There has been a lot of buzz this week about the Desktop Linux Client Survey 2005, release earlier this week by Open Source Development Labs. In this report OSDL says the lack of a quality email program prevents Linux from being widely adopted on the desktop. Some argue that Linux is suitable for the common user while others argue that more application support and training is needed.

My opinion in this debate is simple. I believe users will go where the applications are and most applications developers (including in-house, corporate developers) and ISV are NOT targeting Linux. And the reason many developers do not write desktop applications for Linux is because there are no good RAD development environments like Delphi and Visual Studio to help increase developer productivity for Linux development. Also, there is a lack of 3rd party control support for Linux developers. At the end of the day it is easier to write and go to market with a Windows desktop application than it is with a Linux desktop application (or Mac application for that matter).

Vendors like Borland have tried in the past to sell a RAD environment to developers for Linux, and I for one wish they would continue work on Kylix. Unfortunately Borland has made Kylix a classic product which means there is still some demand but the product is no longer enhanced, supported or marketed. A real shame in my opinion.

Speaking as an ISV I would very much like to see White Peak Software target Linux for its desktop products which includes a new email client for the Windows platform due out next year. However the time to market for Linux desktop applications is much longer. Development tools like Delphi and 3rd party controls like the ones from DevExpress and /nsoftware make it easier and faster to write and publish Windows desktop applications.

Give general development community the tools needed to write Linux desktop applications faster and I believe you will see more Linux desktop applications. And once there are more useful, quality applications available on the Linux desktop the user community will follow. It worked for Palm in the 90's. Palm made it easier for developers to write Palm Pilot applications as compared to WinCE, and Palm Pilot became the preferred hand held for the majority of users.

It has also worked for gaming consoles. The number of titles and the quality of those titles helps decide which console has the majority of the market. Microsoft wants to see the new Xbox 360 beat Sony's PlayStation worldwide. A key point in Microsoft's plan is to attract the biggest names in game development to produce Xbox 360 game titles. Microsoft knows that the best games must be available on its new console for the platform to be a marketing success. The gaming community will go where the best games are. The desktop user community will do the same.

Linux will become more feasible for the desktop once the number of useful software titles has reached a critical mass that attracts the general user. And to make it happen, more software developers need to target the Linux desktop. But first there needs to be better development tools that make it easier and faster to go to market with a Linux desktop application. This will attract more software developers, which in turn will produce more software titles, which will attract more users to the Linux desktop.

posted by Kirby | 02-Dec-2005 10:02 AM | comments (5)


Kirby, you make valid points. There are a growing number of component vendors like /n software that are providing components for Linux developers. Also don't forget that there are tons of Java components out there.

I completely agree with you about Kylix. Its a shame that the product seems to have had such a short life. Until Borland gave it the death nail /n software actually had a Kylix Edition. Technically you can still use our components there using a combination of our Delphi Edition and our Unix Edition library.

Finally, regarding software being the deciding factor - I agree again. I think this is a recent change though. Not too long ago the deciding factor was the Linux distributions that were available and their ease of use. Now there are lots of well implemented, easy to use and intuitive distributions, like SuSE. The next step is great software. It will be interesting to see what long term effect things like OpenOffice have.

posted by Lance Robinson | December 2 04:01 PM


Thanks for pointing out that /n software has components for Linux developers. I'm a Red Carpet Subscriber - great components by the way - so I should have known that.

posted by Kirby Turner | December 2 04:23 PM


Great! Now we have to get to rolling out that Delphi 2006 support, huh? :)

posted by Lance Robinson | December 5 07:40 AM


LOL. Yep. Get Delphi 2006 support in right away. ;)

The /n software components are some of my favorite. I tend to create instances programatically and not through the designer. I'm assuming the /n software VCL components with compile fine with Delphi 2006 but you never know. Those that prefer to drag and drop might want Delphi 2006 support sooner than me though. Of course I have no idea when my copy of Delphi 2006 will arrive.

I preorderd my copy at the beginning of November. I got an email yesterday telling me the order has been backordered for another 30 days! Needless to say, I'm not happy with Borland at the moment. They really should have a download option. /n software sends discs and has a download option. Borland should follow your lead.

posted by Kirby Turner | December 5 08:04 AM


Hmm...I seem to remember BDS2005 only coming in the form of a downloadable iso?

posted by Lance Robinson | December 5 02:14 PM

Add Your Comment

Comment:
(No HTML)

Name:
E-mail/Web site:
Your e-mail/web site will not be published on this site. It is optional and will only be used by me should I need to contact you directly.
 
By checking this option, this site will remember your name and e-mail/web site on future visits. Uncheck this option to have the site not remember who you are on future visits.
 
Enter the code shown above:
Copyright © 1999-2008 Kirby Turner.
Site software written by White Peak Software Inc, a provider of custom software and software development coaching.