Trying to prioritize three steps towards a higher linux based desktop market share, I came up with the three speed bumps on the road leading to "freedom of choice"
#1 - Stability
Following the debate in papers and blogs these days, there seems to be a resistance to replace XP with Windows 7. Now - my experience tells me, that the IT-industry has always resisted upgrades. It's a tough job to find any users out there, who would argue the old system to be worse than the latest. Remember how good OS/2 was? or DOS for that matter? Seems like over the years, it's just getting worse and worse ;)
Of course not! We are improving all the time. I guess nobody would ask to have a new laptop pre-installed with any OS older than the last release. XP's been around for nearly a decade. People know it and are comfortable using it. It might not be perfect, but it works and it has a variety of apps, that can be installed - out of the box. Download, install. So stability - not in the sense of avoiding crashes and the like, but as in "it really works" is important to the end user.
Now whats with GNU/linux?
Most distributions have a 6-12 month release cycle, and only 2 versions of backward support (security patching).
Go ahead - do the math yourself... Why upgrade every other year, if it's avoidable?
I'd like to see the impact of a GNU/linux distribution that remains stable in 10-20 years. I'd be likely to chose it. And I guess CIO's in charge of an install base of tens of housands clients, would be easier to convince to migrate.
Would it be possible? With the architecture of Windows I doubt it - on linux, there might be a chance?
#2 - Data and security
The world is turning fast these days. Remember how much/little data you had stored 10-15 years ago? 50 MB? Might have been a bit more. Today everybody has shitloads of data sitting in their laptops. Prom snapshots, wedding photos, scanned images of the world that was, not to mention music that easily takes up 30+ GB or your home video and DivX library at 100+ GB. That's a lot of data. Sure, that amount of data, is not going to decrease in the years to come. So: Data management is important!
That being said: Dolphin has a looong way to go.
The traditional directory structure of filing data, is starting to lose relevance. Hell, if I am not super structured in building up my directory tree, it ges really hard to find that budget-file I spend hours preparing last year. Not to mention my 13 year old son or my 65 year old mother, who will simultaneously go, "huh?" if I ask them, to name their default directory path. They don't know! All they know is that if the put in a CD in the drive, the music will appear in the player, and later, when they plug in the iPod, the music will magically appear in it minutes later. Great! :)
Some data is private, while most data is meant to be shared amongst users. Does GNU imply some standard for placing (mounting) shared drives and document structure, that is implemented in the applications? Admitted, it does take away a bit of freedom, but as a gain it also takes out the annoying "path to:" text boxes in many GNU applications. And as a bonus you gain access to the users (users!) that don't care about HDD architecture.
But make it a GNU standard, don't have it depend on the distribution flavour. Make sure to communicate the standard to the open source community developers.
I'd like to see a distribution that comes with out-of-the-box policies and directory-structure that allows end user low-brain-activity access to data. :p
#3 - Consumer applications
Hopefully, the presence of a stable distribution that doesn't require regular reinstalls, will grow a number of end-user applications.
The security model of GNU/linux shows the path for managing those. But I don't really see anybody walk the talk yet.
Today root access is required to install/deinstall anything on your linux based PC.
I see application installation happen in three ways (A, B, C):
A. User install - concerning private data management. GNU needs more emphasis on user installs. I am not really sure exactly which applications fall into this category, but I guess mostly smaller (thin client/offline) games and social networking stuff like 'picasa' (though a linux version is not available - but you get the point).
Let users install apps on their own account with no risk to the main system. Application data is private.
And remember, end users 'double click' - they don't fiddle with 'gcc', 'make' & 'install'. Which again means distribution of binaries rather than source code. Don't let the open souce ideals block the growth of Linux. People are willing to pay for almost any application (even if they can have it for free - MS is a stunning example). There are two elements of security here:
1. Open source is secure, because if you are concerned about security issues, you can fix them yourself and distribute your own source 'fork'. Proactive security.
2. Paid software comes with T&C's. If the T&C are broken, you can sue the vendor. Passive security.
Anyway the latter model is very popular with Average Joe and amongst business leaders, because it doesn't require programming skills.
B. Group install - concerning shared data management. To avoid having the same installation in all user accounts larger applications, i.e. office-suites, media players, WOW-like games could be installed and accessed across a group of users, but still without root permission. Application data can be shared by group or private.
Admitted, I have security concerns regarding group-wide installs and you might argue that some kind of admin authorization be required to perform this kind of installation.
C. Root install - concerning system wide administation, security patch and general data management. In the ideal world the end user need not to care about root access. In the real world, the user might occationally have to type in the root password when prompted by security update client. Application data is 'invisible' to regular users.
The question is what happens, when you PC boots up in nothing but a browser? I'll think about that over the next couple of days and get back to you with my view.
Thanks for reading all the way to the end. Post a comment, I'd appreciate it!