 |
Microsoft retrenches around fat clients
Posted by pbrittan on November 11, 2003 at 07:21 AM | Comments (10)
Microsoft is leading a charge back to the desktop. Will the world follow?
Microsoft is placing its bets that fat clients represent the future of software, not browser-based thin clients. In fact, they are phasing out development of a stand-alone Web browser. Longhorn will have HTML viewing/browsing capability built right into the OS, and there will no longer be an Internet Explorer.
I’m not surprised by this. Web clients have never been strategic to Microsoft, which needs to ensure that the Windows desktop remains supremely relevant. The core idea behind HTML browser applications is that they can run equally well on any client-side platform, which thus makes the specific features of Windows unimportant. There is growing customer dissatisfaction with the ergonomics of HTML for enterprise applications, and a number of vendors are offering “rich client” alternatives. Microsoft is offering .NET Smart Clients as its rich client solution, which also happens to be a fat client architecture that runs only on Windows desktops.
That all makes a lot of sense from Microsoft’s point of view. The big question, however, is what will happen to the many Web applications that exist today and to technologies that are dependent on the browser. If Microsoft wants to encourage a Smart Client architecture, will DHTML be phased out?
"Microsoft wants enterprises to write browser applications that take advantage of Longhorn application programming interfaces (APIs), which means that they won't work on non-Longhorn browsers," [Michael] Silver, [an analyst at Gartner], wrote in a research report last week.
The other, related, question is whether large ASP businesses will follow Microsoft’s charge back to the desktop. Apparently, at least some will:
At Microsoft's developer conference, for example, Amazon.com showed how it has used the Avalon graphics technology and the WinFS file system in Longhorn to create an improved shopping site.
Amazon's chief technology officer, Allan Vermeulen, showed how a Web shopper can do a number of tasks that could not easily be done from within a browser, such as rapidly filtering search results for cameras and viewing a photo of a camera as a three-dimensional object. Information from the e-commerce site could also be easily shared with a person's calendar application.
Google also recently announced a desktop application version of its search service.
Microsoft’s competitors, IBM, Sun, Oracle, and others, have been rallying around the Web-based Portal as the primary means for enterprise application delivery. Microsoft’s moves call into some question the future of the standard Web portal, even as Microsoft has started pushing its own SharePoint portal software hard in the market.
Microsoft continues its strategy to cut off Java from end users.
Bookmark blog post:  del.icio.us  Digg  DZone  Furl  Reddit
Comments
Comments are listed in date ascending order (oldest first) | Post Comment
-
Java should lead the way
This is exactly the territory that Java should be able to grab!
I've been thinking for a while now that there would be a move back to rich clients - they're just so much better for doing a bunch of things. Let's face it - the HTML/web interface has been able to serve some good purposes but it's a really braindead paradigm when you want to craft a really nice and useful app. It's crazy how dark ages it is.
From a personal perspective, I wrote AuctionSieve because I wanted to do more sophisticated and *faster* data mining than the eBay website could ever possibly provide.
Posted by: nevster on November 11, 2003 at 06:05 PM
-
Java should lead the way
When are they going to provide a preview function...
AuctionSieve
Posted by: nevster on November 11, 2003 at 06:06 PM
-
Thick client installation...
A primary advantage of browser based applications is the ability of clients to access the application without having to explicitly install it. For many web applications, this is key since the applications are executed infrequently (or maybe just once).
Lightweight DHTML applications have the advantage that they load quickly, but unfortunately they're notoriously hard to write and buggy.
Java applets were a good idea, but they took too long to download... Java web start is not significantly better in that respect.
To get around the download problem, the browser needs to provide better application building blocks (ActiveX controls for example), and a scripting language (JavaScript) to link these components into useful applications. A standard application framework wouldn't hurt.
Unfortunately, all of the above depends on browser support, and that is not likely to happen. For Java-philes, something else is needed on the client side... sort of a "browser on steroids" as part of the JRE to allow us to develop useful applications that can be transparently installed.
Posted by: johnreynolds on November 12, 2003 at 07:12 AM
-
Java WebStart is competitive
With Java 1.4 adding more features it allows a Java WebStart application to be smaller. Plus more people have broadband access.
It is possible to write a complex client and package it in 200 or 300k. It does require some work but a lot of people are working on generating standard screens by reflection and linking with custom screens, thus reducing the overall application size.
Plus Webstart is only downloaded once, and afterwards only updates are downloaded.
Posted by: ineill on November 12, 2003 at 07:36 AM
-
Thick client installation...
Here's a place we see Microsoft flexing its monopoly muscles again.
Part of the reason both Java Web Start has caught on is that Microsoft resists having Java bundled with Windows and on new Windows-based PCs. Downloading Java in order to run JWS and JWS-deployed applicatoins is slow. (Java also shares some blame for this; the Java download could be smaller and Java on the desktop needs improvement). These same issues affect applets, which must either stick with the relatively ancient Java 1.1 or make users go through the same download pain.
Microsoft's heft also prevents the sort of innovation in the browser that you mention. At one point, Microsoft drove development of the web in their typical embrace-and-extend fashion in order to lock up marketshare. Now that they've completed that mission, they've now shifted gears: Stall the development of the web browser completely.
Developments like this infuriate me. The anti-trust case has turned out to be a complete joke. After being found guilty, only to be rewarded by the settlement, Microsoft is now more emboldened then ever.
Rather than give up in frustration, though, Microsoft's rivals, Sun especially, must redouble their efforts. First, they must admit that Microsoft is right: HTML does have its limitations, and Internet enabled desktop applications (fat clients, smart clients, rich clients, whatever you'd like to call them) can provide a better user experience.
Weblogs here and elsewhere show that Java needs some improvement on the client end. Now, this need may have reached the critical stage.
Posted by: jimothy on November 12, 2003 at 07:57 AM
-
Java should lead the way
That's why MS is making so much noise about the superiority of "declarative" approaches to building UIs, e.g. with XAML. That moves them from C#'s "me too!" message to XAML's "that is SO '90's" retort to Java.
The question in my mind is whether the Java community has or should have a "declarative" response to XAML?
Posted by: mchampion on November 12, 2003 at 11:39 AM
-
Developers call the shots!
Developers can either sheepishly follow along, on the M$ leash, or they can innovate on their own!
Guess which approach keeps Bill Gates up at night?!
Take control of your destiny!
Please visit:
https://cajo.dev.java.net
Timing wise, we are at a critical juncture.
John
Posted by: cajo on November 12, 2003 at 03:05 PM
-
Web clients were always doomed
Web clients never really had a chance.. They never worked well and have always been an awkward hack. A web browser is for browsing the web.. not the UI for some other application. Thank god that the awful trend is finally reversing.
If you need to provide a UI via the browser - Java Applets or Web Start applications have always been the best choice. Or Active X controls for those that enjoy the occasional virus.
Posted by: swpalmer on November 13, 2003 at 08:47 AM
|
java.net RSS Feeds
