Originally Published in Linux Journal Issue #64, August 1999
20 Years ago, 15 years before founding Plugable Technologies, Bernie Thompson wrote an article for the first issue of Linux Journal. Over the next twenty years, LJ published 4 more articles including a 20-year retrospective in 2014. Thanks to Linux Journal’s generous terms allowing authors to keep their copyright, we’re able to republish the full series of articles here. If you have your own memories and opinions of the time, please share in the comments. This is Article 4 of 5.
In Making Money in the Bazaar (June 1999), we raced across the landscape of current efforts to drive innovation and make a living in the Open Source market. We now introduce a system for consumer-driven Open Source funding. If successful, it could accelerate the pace of innovation even further and create a small industry around developing free software.
If so, what do you do about it? If you are a developer, you can just write it. The Linux model has been “scratch your own itch” and contribute back to the community.
If you aren’t a developer, don’t have the time or the necessary skill—you’re out of luck. You have to wait and hope that some other sufficiently motivated developer with the same need will take on the project.
This can be quite frustrating—you need that graphics driver now. You could speed things up by hiring someone to develop the software just for you. But paying a fortune to have a custom driver developed for a $50 graphics board just isn’t feasible.
A larger group is needed to share the burden of developing a standard driver. Several thousand people in the world probably use that same graphics card and run Linux. Why don’t some of them get together and share the cost of having the work done?
The same concepts apply to scripts, help files, applications—anywhere there is a demand for software.
A Buyer’s Co-op for Open Source
The Internet is an amazing tool for bringing specialized communities together. A group of people needing the same software is just such a community.
The trick is to attract all of the interested parties to the same web site, where they can pool their resources with others wanting the same thing. This site must coordinate the process of gathering support, selecting a developer, evaluating the resulting software and collecting the funds to pay the developer.
The Free Software Bazaar
Interview with Axel Boldt
The Free Software Bazaar gathers bounties for the completion of particular Open Source projects. It can be found at http://visar.csustan.edu/bazaar/. I talked to Axel on April 27, 1999.
Bernie: What inspired you to create the Free Software Bazaar?
Axel: In a Usenet discussion, the question came up of whether buying a Red Hat CD is a good way to sponsor free software development. I then got the idea that there may be a better, more direct way to induce people to write free software—to “cut out the middleman”.
Bernie: Is money important for the free software movement?
Axel: In the grand scheme of free software development, money does not currently play a big role. Personally, I am not at all unhappy about that fact. I like to think of the Bazaar as a place where programmers can get ideas for projects that are actually needed, and where users can show their appreciation for the wealth of software they get for free.
Bernie: Why would any individual commit money towards funding an Open Source project?
Axel: Two reasons: you need a certain piece of software and you think the free software development model would produce the best results, or you feel the need to “give back” to the free software community. Most of the time, it’s probably a combination of the two.
Bernie: What does the system offer for developers?
Axel: Some money, but mainly ideas for new worthwhile projects.
Bernie: What effect could cooperative funding have on the Open Source community at large?
Axel: It may help establish better communication between users and producers of free software. Users will be able to outline a wish list for a new project as opposed to just provide feedback and patches to already-existing code.
Bernie: What is your background?
Axel: I maintain the Linux kernel configuration help texts, the Linux CD Giveaway List (http://visar.csustan.edu/giveaway.html) and the programs tkinfo and WebFilter. I teach mathematics and computer science at Metro State University in St. Paul, Minnesota.
Bernie: What are your plans for the future?
Axel: Retire early and keep playing with free software.
Axel Boldt’s Free Software Bazaar was the first realization of these ideas. It opened in the fall of 1998; within six months, it collected over $25,000 US in offers and over $1200 in payments towards Open Source projects. The site works by letting users browse a list of existing offers. If a user is interested in sponsoring an existing project or creating a new one, they send e-mail to Axel. He then adds their offer to the listing page.
These offers can be claimed by the first developer to successfully complete the work. Axel then notifies the sponsors, asking them to send payment directly to the developer.
The Free Software Bazaar is a great service to the Open Source community. However, a huge ongoing effort is required to maintain momentum and grow the movement into something powerful.
For cooperative funding to become a significant force in the Open Source market, the achievements of the Free Software Bazaar must be multiplied many times. Managing the demands of so many parties is a difficult problem. A cooperative funding service needs to be innovative in solving the confidence and communication problems between sponsors and developers. It should be convenient and simple. It must be professional and build a strong record of trust. In the end, it is essential to attract and maintain a critical mass of sponsors and developers.
CoSource.com is an attempt to create a service that meets these demands. It is a commercial enterprise created to provide the range of services required to make cooperative funding a success for buyers, developers and the Open Source community in general.
It intends to:
- Publicize to better achieve a critical mass of sponsors and developers for each project.
- Have staff work with corporations directly to encourage sponsorship and development of open source.
- Provide VISA/MC/AMEX credit card processing to make payment convenient (especially for non-US sponsors).
- Automate as much as possible with HTML forms and a database back-end.
- Provide a stable web address and organization for long-running projects.
- Foster trust between buyers and developers through simple, standard and legally binding on-line agreements.
- Provide cash advances to developers with a proven track record when they begin work.
- Provide web space to trumpet the financial contributions of corporations and individuals towards particular projects.
- Prevent duplication of effort by having sponsors collectively hire a single developer to work on a project.
- Allow sponsors to back out any time until a developer is finally selected.
- Allow full control for sponsors to select the developer, development schedule, source code license, etc. that suits them.
If all these goals can be achieved, cooperative funding will provide effective answers to the questions, “How does one make a living on free software?” and “Who is motivated to innovate?”
The answers are that innovation is funded directly by users who pay for new features, which in turn supports a small army of independent software developers.
How Does It Work for Sponsors and Developers?
It starts with an unfulfilled need. Maybe it’s a driver for a USB scanner, a plug-in to convert Excel spreadsheets or a port of some game to Linux. The user goes to http://www.cosource.com/ and finds the project to develop this feature. If it is not there, they can add it with a form.
What is it worth to them for someone to develop this software? Whether it is $10 or $1000, the customer sponsors the project for that amount. This is not something done lightly. A buyer is making a firm commitment to pay up if the software is developed.
Other motivated sponsors come along and do the same. CoSource goes out to corporations and seeks to supplement individual sponsorships with a few large ones. Let’s say the project is an HP scanner driver. While HP isn’t yet ready to pay the full cost of developing a Linux driver, they may be willing to pay for 50% or 25% of the effort.
Developers, meanwhile, browse these same lists to identify projects in their area of expertise. Suppose a developer has done a converter for the Excel file format before. That developer fills in a form to bid on the project, answering the following questions:
- How much would I have to be paid to do this work?
- What is my most conservative estimate for time to completion? What license would it be released under (e.g., GPL, BSD, etc.)?
- Who will judge the final product for completeness and quality (e.g., a known and trusted third-party authority)?
- What will be the URL of the project’s web page?
CoSource then prices the bid for display on the system, marking up the bid for transaction costs, historical sponsor fraud, advance payments, project risk, etc.
As bids are entered, sponsors are notified to evaluate them. They submit a simple yes/no form in response. Voting yes to one or more bids is a final commitment involving a legal agreement to follow through if this developer succeeds. The first bid that gains sufficient sponsorship wins. From that time forward, sponsors are not permitted to back out and shortchange the developer.
The winning developer then begins work on the project, providing updates on their project web page.
At some point, the developer believes he is complete. A release is done and judged to determine if it matches the requirements stated in the original project description. If the release fails, the developer may try again many times until their committed schedule runs out. If that unfortunate event happens, the project goes back to the sponsor/bid phase.
If the release passes, the project is complete! Sponsors are notified to fulfill their commitments. They can easily pay by credit card. Finally, payments are consolidated and a single check is mailed to the developer.
What Are the Goals?
Obviously, this process is more complicated than a typical software-buying experience. In return, the consumer gains much more control over the quality and time frame of work. If you needed one of the new features Windows 2000 provides, you would have to wait two to three years after the initial promised ship date to get them. How can a corporation plan ahead for software rollouts with such uncertainty?
Cosourcing puts more control over features, schedule and quality in the hands of the consumers.
Obviously, this system is not intended for charity or non-profit activity. Rather, it is intended to be the most effective way to outsource development of software and share that cost with other motivated buyers. It is intended to be a way for non-developers to “scratch their own itch”. It is intended to be a fertile breeding ground for hundreds of individual and corporate developers. It is intended to make the funding of Open Source a collaborative effort in the same spirit the development process is in today.
In general, it is intended to empower end users to spend their hard-earned money making free software do what they need.
Making a New Market
On one end of the software market spectrum is “closed” software, where intellectual property is licensed on a per-copy basis. On the other end is free software, where intellectual property is created without payment and voluntarily given to the community at large.
Both of these will continue to grow and thrive. On one hand, closed software will continue to be a billion-dollar market. On the other hand, innovative free software will continue to be developed by students, hobbyists and professionals for various reasons. Both systems make sense and they will continue to compete with each other. But there is a possibility for a vibrant third market. One which blends and bridges the differences between the other two. One that brings the free market to free software.
This market will sell software as a service rather than a product, so it will be compatible with Richard Stallman’s original and ongoing vision for public license code. This market will serve the needs of end users by driving innovation in the areas that matter most to them. This market will bring financial vitality to free software, so thousands of individuals and small companies can make their living developing it.
Software in 2010 — A Look Into the Future
With the rise of the Internet and the Open Source movement in the late 1900s, the basic building blocks of software—operating systems and libraries that applications build upon—started becoming a collaborative effort. At first, few people believed that great software could result. Few believed that a rag-tag collection of individuals and companies, working in parallel, could produce a great platform to build upon. But they did it.
Then, at the turn of the millennium, markets sprung up to collaboratively fund these same projects. Drivers, scripts and middleware to connect Open Source with every kind of software and hardware device were developed. Many small but frustrating problems were fixed. Open Source was now the most interoperable software platform available, and it was getting all the polish and customization needed to appeal to the full spectrum of end users.
Open Source did not win out completely. Rather, the result was intense competition between closed- and open-source platforms that drove accelerating innovation for all.
In recent years, the open model has gone on to tackle problems beyond the platform: highly parallel problems that require a huge collaborative effort; projects that require complete openness and collaboration; efforts that are beyond the resources of a single corporation; modern pyramids of software.
In 2004, the first of these successes—the Interling project—was completed. It is, of course, the software we use to translate hundreds of written and spoken languages to and from the common Interling language. Dozens of programmers were required for each dialect to produce the complex grammar-processing codes. The project was possible only through the participation of thousands of programmers worldwide, with work on each language funded by motivated individuals, corporations and governments.
In 2006, we completed the initial work of the Historica Humanica project. Every piece of writing, every painted canvas and every available oral history was scanned and entered into our huge searchable database. While not every individual has or will publish a full autobiography, many have willed that their invaluable memoirs be made available at their death. What can we learn from history? We’ve found we can learn much, especially at the personal level. The human psyche has not changed dramatically over the ages. We are now able to search our records for others who have felt the same pain or dealt with the same concern. In these writings, we have found perspective and understanding to guide our path forward in everyday life.
Now, in these last few years, we’ve begun to tackle the most daunting effort yet—the Neuroscape project to approximate and emulate the human brain. We learned early in our AI work that no one simple algorithm can replicate the wonder of the human brain. Rather, the brain is made up of millions of flexible, evolving rules and guidelines the equivalent of billions of lines of software code.
It is a project we can hope to achieve only through the most massive parallel effort ever undertaken by humankind…
This vision may turn out to be a pipe dream. Consumer psychology has rarely dealt with a case in which a group of consumers pay for the development of a product, then allow that product to be given away freely from that time forward. Psychologically, this is a strange mix of self-interest and altruism.
CoSource.com and others are going to put it to the test. If you’ve ever complained about some missing piece of free software, now you can put your money where your mouth is. Will you?
Originally Published in Linux Journal Issue #64, August 1999