The desktop transformed user experience through the last 20 years and has been the entry point for applications for a long time now. Come to think of it, even the browser is an application that resides in the desktops of users. The browser helps us access sites that run applications and that may be an aggregation itself, like a portal. The web applications, which traditionally were database centric and form, based data entry and reporting, email and ecommerce type applications are maturing into online office and collaboration applications (the offerings from Google), online gaming and online drawing applications (Gliffy). Interestingly enough though, the desktop has not changed fundamentally in its many years of existence and is still focused on file management, document creation and deletion, and local storage of documents type of functionality. Of course there has been the inclusion of web technologies into the desktops. More recently, Microsoft and Adobe (with its Adobe Apollo) may be examples of companies that are bringing the concepts and technologies of the web to the desktop. An important question then is, what is the future of the desktop and that of the web? Would the desktop be replaced by the web or would the web pervade throughout the desktop? These and related ideas were brought-up by David Temkin (www.davidtemkin.com), the co-founder and CTO of Laszlo Systems as a part of his presentation at the AjaxWorld in New York. David seemed very bullish about the web outperforming the desktop in being the center of user experience in the long run. Guess that is why, Laszlo is taking the desktop to the web with its Webtop offering and betting on it being a preferred solution going forward. Of course they are not the only ones doing it neither are they the first to do it. Laszlo Webtop is a solution that provides the infrastructure to build a web-desktop or a “webtop”. It leverages Java on the server side and is built to take advantage of the established server side java frameworks and mechanisms. (I tried the initial release of this product and maybe would talk about it sometime later or try and compare and contrast it with similar offerings from other vendors.)
Now we are not sure if the Laszlo Webtop is the winner in the long run and I am not trying to present David Temkin’s presentation verbatim here, but I think the questions he has raised and his opinion on the matter are indeed something for us to ponder about. Considering that java developers may still be struggling between old Apache Struts type web frameworks and the numerous hybrid choices that AJAX/RIA have presented to them, are they thinking about where the web-desktop convergence is going and the role they want to play in shaping this convergence’s destiny. Or is it that some of us are happy that such concepts could theoretically be partially or fully implemented as Applets even some 10 years back. Or is it that we are waiting for a JSR to be initiated for such a purpose :)
Talkib about web-desktops, and doing this using server-side Java, Pramati's Dekoh platform takes it a step further by bringing the server-side Java right onto the desktops. Unlike most other web-top approaches that focus on "bringing web to the desktop", Dekoh does this well and does the reverse equally well- of bringing the desktop also onto the web. Just as webtop apps can access resoucres/services over the web, the apps running on the desktop can also be accessed from anywhere on the web (subject to security permissions being set).
Essentially, bridging the web desktop divide by blurring the lines both ways. Web into dsktop. Desktop into web. (www.dekoh.com)
When I cut my teeth as a Unix sysadmin 16 years ago, I quickly learned that the distinction between a workstation and a server is only one of arbitrary configuration and nomenclature. In fact microsoft's fixation on maintaining the myth that servers and desktops are fundamentaly different things has impaired their products.
I think the way forward is to realise that modern desktops are just as good at serving up web applications as web servers are. Why do we need to distinguish between web development and desktop development? If my documents are on my workstation, why can't I access them via the web? Why should I care where they are stored?
Yes I want to take advantage of the power of my workstation and hence local apps, but realy that only means local execution. I don't want my workstation's network connection to be essentialy a one-way link. Forget Client-Server, the future is peer-to-peer for data and applications. I don't expect this to happen for a while, but ultimately it's the only way to go.
Simon, I agree with you that these arbitrary borders need to go away. However, there are very hard problems to solve. Data replication can lead to nasty logical issues and complex architectures, because as soon as you have two copies of some piece of data, they tend to diverge and you need a mechanism to merge them. If you put in place some kind of locking, you impair concurrency and you incur all kinds of nasty issues like deadlocks. Restricting access to stuff according to some possibly complex organisational ruleset is also hard if you have physical distribution and distributed responisbility for security. So I think, it's very easy to allow access to your own personal data on your desktop from anywhere. That's a solved problem. But it's hard to have security and data consistency on a level that's required for some (not all) kinds of data.