Jason Kottke offers a clear and concise discussion of how the underpinnings of personal computing are evolving. In essence, we’re moving from a PC-centered to a network-centered world. More and more applications will run in browsers (or browser-like platforms such as Google’s Desktop, Yahoo’s Konfabulator and Apple’s Dashboard) rather than as traditional, standalone programs installed on individual hard drives, and whether your data resides locally or on some distant server will ultimately be both transparent and irrelevant. As Kottke shows, most of the dominant players in PC software (at least on the consumer side) are embracing this vision, happily or fearfully.
Kottke notes that one of the hurdles to this model is that the browser provides a primitive user interface, lacking the robustness and responsiveness of standalone applications. (Whatever the faults of Word, do you really want to do your word processing in a browser window?) He quotes Paul Graham as arguing that while the browser interface has a lot of problems, it’s “good enough” for most people’s purposes. That’s probably an overstatement at the moment, but it won’t be an overstatement for long. The browser interface – or what I think would be better termed the utility interface – is going to advance rapidly, as many companies compete for control over it.
A couple of glosses on Kottke’s thinking. First, he doesn’t pay enough attention to the interests of businesses in the evolution of personal computing. Photo-sharing, music-playing, and blog-reading applications are certainly important and interesting, but most people’s computing decisions are determined as much or more by the needs of their employers as by their own personal interests and diversions. Cost, security and reliability concerns will ultimately lead companies to demand even less local data and applications than Kottke imagines. The user devices will become ever thinner, as applications, data and even the user operating system resides in central servers. Whether you’re using Word on Windows or Google Word Processing on Google Desktop, it will all be supplied over the Internet rather than from the bowels of your own PC. Second, and related, Kottke’s assumption that a lot of content needs to be replicated locally (so you can continue working when, for instance, on a plane) reflects the current gaps in connectivity. Those gaps aren’t going to be around forever. In five or ten years, we’ll have ubiquitous high-speed network connections, and at that point the need for local data, apps and web servers goes away. Your personal desktop, residing entirely on a distant server, will be easily accessible from any device wherever you go. Personal computing will have broken free of the personal computer.
I agree with the gist of this article, but I don’t think we need to wait for the browser to improve as a UI engine. We already have the tools we need to deliver “rich applications” over the web (e.g. Java Webstart, Flash, etc.) These applications can be located and launched from a browser but they don’t have the UI limitations of a browser. HTML and HTTP are good at what they are designed for (transferring and rendering text and images) but are really not designed to generic UI protocol in the way, say X, was designed to be. I think its time to stop pushing that square peg into this round hole and move on with our thinking.
“agree with the gist of this article, but I
don’t think we need to wait for the browser to improve as a UI engine.”
Yes, I agree with this statement – but, we still are in the age of service-oriented-server-side-applications beyond flash and browser plug-ins.
Have a look on sharepoint. It is not the best application I would recommend but it offers the possibility to open documents from a server in word or any other office application on the pc and save it back to the server [without saving it on the pc or uploading it manually!].
More and more so called rich clients will call their data from and save it back to a server over the internet. Also the format of the files will change from a proprietary format to XML (like word can do today and will do by default in the next version :-).
The service oriented architecture will enable rich clients to communicate with servers and hopefully products designed with a rich client will start to use this possibility!
Most of Google’s applications run exclusively on Microsoft Windows. Many of Google’s applications install locally and connect remotely for data. Google understands the easy “click-once” deployment concept well. When Microsoft launches it’s new Microsoft Windows Vista OS by the end of 2006, it will become possible to build Rich Internet Applications (RIA) faster and better than ever before. Internet applications can look and feel much like local applications, and they will not need a browser. It is possible Google will take advantage of these capabilities to further reduce the need for local installation. By taking advantage of new OS features, the application needs not be installed on your PC…
I guess before we can discuss where the OS is, we’ll need to define what the OS actually is. That gets harder every day. Certainly, the local OS is still evolving, to better support the needs of the “Internet OS”