Linux Today: Linux News On Internet Time.
Search Linux Today
Linux News Sections:  Developer -  High Performance -  Infrastructure -  IT Management -  Security -  Storage -
Linux Today Navigation
LT Home
Contribute
Contribute
Link to Us
Linux Jobs


Top White Papers

More on LinuxToday


Editor's Note: Agnostic Apps Make Desktop Arguments Moot

Jul 29, 2005, 23:30 (1 Talkback[s])
(Other stories by Brian Proffitt)

By Brian Proffitt
Managing Editor

I think I have related this story once before, but I will do it again:

When I was a lowly development editor working for Sams Publishing, I had the opportunity to travel to one of Netscape's conferences in Manhattan to do some background research on the technology and (hopefully) find some authors to write for Sams.

I found a couple of authors, but the one thing I remember most of all from that trip was this statement made by the keynote speaker: that one day Netscape would eventually replace Windows as the desktop platform of choice.

A browser? Replace Windows? Being younger and stupider, I scoffed at the notion. Desktops were synonymous with operating systems, and the Netscape API, no matter how robust, was no operating system.

(Keep in mind these were the Dark Times, before my introduction to Linux.)

This week, I was reminded yet again just how dumb I can be, when I moved to a new e-mail application and a new home page and found more evidence to dispel the notion that desktop=OS. They are not the same at all, Microsoft is about to suffer for it, and Linux will reap the benefits.

First, I learned that Google had improved its personalized home page so you can add your own RSS feeds to the page. Since they don't offer LT as a standard offering, I was not interested in it until this functionality was offered. Now that it has been, I can slap a multitude of information on my new Google page and digest it all in one fell swoop. (Note to Google: find a way to add my favorite comic strips to the page and I will be happier than a pig in--well, just really happy.)

I realize that a customized feed interface on a Web page is no big leap in technology, but when taken as a whole and in its simplicity, I realized that Google is in a great place to really jump ahead on the latest paradigm: the widget-centric desktop.

Whether you run Linux, OS X, or XP, widgets that provide information feeds are fast becoming the Next Big Thing on the desktop. News, weather, traffic... information feeds are filling up our screens and making the desktop more than just a place to display cool wallpapers. The desktop is becoming an application platform by itself. Google's personalized home page is an easy portal to such widgets and though it may not be as flashy as something like Konfabulator, it's simple to configure and it runs on any browser.

I believe widgets are the future, at least in the near term. Is it any wonder that Yahoo! just acquired the aforementioned Konfabulator? If they can control the content of the feeds these widgets produce, then Yahoo! gets a big set of eyeballs to sell to advertisers and sponsors.

For Linux, a widget- and browser-based desktop can only be a positive. If Google's tools take off, then it becomes less and less important which OS I am accessing Google from. All I need is a browser and an operating system that connects me to the Internet.

Faced with that requirement, then spending money on XP or OS X quickly becomes unpalatable. If needs are met be something that is just as good and costs a lot less (indeed, nothing at all), then users become that much more attracted to Linux.

At this point, this line of reasoning sounds a lot like the "who cares about the application base, because it's all going to be Internet-based anyway?" argument that some use to justify Linux on the desktop. It's a solid line of reasoning, but I have never much cared for it, personally, because it seems like Linux is just getting users as a sloppy second. Frankly, this should be a secondary argument for Linux on the desktop: valid, but not the one that should be shouted from the rooftops because it ignores all the other things Linux has that makes it an exceptional OS (security, network tools, filesystem...).

Besides, I found evidence this week that the widget/modular approach is something that is working very well for native Linux apps, not just Web-based programs.

The evidence is the application known as Thunderbird.

Thunderbird, in case you have not heard of it, is Mozilla's e-mail and news client. Like the companion browser Firefox, Thunderbird is small and basic in its original form. But with the availability of extensions and themes, users can trick it out and make it run just how they'd like it.

I have had Thunderbird for quite some time, but I have never used it as my primary e-mail client, preferring to use Novell's Evolution. But, ever since I upgraded to Fedora Core 4 and a faster machine, Evolution has given me nothing but trouble. Fetching POP mail took forever thanks to the spam checking, and at times it would hog so much system resources, everything else would grind to a halt. I figured (wrongly as it turned out) that the new machines and its better resources would fix the problem. No such luck. Things still ran much too slowly and, to top it all off, Evolution would spontaneously crash and die for no apparent reason.

After the fifth time of this nonsense, I started up Thunderbird, set up my e-mail accounts, and started using it full time. Migration was easy--all my contacts are on my Palm, so I did not use the Import feature for that. Both Thunderbird and Evolution use the mbox format, so it was really only a matter of copying the mbox files from one directory to another. The only glitch in the process was that since Evolution uses Vfolders to file things away, my actual main Inbox mbox file had all of my messages even if they had been moved to another folder back in Evolution.

Seeing 51,672 unread messages makes for a small scare, but once I realized that all but 250 or so were duped in the subfolders, I was able to quickly clean out the Inbox to something less than insane.

On the whole, I am very happy with this move. The spam filters work very well (and very quickly), my filters were easily recreated, and best of all my system resources are barely breaking a sweat, so my entire system is humming along. The only thing that I am wrestling with now is Thunderbird's apparent lack of ability to save more than one message into a single file. (This needs solved right away, since this is how I start create the daily security digests: save the day's security alerts to an HTML file, run some macros in emacs, and post the content. Suggestions welcome.)

This problem aside, the modularity of Thunderbird has impressed the heck out of me. I had some duplicate messages in many of the subfolders and all it took was a quick download and install of the Remove Duplicate Messages 0.0.6 extension and I was good to go. To me, extensions and even themes are basically just another form of widget. I think this paradigm is the best out there right now: developers build a cross-platform application, keeping it small and relatively simple. Then they open up the code and let anyone build what they want to build. Since everyone is coding to the application and not the operating system, these extensions run on the app whether it is sitting on Linux or Windows.

Mozilla is not the only one doing this, either. Mambo works much the same way with a huge base of add-ons for content management and Web page design. I hope that other apps, such as a word processor, spreadsheet, presentation, yada yada... will be developed along this model too.

For now, Firefox, Thunderbird, and Sunbird are the fulfillment of the promise of that long-ago Netscape keynote and I now understand how an application can become the platform. More importantly, the presense of these apps (native or browser-based) demonstrates more and more that any desktop can have a brilliant and functional user experience. So anyone that tries to sell you on a proprietary OS based on the merits of its actual GUI design is really just selling you vapor.

Applications are becoming more and more platform-agnostic, which means that in order to compete against Linux, other operating systems are going to have to provide better performance and value in other areas like security and stability. And, gee, that's worked out so well for them so far, hasn't it?