Linux on the desktop has seen some significant successes over the years, from improvements with hardware detection to user adoption. Yet despite these successes, the single sticking point I find myself arguing with people over the most is the idea that existing methods of software installation are ideal.
Installing software with most distributions is pretty brain-dead simple. With command line options and a variety of GUI solutions to make the process even easier, I genuinely don't think theres a problem with the ease of software installation.
Dude, where's my software?
For moderately experienced Linux enthusiasts, most software is a stone's throw away. But even the more experienced desktop Linux users have been known to discover a new application from the most inconvenient sources.
Often these discoveries take place long after the user has given up locating such an application when they needed it most. Where this becomes truly problematic is when the application was available from the software repositories used with the user's own distro all along. Yet the app went totally unheard of because the user didnt know which category it was featured in!
This is hardly an isolated incident, mind you. I can count at least seven individuals who I know personally who have been in this situation. Is it office, business or communication related? Also, how is the performance rated? So many questions often going unanswered.
Well, at one time there was a solution to this problem on the Linux platform. Most people within the Linux community scorned the solution at the time, due to strong opinions of the Linux distro this utility was bundled with. Regardless, the utility itself has yet to be matched.
The utility was known as Click and Run.
Enter Linspire's CNR software installation utility
Back when Linspire was still known as Lindows, they introduced a software installation utility that changed the way people looked at installing software. Click and Run (CNR) at the time, was the simplest way to discover and install new software applications ever seen on any platform.
Sadly thanks to the evolution of the company that created the utility, CNR of today is not nearly as compelling as it once was. Now its merely another application that must first be installed, then used to install software. It uses basically the same methods employed already by a number of popular Debian-based distros.
With the exception of making applications (such as LinDVD, among a few other proprietary applications) available without much searching, there is really no clear advantage that I can see.
Perhaps the final nail in CNR's evolutionary coffin is the missing software aisles that were big with the original CNR utility built into the Linspire 5.0 release. At that time, not only could a user keep track of which software is in their preferred list, they were free to share this list with others.
Sadly since the move over to the new CNR utility, I have yet to see evidence of this function.
Clearly there was something quite user friendly here. Seems to me that the idea was right at one time, now it simply needs to evolve with the times.
From CNR to a Web based App Store
A headache I used find myself frustrated with was a lack of applications designed to fit certain needs with specific levels of functionality.
Sure, more often than not there was something GTK- or QT-based out there that would give me the basics of what I was looking for. However in rare instances, I found myself needing software with a more razor-focus to handle specific tasks.
Then Adobe AIR came out for Linux. Almost immediately I found myself running a multitude of applications on my PC that were unavailable previously. It took some searching, but there are some fantastic AIR apps out there that are worth a look.
For various web site endeavors, I found myself using an app known as Market Samurai. I also run specific apps for Twitter and Facebook.
Productivity apps I fell in love with include "Klok" and Present.ly's own AIR application. In each instance, the natively available Linux software did not hold a candle to what was offered for Adobe AIR. Not even close! Adobe AIR really opened new doors for me.
So this got me thinking. Why not have some kind of Linux friendly App Store that included both Adobe AIR apps and native Linux software? Seemed doable.
Perhaps the problem is that there's not a definable market for it? After all, Linux has long since been perceived by outsiders as a platform used by cheapskates, at least from a software perspective.
And web-based software, while plentiful, is rather limited in what its able to do. Besides, there is not really any challenge in finding web-based apps. Just go to Google.com and 99% of them are right there.
After rolling the idea around for awhile, I finally came to terms with the issue that an age old problem exists and is showing no real signs of ever changing. Most folks are not interested in mixing proprietary software in with their open source software repositories. Yes, Ubuntu among a couple of other distros do this. But they do so sparingly. And even then, they get a lot of grief for it.
Realizing this, I believe that there must be a way to keep the proprietary apps completely out of the repositories. Yet at the same time, allow an app store to offer these closed source Adobe AIR applications while also utilizing the default Linux software repositories provided by your specific Linux distribution.
It's a tall order, I wonder if it is doable?
The key to a Linux App Store
I think the key to putting together an app store for Linux stands with being able to take what works, then making it better. This means accepting reality.
People are going to generally seek out software based on its function, not its license. This means providing access to both free and paid software. Obviously the bulk of the paid stuff coming from Adobe is AIR-based solutions such as iplotz, as one such example.
Clearly narrowing down to a business model that is both sustainable and something palatable for the existing Linux user community is no small task. I believe the biggest issue is to offer definable value where there is none at the moment. Figure out what is missing from the typical application installation methods, then enhance around them without stepping on any toes along the way.
Dreaming of Linux Delivery
This article has hit upon a number of areas that might be defined as sensitive to a number of people. At no time am I claiming that existing methods of software installation need to be replaced.
Rather, I see existing software installation being enhanced, perhaps made a bit more consumer friendly -- especially considering the sheer number of applications in the Adobe AIR realm that are not available in a GTK/QT build, yet would be of real benefit nonetheless.
I don't pretend to have the magic formula in which GTK/QT based software can live in the same marketplace as software using options such as Adobe AIR. Even though both software families appear to run quite smoothly on each of my desktops, there may be challenges here I havent anticipated.
I have a dream in which application discovery is free of licensing concerns and instead is based on what it can do. I dream of a Linux software marketplace that is able to make this a practical event for all to enjoy.
Sadly, at this point any thought on how to best merge these two very different worlds of software has left me high and dry. Nothing. No idea how such a thing would come together. Maybe shooting for something like what we see with Google Android?
Perhaps someone out there reading this thinks there is some vague merit in my vision? One can only dream, I guess. Then again, stranger things have happened.