With open source the e-darling of today's e-press, it may strike many as heresy to say that the movement
lacks sustainability. In these days of disintermediation, melting borders, and death to old-line institutions,
some of my more libertarian colleagues may blanch when they hear it stated that the lack of sustainability
in open source is due to the lack of institutions. More institutions?
Gag me with a civil service manual!
If you listen to self-anointed open source popes (albeit gun-wielding popes) like Eric Raymond, author of
manifestos and the occasional snippet of code, what we are witnessing with miracles like Linux and
Apache is nothing less than a Revolution, an official Whole New Paradigm, the End of the World as You
Knew It. Microsoft will only be a memory and software will enjoy the
inalienable rights of life, liberty, and the pursuit of packets.
I regularly attend the Reformed Church of Open Source and occasionally go to services at the ultra-orthodox Old Church of Free Software, but I confess I don't always believe that every program will be
saved if only it would strip off its compiler and walk naked through Comdex.
Open Source Miracles
The PC view these days is that Open Source is a self-sustaining system. Look at PERL. It was incubated by
Larry Wall at the Jet Propulsion Laboratory, then moved into a somewhat fuzzy middle stage, and has
finally ended up as a stable, mature language supported by O'Reilly & Associates. O'Reilly sells millions
of dollars of PERL books, plus helps maintain the CPAN library and does a nifty yearly conference. PERL
is saved, the language (and libraries) are dynamic, and the language has a long-term future. It certainly
is an essential part of my computing infrastructure.
Or, look at Linux. Incubated by
Linus Torvalds and a huge cast of volunteers, the software reaches
millions of computers and is under active development. With companies like
VA Research selling Linux-based computers and
maintaining, supporting, and distributing an
operating system distribution, Linux is now in the mainstream.
Finally, look at two stalwarts of the movement, TCL and Sendmail. TCL was incubated at Sun and then moved into a fuzzy middle period where the future direction
was unclear. Now, TCL is a rousing success story at
Scriptics, which makes TCL widely available
and sells a great development kit called TclPro that features wrappers, compilers, and debuggers. Great
stuff! We bought TclPro and will be buying lots more copies.
Sendmail, the macro language we all love
to hate but all use, was incubated at Berkeley. The software has recently ended up in
which keeps the base version widely available and sells value-added support and premium versions. I must
confess we haven't sent Sendmail, Inc. any money, but the possibility is certainly very real.
So, what's wrong? I would submit that the focus on the open source movement, with
Linux and Apache as poster children, is not addressing the problem. For every success story, there are
dozens of failures. One thing is apparent, however. We are witnessing the blossoming of a new
phenomenon at a level not seen in a long time: social entrepreneurs willing to put in huge numbers of hours
to make their community a better place.
This is not the first time we've seen this. For thousands of years, there have been periodic outpourings of
people willing to pitch in and improve their communities. Indeed, the Internet itself was built this
way. One can think that the Internet was the creation of NSF and ARPA and our brave private sector, but
the real reason the Internet didn't falter in mid-stream is that it was the product of an intense volunteer
movement. Thousands of people got day jobs in various capacities but made their real gig building
protocols (and implementations of those protocols) that worked. The Internet Engineering Task Force was
the meeting ground for this volunteer movement in the same way the Institute of Radio Engineers was the meeting
ground for the social entrepreneurs that built radio in the early part of the twentieth century.
But, it takes more than just human capital to build real, functioning services. The Internet lowers entry costs allowing
well-intentioned volunteers to start a new service with very little monetary
capital or computers or other expensive barriers to entry. But, when these enterprises become bigger, real resources are needed. The Internet means you don't need huge computers to start a service,
but when you reach millions of people, human capital needs to be supplemented with other forms of
So, what's wrong with the system adopted by Open Source advocates? Open Source is good for business.
Now that business understands that, everybody will live together in a happy coexistence. PERL is great for
business, lots of books get sold, the language is saved. Problem solved.
There are two problems with relying solely on human capital in the form of Open Source
hackers. First, the process is accidental. Sometimes it works, sometimes it goes dreadfully wrong. The
second problem is the process works fine for certain specialized, mission-critical software poster children,
but the focus on open source leads to the neglect of less glamorous but equally critical pieces
A Better E-Label
Open Source is a diversion from a broader class of programs: public works projects for the Internet. With
some trepidation, let me attempt to consolidate this concept into an e-trendy moniker: "E-Works" (not
affiliated with Ewoks or
STAR WARS in any way except in search engine results).
The E-Work concept is different from Open Source in that it
emphasizes the overall service provided to the community, not the fact that somebody wrote source code and
published it. PERL is an E-Work because it consists of a language and the things that make the language
useful: ongoing development, ports to multiple systems, training, books, and a community of users that
depend on PERL. By contrast, companies release the source code to a device driver for an obscure
product that only they sell (and which has no conceivable broader use), it may be open source but it certainly
isn't an Internet Public Works project.
Open Source defines its product by the fact that software is visible. Free software is defined by the fact
that the software is free. Public works means that a valuable service is being offered to the general public.
Think New York City. We need Bloomingdales and Macy's (and indeed these are part of the New York
community and do public projects such as the Macy's parade). But, New York also needs public works
projects like Central Park, Lincoln Center, the library, and parkways.
In the Internet, the focus on things like source distribution takes a focus away from a more important
criterion: is the service a useful one that increases equity and access for the public?
If you buy the global village metaphor, you have to buy the rest of this communitarian metaphor: we need a mix of public
and private activities on the Internet. PERL may be working, but look at the Domain Name System. The
NIC is now the intellectual property of Network Solutions and new top-level domains are being promulgated to
registrars that can meet very high entry costs. The only part of this essential service that seems to be
working well now is the informal operation of the root servers by a group of volunteers and the fact that the
BIND software is still widely available, thanks to the
Internet Software Consortium.
The Domain Name
System clearly isn't being run as a public works project and the U.S. Government, after an abortive attempt
by Ira Magaziner that was as successful as his efforts in health care reform, has become a primary stumbling block in
reforming the way our DNS is run. The DNS has been turned into the whipping boy for agressive corporate interest
under the shell of the
Internet Society and an
ICANN that seems overly
attuned to not making waves and keeping the Department of Commerce
If you think that public works are well cared on the Internet, broaden yourself from source code to
data. After all, the whole point of software is to munge content. Look at the Federal Government, for example.
While some progress has been made in making core public information from the feds available to the
public, there is so much remaining to be done it is a true shame. The Patent and Trademark Office has been slowly getting itself
on-line, but the rest of the Department of Commerce seems to be pushing ever closer to a commercial
model where their data is a product, not a part of the infrastructure for the general public to build our cities
on. The retail information industry still rakes in hundreds of millions of dollars per year reselling
information that we already paid for with our tax dollars. That information
should be a lowest common denominator fundamental utility, a public good not a product.
The federal government's efforts to build an information economy infrastructure
of crucial data has been slow. Efforts by other
groups that should be building up these databases have been almost
non-existent (and highly unwelcomed by the relevant MIS departments in
the federal government when they do occur). Why? The barriers to entry are too high: you can
build an operating system with a few people, but it takes much more to
run a 500-gigabyte service with a user
population of several million page impressions per day. You also need a few
computers, a few T1 lines, and, unfortunately, a few lawyers to stay on top of misguided effort by federal
MIS czars to stop access.
In the business world, you start with an idea and some people willing to
take a risk and stand behind the idea. These are the "entrepreneurs,"
people with ideas and the energy to implement them. Those ideas are supplemented by financial capital, which in turn becomes resources
such as offices, computers, and all the other things a real functioning institution needs. The financial
capital comes from angels or venture capitalists, and the entrepreneurs have to convince these people that
they are building a sustainable operation that will have two ingredients:
- An ongoing stream of profits.
- An exit strategy.
The ongoing stream of profits means that the institution has sustainability. The exit strategy means
the investors can get their money back out, either through an initial public offering (e.g., sell their shares to
the general public) or through an acquisition (e.g., sell their shares to another institution, which may in turn
have sold their shares to the public or may be privately owned).
For Internet Public Works, we've witnessed a unique wave of entrepreneurs, people willing to put their
ideas and their time on the line to create something real. But, how to build these projects up beyond the
first stage? You can try to find a corporate sponsor who has an interest in the area, but I can tell you with 5
years of fund-raising for public works projects for the Internet behind me that the process is hit or miss. If
you can impress the corporation with the PR potential, you might win. But, you can't argue the merits of
why your public works project is needed for the public at large. That's just not good enough for corporate
Sometimes things go really right, like with PERL; other times core software
just disappears or freezes functionality.
Look at multicasting, for example. Progressive service providers such as Verio are building
native multicast into their infrastructure. But, what about multicast applications that would allow anybody to start
doing collaborative audio and video across the Internet?
Multicasting started out just fine in the early stage. Lawrence Berkeley Laboratory, ISI, Inria, UCL, and a
host of other groups provided the incubator stage, coming out with a whole slew of software systems such
as a whiteboard, video tools, audio tools, and session directory tools. Many of those efforts have now
stalled, as people leave their institutions and move on to other jobs. Where are multicasting applications
now? Cisco helped start up an Internet Multicasting Initiative, but the initiative is a classic industry group
with relatively high entry costs, a focus on high-priced conferences, and no attention to getting an essential
core of software tools available to the general public. The Internet Multicasting Initiative is building an
industry that can sell things like Cisco's IP/TV, but (like operating systems and Linux) there is a need for
ongoing development tools that are available to a broader audience.
Institutionalizing Open Source
If we need parks and schools and libraries, or their digital equivalent, then relying on
occasional wins when companies see immediate benefit in funding a public works project is haphazard.
Sure, we got lucky in the early part of the century when Andrew Carnegie decided the world needed libraries as
well as steel plants. But, what if there had been no way of funding other public works projects? Do you
think we'd have Lincoln Center if we had waited for Coca-Cola to sponsor it? Or Central Park? Do you
think New York would have had parks at the side of parkways if we had waited for the Ford Motor Company to
The Internet needs a more systematic way of encouraging the creation of public works. Let us posit
for just a bit that such an aim might be accomplished by an Internet Public Works Commission (or by many
such commissions). The goal of this commission would be to attract investment capital and manage that
capital by investing in a series of public works projects for the Internet that
maximize equity, access, and sustainability.
It is my contention that many of the current public works projects on the Internet, such as the Internet
Software Consortium, can certainly become sustainable enterprises in the long run, but requires capital
when they become middle-stage enterprises. How do we attract capital to the building of public works?
That requires a way to give the investors their money back, or else we are talking about donations, not investments. And, I can tell you definitively that
investment dollars come from a different pool (and a much larger pool)
than donation dollars.
Here is one way to do this. The Commission is formed with a pool of capital, say $20 million. Perhaps this
comes from foundations. The goal of the commission is to attract another $200 million in capital from
investors who are interested in investing in Bit Bonds, a financial instrument that has a long period of
return (say 30 years) and is perhaps even guaranteed by the federal government, just as highway
construction bonds are often guaranteed by the federal government.
The commission takes this pool of capital and, with the help of a full-time staff drawn from both the open
source and financial communities, looks at the public works projects that have already been started. The criteria for investment might include:
any of these projects show the potential to move to the middle stage and become sustainable enterprises?
the commission invest in these projects and thus encumber the intellectual property so that it is always
available to the general public?
- Does the enterprise have a way of keeping going despite these intellectual
- Do they have a management team willing to stick with the operation?
- Are there sources of
revenue, akin to the Museum Gift Store, that might be enough to sustain the operation?
In the case of the Internet Software Consortium, they were able to accomplish these goals. The investment
money came from a group of Unix vendors. The management talent came from hiring David Conrad, the
person who created the Asia Pacific NIC, as the executive director of the consortium. The ongoing support
came from selling support contracts to large corporations. However, as former chairman of the Internet
Software Consortium, I can tell you that getting to that point was not easy. Paul Vixie and I spent several
years before David Conrad was lured in and got things moving. During that time, ISC was not able to
advance the DNS to include security, build up the number of ports, or do any of the other special work
needed. ISC spent several-years in middle-stage purgatory before emerging successfully.
In many other cases, the projects just never made it. And, they never made it despite the fact that talented
engineers were willing to work hard on the projects and didn't really care if they might make more working
elsewhere. There just wasn't any place to find capital. This is not some assumption, this is based on
several years of getting email from people trying to figure out how to get their projects going.
How does this commission become more than just a donation? It only takes one web-like success.
Venture capital firms invest in dozens of firms, hoping one will make money. If you look at the dozens and
dozens of public works projects already on the net, it is highly probable that one of those will not only meet
the goals of equity and access to essential infrastructure, but might even make some real money for the
organization running it, and hence for the investors (including the commission). Just think how many hot
dogs are sold in Central Park. Do the commissions from hot dog sales take away from the park's general
mission? Not a chance. Is there an alternative? Sure, Central Park could have been designed as a
corporate sponsored park with admission fees.
Dodging Rat Holes
An Internet Public Works Commission would take a very special confluence of forces to happen, but it
could certainly happen. Investors with both vision and money would have to be convinced that the money
is managed in a way they'll get their money back on Bit Bonds. The staffing of the commission would have
to be done very carefully to make sure that these are people that understand what public works are and how
to work effectively with social entrepreneurs.
There is one easy way for such a project to go wrong, however and that is the dreaded Kitchen Sink
Syndrome. We need more artists on-line. Perhaps the commission should fund more artists. We need
more government data on-line and more people should know how to use it. Perhaps the commission should
sponsor training classes on how to use government data. Our schools have to be on-line. Perhaps the
commission should wire all the schools.
The key to a public works commission is focus. Investments in middle-stage enterprises that have
demonstrated they can move into sustainable operations with an investment in capital are the key. And, if
the commission is hard-headed in focusing on public infrastructure, many of the kitchen sink goals are
going to be easier to attain. Think public works, not a social incubator.
Stimulating Early-Stage Investment
But, there is one advantage of a middle-stage operation like the Commission.
This is what might be needed to stimulate the flow of capital from early-stage investors, like venture capital firms and angels.
Today, if you go to a VC with an idea for an Internet Public Works project, you have no exit strategy. With
no exit strategy, the person money sees a donation, not an investment.
What if a $200 million commission were making strategic investments? You can go to a VC and explain
your plan. Give me $200,000 and I will build essential public infrastructure up to 2 million page
impressions per day. I will then sustain the operation through a combination of revenue opportunities, and
fuel continued growth through an investment from the public works commission, which will also pay you
back your early-stage investment money. A $200 million commission could easily stimulate a $1 billion inflow
of capital from early-stage investors. Coupled with federal incentives, such as a hacker tax credit
which allows people to write off development costs for certain classes of open source projects, the foundations for a financial
system adequate for properly supporting the open source movement (and other public
works projects) begins to take shape.
What might a Public Works Commission invest in? I would hope that
this question would not be answered without some real e-work happening
first, but here are five categories of projects that are currently
- Open data. A huge amount of data, from libraries to government,
is being put on-line in a haphazard way. This must change.
- Operating systems. Any high-school teacher ought to be able
to down-load an OS, slap it on commodity PCs, and begin putting
together a network.
- Network configuration, control, and administration. The set top
box is one way to control your LAN, but you shouldn't have to
license technology from 12 different companies to set up, control,
and administer your computing devices, including your television,
your fridge, and all the other components that form your home
- Core infrastructure. The DNS, the routing arbiter, root
servers, and a host of other key components are absolutely
essential to the proper functioning of the Internet and
these infrastructure components are being funded and operated
in a catch-as-catch-can manner. Big bad.
- Communications. Multicasting, IP telephony, radio on the
net, and IP/TV are all being developed as strictly commercial
products. Core applications (and support for those core
applications in the infrastructure) need to be much more
Shouldn't the federal government be doing all this? Why bother with
intermediate institutions like a public works commission when our federal
government is sitting there looking for a mission in life when it comes
to building up the information highway? There is a well-known
government can fund certain classes of work effectively, but is totally blind to others. Public works
projects in the real world are certainly supported by the feds but so far
our brave federal government silent has been incredibly silent on
the question of building real infrastructure. I tried to send e-mail
to Internet creator Al Gore, but he was busy
hacking some open source html on his
In the real world, public works are created because a few citizens get together and decide that what their
community really needs is a park, or a library, or a museum, or a new parkway, or free concerts for
teenagers. Public works are local. The Internet doesn't change that, it
just changes the definition of what local is. It takes individual action to start
useful public works projects, and this proposal is an attempt to figure out a more systematic way to support
those individuals who have the willingness and ability to make public works
integral part of the Internet landscape. If we want an Internet composed
exclusively of shopping malls like Amazon and bedroom communities like AOL,
then nothing is needed. If we want to sustain the open source movement
and other public works projects, it is time for the revolution to grow
We can wait for the federal government to build our communities, but they
are too busy reinventing their next election campaigns or preparing themselves
for the walk through the revolving door. We can wait for corporations
to magically build our parks. Or, we can do what communities have always
done: combine social capital with human labor and build ourselves the
kind of global village we want to live in.
If the Internet is a bedroom community, a place we go to make our
money and then exit after the gold rush, public works don't make
any sense. If we plan on living with this infrastructure for a long
time, it is time to start planning our infrastructure on a more
systematic basis that allows people to build the parks, schools,
libraries, and other public works that make our cities places
Copyright © 1999, 2000 media.org.