Time for a Web OS?

Multics, Unix, OS/360, Pick, George3, RSTSE, RT-11, RSX-11M, CP/M, MP/M, PC-DOS, Macintosh, OS/2, Windows, OS X, Linux, Beos. That’s a very short list of operating systems past and current for mainframes, minicomputers, early PCs, later PCs, network servers, clusters and PDAs. So what’s left to stick an operating system on?

How about the Web? The idea of a Web operating system has been bandied about for some time, and a few systems are appearing.

The central idea of a Web operating system is to provide a desktop (or Webtop) in a Web browser that is similar in function to a PC desktop. To be an operating system, it needs to have applications (or Weblications, as they have been called) along with management tools. Exactly what constitutes a Web operating system is debatable, but in general at the client end a Web browser is responsible for rendering the user interface and handling user interaction as well as performing some or all of the processing.

At the other end is a remote Web server that provides data storage and retrieval for the client application that includes configuration and working data as well as database support, higher-performance processing when necessary, access security, authentication and any other services that the client hasn’t got the oomph to provide.

But Web operating systems are about more than simply spreading the processing load around. They are also intended to provide a consistent, centrally updatable personal-productivity environment no matter where you are, what type of computing hardware you’re using or what type of client operating system the browser runs on.

Whatever browser you use is supposed to act as a universal abstraction layer hiding the ugly realities of the client. So the Web operating system’s goal is to provide a personal-computing environment that runs anywhere, on any browser. This sounds like a promising idea, but there are drawbacks. For a start, the practicality of the solution depends on the reliability and availability of the remote services, and you have to have enough bandwidth to make usage about as frictionless as it would be using a regular PC operating system and regular applications.

Some critics of Web operating systems (and, for that matter, software-as-a-service) argue that using a Web-based service of any kind requires a level of blind trust that is inconsistent with IT objectives. We’re not sure this is a good argument, because that same kind of trust is required when we entrust our business processes to regular software.

What’s the commercial drive for Web operating systems? The anytime, anywhere access to a complete, low-cost personal-productivity environment, along with no installation, backup and upgrade concerns, means there’s a great fit for the mobile-worker market and for consumer use (particularly as a replacement for Windows and Microsoft Office on low-end PCs).

It is essentially the same argument that applies to any software-as-a-service offering with the added benefit of integrating two or more personal-productivity applications into a manageable system.

We suspect we won’t ever see rational companies or individuals trust all of their personal or corporate productivity and data to a single, purely online, Web-based solution. What will most likely happen is we’ll see hybrid systems where we can move and synchronize our content between remote servers, desktops and mobile devices more or less transparently. We’ll also use browser-based Web applications and Web operating systems that are smart about available local processing and storage capabilities and network bandwidth to provide optimized user experiences.

But even though that nirvana is some time away, there are some beta and Version 1 attempts out now that look promising. Next week we’ll explore what a Web operating system looks like today.