the projects that do something that already being done by a successful commercial application, only doing it cheaper and very well.
The ones that do the same thing, only poorly, will fail.
The ones that end up costing more to implement than the commercial application, even if they do it better, will fail.
The projects that do something new, something people don't know they need, are doomed to failure from the start because your typical open source developer doesn't have the resources to market the product. There was a time when people didn't need sliced bread. Bakers didn't need bread slicers. But the bread-slicer-makers had the resources to market their product and convince the bakers and public it was needed. So now we have sliced bread, and nothing greater since.
Give me my freedom, and I'll take care of my own security, thank you.
Profit indicates success of the marketing plan, not the software development effort. It doesn't matter if you're coding for love or for money, there are some things that apply to both. Take a look at process and product. How is the process? Are there goals and are they being met? How is testing coverage and how often is testing being done? Is the code maintainable? Take a look at the end product. Does it do what it's supposed to without too many bugs? Are issues being addressed in a timely manner? Most importantly, how well does it fit the need for which it was designed?
Milestones: establish concrete goals when you start the project, along with a timeline. Of course, these may evolve over time. Happens for commercial apps, too. If concrete milestones aren't met at some point, it's just vapor.
Traffic: both developer and user. Is there a relatively continuous level of input/interest in the project? If developers don't want to develop, and users don't want to use, it's probably going nowhere, even if it's the best thing since the BeOS.
"If, therefore, any be unhappy, let him remember that he is unhappy by reason of himself alone."
~Epictetus
Stallman demands that people call it GNU/[Foo]
As an end user of successful open source software, I would have to say that the best way to get feedback on the success of your project is through an X-users mailing list where X is your project. It seems to me that the more activity in a mailing list usually indicates the size/success of a project. I have spent time in the past simply lurking on a mailing list for a while for products that I am evaluating, and in other cases, I join the mailing list to lurk right away on desktop software that I am not already familiar with. It is the best measure thusfar that I have found.
Stable in relation to the time invested.
Useful...to someone (this is open to broad interpetation).
It should have the goal of attaining at least as much funtionality as any of the software that it is replacing.
Other examples of good OSS is squid, openoffice and, yes, even Linux (Red Hat 9 is least as functional as Win98SE, and that is just from an end user standpoint).
Just my W.O.
WAR TUX!!!
to understand what developers look at to know when things are going well and when they're not
The bug list and feature request list are one way. Strong feedback implies interested users. Also adoption by other developers into the development group shows others are interested, so you must be doing something right.
Developers: We can use your help.
on one of my open source projects I used (more accidentaly than deliberately) the technique which is standard among people who write exploit. I have a small error in the makefile which causes something liek 50% of people come back for help on compiling it. This gives me pretty good estimate of how many people are actually using the package :)
Of course this leaves out win32 users who just download the binary, but oh well.
I passed the Turing test.
You have been sued by a huge mega corp with a team of lawyers over patent infringement and the EFF comes to your rescue.
-- Thou hast strayed far from the path of the Avatar.
Some projects are simply on the right spot. Good examples are X11, SDL and Mesa. There was overwhelming need for it, so more developers quickly joined ranks.
Some projects are outright glamorous in a geeky way. Anyone working on the Linux kernel enjoys the respect of any geek for instance. Stuff like drivers and VM are supposedly tough subjects and anyone involved in ANY way is much more kool than someone making widow managers, no matter how complex.
Some projects provide the much needed high of bashing the Goliath. Wine and Samba fall in this category. Look ma! No windows. And seeing Bill Goates and Balmer try and pull the rug under a project that makes no money is just glorious.
Projects really attract various developers for various collections of reasons. The best reason is the most original.. to scratch that geeky itch. Thats how Linus started the kernel and how others like Alan Cox joined in. Thats how UNIX was originally created and BSD nurtured in the universities. Being so big now, the opensource world has other reasons kicking in, like a smart student seeing the market is kaput, realises he needs something big put on his resume fast. Thusly security and networking projects boom! Included here are also java-related projects.
The most popular projects reach there because theyre there at the right time. Apache didnt quite start out with the best design, but a good webserver was NEEDED, and apache most of the time had more features than the rest.
How do popular projects maintain their status?? Momentum of course. Both apache and the Linux kernel are good examples. FreeBSDers fume on why dont teen hackers flock to BSD. Everyone knows Linux, and once its in the upper parts of the corporate, everone needs to learn it. The media follows it and the natural positive feedback keeps it going. True also for proprietary software, like the most used OS out there for example. Bad quality but who can stop THIS momentum easy??
Yet some softwares quality and design are simply good. They have the power to dethrone the champion. Qmail simple came and is gradually removing sendmail from its position. Proftpd is removing wu-ftpd, and we can only hope Linux or FreeBSD does the same to Windows.
"Give orange me give eat orange me eat orange give me eat orange give me you." -Nim Chimpsky
Hmm, a lot of the posts seem to be missing a big point.
A good metric for "success" in an OSS project must be whether the developers have fun hacking on it. Even Linus has said repeatedly that he made the kernel "just for the fun of it".
Most of the projects are hobbies, and the point of a hobby is to provide an interesting diversion for the hobbyist. If thousands of people get to enjoy a web browser/OS kernel/game/whatever as a side effect of the hobby, well that's just dandy. But if it isn't a commercial product, then who cares about market share, step-3-Profit!, or any of that other nonsense?
Liberal (adj.): Free from bigotry; open to progress; tolerant of others.
Wearing pants should always be optional.
Are you still having fun?
I've commited some of my spare time to open source projects and even started a few pet projects of my own. While success can sometimes be measured by number of users, or downloads, or mailing list traffic, I think it's worthwhile to step back from the project and make sure you're still having fun. At least that's important for those of us who develop open source software as a hobby as opposed to those who do it for a living (and there are many more hobbiest out there). If suddenly you find yourself dreading to read your mailing list or fire up you text editor or IDE, then you know it's time to take a break or re-evaluate the project.
Then again, every developer and project has different goals and really it's only by these individual metrics that a project or individual's success can be measured.
There was an interesting thread on the Jakarta general mailing list about this a couple months ago. You might want to check it out.
Who said Freedom was Fair?
The important gateing factors on any Open Source project are:
1) Motivation (a problem to solve, that people
can agree upon)
2) Working code (something that comes close to
solving the problem, or from which people can
see a solution)
3) Community (communications and peers to provide
a context in which the work can take place)
A lot of people have #1, so they declare a Source Forge project, try to cookie-cutter #3 (impossible to do), and leverage having #1 and #3 into someone creating #2 (also impossible to do).
Mozilla had #1, some of #3, and almost none of #2 for a very, very long time, and it's still suffering the backlash from it (for example). BSD did not take off until Bill Jolitz made it boot. Fetchmail sort of works, but no one cares. Etc..
As a matter of fact, I claim that, given any #2, I can *find* #1, and *create* #3.
It's trivially easy to start Open Source projects by the dozens, if you are even a halfway decent coder: just make something good enough to work, but lacking enough to convince a group of people that they could (and should) improve it, rewrite it, or otherwise do better.
That sounds like most modern commercial software, to me, since it has legacy design factors from the 1980's/1990's causing it to need documentation, support, and training materials as part of the (no longer relevent) copy protection systems that grew up around the software developement process.
Seriously, it took a *lot* of skill to come up with the first Word Processor that needed documentation for people to be able to use it ("PC Write"). The author, Bob Wallace, said at one convention where he spoke, "Software...", gestured expressively above and to the sides of his head, "...is all up here. I sell manuals.".
-- Terry
There is no one definition of success for an open source project. Anybody who starts one should have some goals in mind (e.g., hack on cool code, make something which solves a problem for me, make something which is used by 100/1000/1,000,000 people). Success is meeting those goals.
Here are a couple of examples.
I wrote GNU/Taylor UUCP. When I started, success for me was to develop a UUCP package which would be widely used by people without the money to spend on AT&T UUCP, and to be the premier UUCP package on free Unix systems. I met those goals.
I was the GNU binutils maintainer for a few years. During that time, success for me was providing, on multiple platforms, 1) an assembler which could handle whatever gcc generated; 2) a linker which was compatible with the system linker (on a non-free Unix system), and was faster; 3) tools which were very fast on free operating systems--specifically, much faster than gcc so that they were not the bottleneck for development; 4) adding full support for shared libraries. Those goals were only partially met--on Solaris, in particular, the Sun linker was better.
If you don't have any goals, then you can't succeed. If you can't measure your goals, then you can't know whether you have succeeded.
What makes open source projects successful is obvious. Look at things like Mozilla, gaim, DC++, CDex, etc. What do they all have in common?
Most open source projects fall into one of the following categories.
1)A program someone wrote for themselves, and decided to make freely available for the heck of it.
2)By geeks for geeks.
3)Done by a group, for free and open, but thinking like a commercial product.
3 are the succesful projects. They have good GUIs, they don't crash, they have features that make them better than commercial alternatives, they install easily, they work on many OSes, and they are generally useful. They are often mistaken for commercial products. Slick interface is key. They just happen to be free and open.
The GeekNights podcast is going strong. Listen!
As I and others have said, success for an open source project is defined by meeting your goals.
But let's say your real goal is to be a respected member of the open source community (which, as we all know, leads to fame, groupies, and vast wealth). What should you do to meet that goal? (Actually, there are several ways, but I'll only talk about ones which involve starting an open source programming project, since that is what the original question was about.)
First, your project needs to be something which other people will want to use. Don't write another mail reader. Write something new, at least new to open source. If you don't know what people want, you'll have to ask them. In general, your project needs to either be an open source replacement for an existing proprietary program, or it needs to create a new and interesting niche.
Second, your project needs to work, at least minimally. You have to be able to get it to the point of working, either by writing it yourself or talking people you know into pitching in. If your project doesn't work at all, few people will contribute to make it better.
Third, you need to sell your project by mentioning it on Slashdot, on relevant mailing lists, and on relevant web sites. You need to do this respectfully. One approach is ``I'm looking for suggestions on how to improve my FOOBAR program. It can already do AMAZING THINGS, and I'd like to know how to make it work better for specific users.''
If you follow these simple steps, you too will be on the road to fame and fortune! When you get there, just don't forget the little people who helped you along the way.
Good forks have the following in common:
Consider some good forks:
- mozilla -> phoenix -> mozilla(whateveritisbird)
- X11 -> XFree
- GCC2 -> EGCS -> GCC3
- linux is perhaps the best example - two major branches running all the time, and both
(particularly the 2.3, 2.5, etc. dev fork) heavily forked themselves. 2.5 changes are
often backported to 2.4, even to 2.2, and the maintainers all still talk to one another.
By way of contrast, the GNUemacs/Xemacs fork is a prime example of a bad fork. Bad blood, wilful incompatibility, divergence, duplication of effort.If XFree's current "governance fork" turns into an all out code fork then that would, I fear, be a bad fork - all that bad blood will surely make things very difficult technically.
So perhaps the best advice to a successful project is "encourage forks, and provide a safe environment for them". Apache and Mozilla both do this, to their benefit and credit.
## W.Finlay McWalter ## http://www.mcwalter.org ##
Specifically, when you see other projects you think of as being successful, including references to your project in their documentation. Even if it's telling users how to workaround your program and its, um, features.
It was a strange thrill to see ESD called out by name in the Quake (2?) for Linux documentation from Id software. I knew the project was onto something when Id deemed it necessary to warn people that my simple software audio mixer would interfere with the audio in Quake. They were expecting it to be enough of a problem to head it off in the official documentation. That's a user base.
If you have a program for Linux, inclusion in one or more major distributions is also a sure sign that lots of people are getting some use out of your program. If that many people are using your program, it may even outlive your ability to contribute to it...
How's my programming? Call 1-800-DEV-NULL
However, those are largely things we don't control. The controllable factors of success are more interesting to me. I guess it's because there are lessons on software engineering here. Cool projects can be run into the ground, and tiny niche projects can do well if they're well-run.
Hands down, the best nuts-n-bolts coverage I've ever seen on important issues to successfully developing open source is in a book by Karl Fogel, Open Source Development with CVS. Fogel's one of the developers behind CVS and it's planned successor, Subversion
The book is an interesting paradox: it has 1/2 the chapters GPL'ed. When I started working with CVS, they were useful enough that I bought our development team two copies of the book. Then I read the rest of it... and those are the chapters I'm talking about. Absolutely, they're the best summary of what it takes to successfully run a GPL-ish project. (Ironically, they've GPL'd the technical detail chapters and you have to buy the book to read the parts that talk about things critical to the success of an open source project).
Success is helped by things like doing lots of releases (seeing progress gets others to buy in, and not seeing progress leads to people quitting in frustration) and only adding features you need (let someone else add the features they need). There's a lot more here, but I'm not about to steal Fogel's thunder. Many of these are ideas that are effective in regular development, especially on custom coding projects within big companies.
The focus on GPL code is not the same as on shrinkwrapped products: you're not trying to add features just to add selling points. You're trying to get more people to use the project.