|
Is the rediscovery of the desktop just the latest swing of some tech-trend pendulum, or is there something more going on here?
August 02, 2007
URL:http://drdobbs.com/web-development/return-of-the-desktop/201202675
This year, some of the big boys gave every impression of having suddenly and simultaneously remembered that there is such a thing as a desktop. Google got Geared up, Adobe announced AIR, and Microsoft saw the light with Silverlight, all of which are tools to help web developers integrate operations on the Web and the desktop just a little better. That oft-repeated mantra that the web browser is the new operating system? In 2007, not so much.
But is this rediscovery of the desktop just the latest swing of some tech-trend pendulum, or is there something more going on here? While it's probably too early in this shift to make any grand pronouncements, we can at least trace its history and identify some of the key players.
The pendulum has been swinging away from the desktop for some time now.
All the way back in February, 1995, not long after Marc Andreessen and Eric Bina unleashed the first visual web browser, no less an authority than Bob Metcalfe, the developer of Ethernet, predicted that the web browser would become the dominant operating system for the next era of computing.
He wasn't alone. Pundits and academics rushed to assert that the desktop operating system was dead, replaced by the browser. This was, of course, nonsense if taken literally, but it was not meaningless. The very real phenomenon that these observers were pointing to was a shift in the center of gravity of computing, from the desktop to the Web.
The swing away from the desktop continued unabated in the succeeding decade-plus. Initially, it was all about end users, about the Web as a place to find files, information, pictures. It was static websites and file downloads. But then the Web began to be a locus of application deployment as well (beyond mere HTML coding of static web pages, that is). There were clear advantages to putting the app online. For one thing, you could ensure that everyone was using the latest version. Developing for the Web meant using scripting tools, primarily, which translated to rapid deployment. Given the speed of web connections, this did not generally result in unacceptable performance.
Adobe Flash-enhanced websites were everywhere, and the term "Ajax" was coined to capture the collection of technologies behind most of the new web apps. These things began to be less like web features and more like real applications that happened to be hosted on the Web. Google's web apps, we were told, threatened Microsoft Office, despite the fact that Google invariably insists that it doesn't use Microsoft as a guide in developing products and services.
A decade into its life, the Web became a locus not just for application deployment but also for application integration, leading to a new kind of application. In 2005, Paul Rademacher found a way to integrate Google Maps with Craigslist, convincing Google to open its APIs. This led to an ecosystem of mashups, applications assembled dynamically from online components. Small pieces loosely joined, in David Weinberger's apt phrase.
And now there is something new, or at least nuanced. The rediscovery of the desktop could be described as doing homage to the Microsoft model of (desktop) software plus (web) services, rather than the Web-replaces-desktop model of the web apps. (Microsoft's term "software plus services" is suggestively agnostic regarding the relative importance of the Web and the desktop, and tells you something about Microsoft's priorities.)
But the new thing brings with it a different center of gravity. The Web is still the hub, but the desktop is acknowledged. The idea seems to be, yes, use the user's desktop and computer and filesystem, but use just what is needed, and no more. And the emphasis in development remains on scripting, at least for the GUI.
"Break free of the browser" is a telling phrase in the age of Web 2.0. Joyent's Slingshot, essentially a VM for Rails apps, uses that phrase. Despite the benefits of the Web for at least some kinds of applications, Joyent realized, there is a need for a desktop connection. The decades-plus barrage of hype for the Web as a platform makes this phrase seem all the more telling: The companies in the back-to-the-desktop movement aren't anti-Web, but they have seen its shortcomings clearly enough that they welcome such language. Joyent's technology lets Rails developers create apps that work the same online and off, with seamless drag and drop across the membrane. Corel uses Joyent's technology for its line of integrated web/desktop apps. Zimbra (online/offline e-mail) and ThinkFree online/offline office apps) are two more companies riding this Web-centric but desktop-empowered trend.
And then there's Dojo. "Imagine if web applications could store megabytes of data on the client side...persistently and securely." Dojo tackled that challenge, to deliver an API to provide JavaScript web apps with offline storage capability "whether this JavaScript is in a browser, a Firefox plug-in, an ActiveX control, using Windows Scripting Host, etc." Cleverly, dojo.storage automagically detects available storage options and selects the most appropriate one.
It's free, it's open source, and it consists of just two pieces: a JavaScript library bundled with your web page and a 300K cross-platform, cross-browser component to cache your web app's UI for use offline. But they now find themselves swimming in deep waters with some big fish.
Then Google announced Gears. Dojo, its developers were quick to point out, will be recrafted to work with Google Gears, Dojo's old storage mechanism deprecated almost before it was seen.
Gears consists of three more or less independent components that plug the holes in online apps. The three components are LocalServer, a local proxy server with a JavaScript API; Database, JavaScript-accessible SQLite local database storage; and WorkerPool, a mechanism for running web app JavaScript code in the background, so it doesn't block browser script execution. Just what you need on the desktop and no more: It's clear that Google was shooting for just that target with Gears. Very cool, early observers have called it.
Gears aims to removes all the arguments against web apps by giving them the needed offline hooks, and it doesn't tie them to any other Google technology. It's open source, platform neutral, browser neutral, and free as in beer.
The first Google Gears app is Reader, an RSS reader.
In June, Adobe rebranded its Apollo download as AIR. Adobe already has category-defining tools for both desktop and web development, and AIR leverages these, letting developers use Flash, Flex, HTML, and Ajax skills to develop web apps with desktop hooks.
Like Gears, AIR brings an offline component to online apps, but it doesn't require a browser. Apps created with AIR can run independently of any browser and have all the status of "real" apps on the desktop.
Although it's not entirely open source, key elements of it are, including the Webkit HTML engine, the VM, and SQLite local database capability, and the runtime is free. Although Adobe didn't initially plan to include a database component, the company changed its mind on seeing that this was the missing feature most cited by developers after the alpha release. AIR's APIs will work with Gears' APIs, Adobe assured developers, and Dojo will be as supported a development framework as Adobe's own Flex. Those who have been playing with it are impressed and imagine that entirely new classes of apps will be created with AIR.
Although Microsoft's high profile Silverlight tool is being portrayed in media reports as in the same mold as Gears and AIR, it is probably fair to say that Microsoft has different goals than other companies. Microsoft is so invested in the status quo that it can hardly embrace a truly iconoclastic paradigm-threatening strategy as a corporate policy. When Bill Gates or Ray Ozzie lays out one of those grand visions for the company, as in 2005 when Ozzie declared the company's business at risk if it didn't change direction, he's really trying to steer an ocean liner. So it appears that Microsoft is still following some sort of software-plus-services map, and still working intently on Live thisandthat under the 2005 course correction, notwithstanding the fact that the market is sending signals that the road has shifted under the map.
That said, no company is more competitive than Microsoft, and it is looking for how to compete as the Web becomes more of an app space (hosting, running, developing). Gears is a signal that the model Microsoft was competing against or hoping to embrace and extend or to leapfrog or whatever—has changed.
Silverlight, announced early this year, is a tool for developers in the realm of the Web, although not so clearly a tool that forces a new way of developing. This is not to say that it can't be used to create new kinds of software, with a different relationship between the Web and desktop. Those who have seen the presentation use language like "Flash killer" and "Microsoft has rebooted the Web."
Silverlight will work with Gears and AIR. It is proprietary, cross-platform, and cross-browser. It is a runtime for browser-based apps and provides the developer with a subset of Windows Presentation Foundation capabilities for animation, vector graphics, and video playback. It also includes the .NET CLR, which means that developers can work in any .NET language, which is a big deal.
The bottom line is that Silverlight is a very impressive tool that will probably lead to new classes of web/desktop software; compete with Flash, Ajax, and JavaFX development; and shake up web development. It just won't shake things up in the same way that some of these other tools will.
And what about iPhone? It's an obligatory question these days, but a good one nevertheless. So far, it appears that Apple does intend to force developers to embrace a new way of developing software, or at least a particular way, when it comes to iPhone. The platform will support no Java, and no Flash. Any phonetop hooks that are provided will be Apple's own, and web apps will run in the browser, not as empowered first-class apps. At least, that's the way things looked in the first few hours after launch. It is not clear that something like Sun's JFX won't come to iPhone eventually, but Jobs strongly stated that this is not Apple's direction.
Perhaps the most interesting thing about the return to the desktop is that almost everyone seems to have learned the lesson of mashups. (Which is what, exactly? That the market demands that web tools work together? Something like that, I suggest.) It's a love-fest of cooperation: Gears, AIR, Silverlight, Dojo, and Slingshot all empowering you to create new classes of software by working together in some combination. Using SQLite. Just not on your iPhone.
Terms of Service | Privacy Statement | Copyright © 2024 UBM Tech, All rights reserved.