We'll see | Matt Zimmerman

a potpourri of mirth and madness

Embracing the Web

The web offers a compelling platform for developing modern applications. How can free software benefit more from web technology, and at the same time promote more software freedom on the web? What would the world be like if FLOSS web applications were as plentiful and successful as traditional FLOSS applications are today?

Web architecture

The web, as a collection of interlinked hypertext documents available on the Internet, has been well established for over a decade. However, the web as an application architecture is only just hitting its stride. With modern tools and frameworks, it’s relatively straightforward to build rich applications with browser-oriented frontends and HTTP-accessible backends.

This architecture has its limitations, of course: browser compatibility nightmares, limited offline capabilities, network latency, performance challenges, server-side scalability, complicated multimedia story, and so on. Most of these are slowly but surely being addressed or ameliorated as web technology improves.

However, for a large class of applications, these limitations are easily outweighed by the advantages: cross-platform support, instantaneous upgrades, global availability, etc. The web enables developers to reach the largest audience of users with the most compelling functionality, and simplifies users’ lives by giving them immediate access to their digital lives from anywhere.

Some web advocates would go so far as to say that if an application can be built for the web, it should be built for the web because it will be more successful. It’s no surprise that new web applications are being developed at a staggering rate, and I expect this trend to continue.

So what?

This trend represents a significant threat, and a corresponding opportunity, to free software. Relatively few web applications are free software, and relatively few free software applications are built for the web. Therefore, the momentum which is leading developers and users to the web is also leading them (further) away from free software.

Traditionally, pragmatists have adopted free software applications because they offered immediate gratification: it’s much faster and easier to install a free software application than to buy a proprietary one. The SaaS model of web applications offers the same (and better) immediacy, so free software has lost some of its appeal among pragmatists, who instead turn to proprietary web applications. Why install and run a heavyweight client application when you can just click a link?

Many web applications—perhaps even a majority—are built using free software, but are not themselves free. A new generation of developers share an appreciation for free software tools and frameworks, but see little value in sharing their own software. To these developers, free software is something you use, not something you make.

Free software cannot afford to ignore the web. Instead, we should embrace the web more completely, more powerfully, and more effectively than proprietary systems do.

What would that look like?

In my view, a FLOSS client platform which fully embraced the web would:

  • treat web applications as first-class citizens. The web would not be just another application, represented by a browser, but more like a native application runtime. Web applications could feel much more “native” while still preserving the advantages of a web-style user experience. There would be no web browser: that’s a tool for legacy systems to run web applications within a compatibility environment.
  • provide a seamless experience for developers to build web applications. It would be as fast and easy to develop a trivial client/server web application as it is to write “Hello, world!” in PyGTK using Quickly. For bonus points, it would be easy to develop and run web applications locally, and then deploy directly to a PaaS or IaaS cloud.
  • empower the user to manage their applications and data regardless of where they are hosted. Traditional operating systems act as a connecting fabric for local applications, providing a shared namespace, file store and IPC mechanisms, but web applications are lacking this. The web’s security model requires that applications are thoroughly sandboxed from each other, but a mediating operating system could connect them in meaningful ways, just as web browsers store cookies and passwords for various websites while protecting them from each other.

Imagine a world where free web applications are as plentiful and malleable as free native applications are today. Developers would be able to branch, test and submit patches to them.

What about Chrome OS?

Chrome OS is a step in the right direction, but doesn’t yet realize this vision. It’s a traditional operating system which is stripped down and focused on running one application (a web browser) very, very well. In some ways, it elevates web applications to first-class status, though its paradigm is still fundamentally that of a web browser.

It is not designed for development, but for consuming the web. Developers who want to create and deploy web applications must use a more traditional operating system to do so.

It does not put the end user in control. On the contrary, the user is almost entirely dependent on SaaS applications for all of their needs.

Although it is constructed using free software, it does not seem to deliver the principles or benefits of software freedom to the web itself.

How?

Just as free software was bootstrapped on proprietary UNIX, the present-day web is fertile ground for the development of free web applications. The web is based on open standards. There are already excellent web development tools, web application frameworks and server software which are FLOSS. Leading-edge web browsers like Firefox and Chrome/Chromium, where much web innovation is happening today, are already open source.

This is a huge head start toward a free web. I think what’s missing is a client platform which catalyzes the development and use of FLOSS web applications.

About these ads

Written by Matt Zimmerman

July 26, 2010 at 10:43

7 Responses

Subscribe to comments with RSS.

  1. nuBuilder is a great FLOSS.

    I think its ideal for creating the next generation of web-based back office applications -(sure I wrote it but we’ll happily offer a full money back guarantee to anyone that doesn’t like it)

    Steven Copley

    July 27, 2010 at 00:03

  2. So does this mean you are going to be pushing for UbuntuOne server source code to be open sourced?

    Corey Burger

    July 27, 2010 at 01:30

    • I agree, and I don’t think this can be understated. Ubuntu One is neither open nor federated; it “breaks” the web and the Ubuntu experience.

      It’s the kind of web service we need to get away from, and encourage people to make open and federated replacements for. Which I’m very much in favor of.

      Jared Spurbeck

      July 27, 2010 at 23:06

    • That is not what this post was about, and I don’t want to (re)open that particular discussion here on my blog.

      I do think that free web applications are important to the advancement of free software, and I also believe that profitable businesses can be built around them.

      Matt Zimmerman

      July 29, 2010 at 11:00

  3. [...] been via Glyn Moody’s endless twitter stream — I ran into Matt Zimmerman’s "Embracing the Web". Another "wait, what? we do that." To some extent, KDE’s project Silk aims to [...]

  4. Good idea. What about free servers for these apps? free servers supporting all the languages and frameworks which make these apps so wonderful. If I host on my own server, what if I die, can no longer afford hosting. What becomes of my app?

    allenbeme

    July 28, 2010 at 13:29

  5. [...] and imagination. This includes considering how we package and distribute software, how we adapt to technological shifts, and highlighting opportunities to cooperate with other open source [...]


Comments are closed.

Follow

Get every new post delivered to your Inbox.

Join 2,245 other followers

%d bloggers like this: