Can free software transition to the web services world?

December 18th, 2009 in open source

Etherpad is being released as open source software because the team is moving onto Google Wave. As an open source web hosted project, without a company behind it, it is unlikely to succeed.

Back when I used to talk about the business reasons for open sourcing code, "end-of-life" was always one of my first examples. People think, "oh, I have this project that I no longer want to work on but people still want to use it, so I'll open source it!" They have the (often mistaken) idea that somebody out there will just start working on it.

This is exactly what Etherpad is doing:

Our goal with this release is to let the world run their own
etherpad servers so that the functionality can live on even after we
shut down etherpad.com.

They are likely to lose a lot of users immediately as hosted software is very hard for end users to use without hosting.

There is a real need for hosting of open source web services.

Open source web service projects that have been successful like WordPress and Drupal have companies that can host it for end users. Without hosting, I think it's unlikely that projects like Etherpad will be widely successful. And we should now ask what happens to the companies behind WordPress and Drupal when they get acquired?

The other type of successful open source web service are projects started by companies and open sourced as part of the business strategy for creating community or
encouraging developers to help them out. In these cases the code is
often written entirely by company employees. And the hosting is done by the company as part of that same business strategy. Projects like SugarCRM and
Alfresco are successful open source web services but if the companies
went away (or were acquired), the projects would have a hard time
living on in the open source ecosystem.

Hosting for free and open source software projects is hard.

While it's virtually free (other than labor) to start a free software project, hosting a web service costs money, especially if it becomes popular. And consumers are used to free web services. (Web service companies often make money off of advertising (on the free version) or by charging for a supported version or a version with more features or by hoping to be acquired some day.)

So if it takes a business model right now for web hosted projects to be successful, how do free software projects transition to the web services world? How does a project start as a free and open source software project and transition to a hosted web service without becoming a company?

I don't think it's enough to just make the hosted version available for people to install and use on their own. I think we have to figure out a way to also host the software. If you require people to host the web services piece themselves, you are confining your market to technical people or companies with IT staff. The project might become successful but it's unlikely to be used by the average end consumer as part of their daily life.

So how do free software projects develop and host web services? Do they use business models like advertising to be self sustaining? Do they start foundations like the GNOME Foundation that will run as a nonprofit but make enough money of the hosted version to at least cover expenses? Or will all hosted services be essentially startup companies done by free software developers hoping to create a successful company?

Can we have self sustaining web service free software projects?

16 Responses to “Can free software transition to the web services world?”

  1. Kim Weins says:

    Interesting topic Stormy. Maybe the cloud can come into play here. If the open source project could be “ready to deploy” at the push of a button into the cloud, then one or more companies could use it. Obviously there would be a cost, but that cost would be in a pay-per-usage model which might be attractive for some potential users.

  2. Sander says:

    It would be great to see more free software web apps. I hope to see GNOME moving more towards the web: a lot of apps would be much more powerful there, and there is a need for a well integrated web app suite following the GNOME philosophy.
    Mozilla is doing interesting things with regard to web apps lately. They are working on Bespin, a mozilla.com-hosted IDE that can quite easily be hosted on your own pc too, and Raindrop which seems to follow the same model. I think a lot can be learned from Mozilla.
    Also, in Ubuntu there’s work on CouchDB and Ubuntu One. They make it easy to build (web) apps that run on your own pc and store their data there, syncing it with your free Ubuntu One account. (Users only pay for extra storage.) It’s not really free software, but these people might have good ideas about how to shape a more viable free web service world.

  3. Sandy says:

    One other thing to note is that Etherpad is not trying to build a community here. They are pretty much throwing the code over the wall, so to speak.
    This is exactly what they were asked to do, and it is a great thing that other folks will be able to use their hard-earned technology. I applaud them for it, and in fact I may try to use it in some of my own projects.
    But let’s not compare this to the idea of community-developed free software surviving in web app form. We’ll have to find other case studies for that.
    Obviously, I’m hoping that Tomboy Online will be one such case study…but we have to launch it first before we can really figure out how screwed we are. ;-)

  4. stormy says:

    I too am rooting for Tomboy Online! How do you plan to do the hosting?

  5. Sandy says:

    We really don’t know yet. Either sharing a server with some like-minded projects, or hosting on GNOME servers if that’s possible.
    The things is, I have no experience with deploying web apps, so I can’t reasonably estimate hardware requirements, or even assert anything about security. We’re going to have to learn these things as we go.
    I was thinking of hosting it myself until we understood these things better, but with recent encouragement from the GNOME infrastructure team, we might try there from the get-go.
    We still have a little work to do on the Snowy software before we can think about deploying even an alpha of Tomboy Online, though.

  6. tonfa says:

    Or maybe people could learn to self-host, that was the idea behind internet. Launching a local instance of etherpad is easy, and it is usually used by few people to collaborate on a document in a short period of time.
    So it’s find, launch a local server and you’re good to go, you’re already paying for internet access, why would you need more?

  7. tonfa, the problem about that is IPv4 and firewalls-installed-by-default. When IPv6 happens and people start trusting their OSs and applications rather than relying on network security systems, that may be feasible.

  8. Rym says:

    This is the idea behind http://autonomo.us/ of course.
    I wonder how long before the open sourced Etherpad gets Wave support ;)

  9. stormy says:

    A lot of technology would have to improve for people to be able to self host. Most people I work with that don’t work with technology have a difficult time logging into a new blogging service much less hosting it themselves.

  10. Don Marti says:

    There has to be a profitable middle ground between VPS (you’re root, in control of and responsible for everything from kernel updates to spam filtering) and a basic web hosting plan with PHP, MySQL, and not much else. I’d pay for a plan that gave me git, ikiwiki, screen, and a reasonable selection of up-to-date tools, well-administered, but not root. Once you had that, you could deploy (or let users share) canned offerings of a variety of web apps.

  11. Tom says:

    I think about this all the time with Gourmet (http://grecipe-manager.sourceforge.net), the free and open recipe program I develop. If Gourmet is to have any chance of not becoming obsolete in the next five years, it seems it will have to move to the web. There are also numerous features that I could begin implementing if I went to a web-hosted architecture.
    That said, I write Gourmet for fun, for free. Sourceforge made it possible for me to share my work at no cost to me and as a benefit to the broader community. What service would make it possible for me to do the same with a web application? How could we build such a service?
    Otherwise, it seems impossible for devoted programmers to get their work out there in this new web-based world unless they also want to be businesspeople. It’s not even that I’d need a business plan to make a lot of money, I just can’t afford to take a loss to distribute software that, fundamentally, I develop for fun.

  12. Www says:

    Arguably, I think wordpress’ great success is not that it has Automatic behind it but rather that it’s *incredibly easy to install*. The answer for sustainable web services in the open source world isn’t finding a way for companies to take responsibility for their contributions, but rather to make it possible for normal folk to run the software on infrastructure they already own or can easily acquire.
    We don’t just need the etherpad code, we don’t just need some company to run the code, we need to write software–web based or otherwise–that non-technical folks can administer with relative ease.

  13. stormy says:

    Let me know how I can help!

  14. stormy says:

    So I do do some self hosting through DreamHost. But I still think the way it currently works requires too much technical knowledge for the average user.

  15. stormy says:

    I’d love to see a web hosted version of Gourmet – even if it’s just the data. I want to share my recipes with my SO at work, with the kitchen in the computer, with my friends.
    I’d love to talk to you about the idea.

  16. concerned and optimistic says:

    The solution: Distributed Self Hosting (DSH) a cross between Wuala and Opera Unite.
    An easy to install and use application to host the web application. This would take of everything related to hosting the applications. It would take drag and drop web apps. It could even use an installation system like the Linux Distros or an AppStore like system.
    This is very important: the user would control everything. No censorship, no privacy issues, security(encryption), the user data would be held hostage or delete.
    This last point is very in regard to the recently announced GoogleOS. GoogleOS will store any data permanently on the device. If the user stops paying the user data is held hostage; if the user don’t pay for x amount of time the data is delete. Ransomware is born!
    As for Open Source Web Apps (OSWAs) financing models the “web app model” is the ideal opportunity. users could self host or for x amount of money use the third party.
    Even Google have pay for versions of their apps, as several other companies: Dropbox, UbuntuOne, Flickr, etc. In fact the future will be very bleak for free apps.
    WE NEED FREE SOFTWARE/OPEN SOURCE applications NOW!
    Everything from notetaking applications like BasKet or the above mentioned Gourmet to games could self financing by creating parallel local/web versions of their software and give the users a way to pay for their software and guaranteeing a faster development.
    Right now projects are looking for models of financing: donations, pledjies, bounties, etc…
    The open source web app ecology has been an wasted opportunity. While most of the infrastructure of the Web is built on open source software, from web servers to programming languages, most applications built on them are not. Those that are, are either too are to install to the average user or too simplistic.
    Some are trying to overcome this buy distributing them in Virtual Machines (Appliances) but this is far the ideal solution as the applications will much, much bigger in size and consume much, much more resources without any need.
    An web app in 5MB or less can be using this system grow to several hundred megabytes. Really poor solution, only good for testing.
    As I said what is needed is a small, fast, easy to install and use Host App. Distributed. Something like Wuala for web apps.
    The the web apps themselves would be distributed as self contained files that a user would just drag and drop inside the Host App. Then a wizard would appear to configured that app and then is ready to go.
    With the technologies in HTML5 the browser could be that Host App. Google, Mozilla and Apple are adding these technologies rapidly. And they are for the most part open source.
    It’s time for open source to stop playing catch up to the proprietary world and lead.