Username or EmailPassword
Web applications diminish the incentive to make native desktop applications on all platforms, not just on Linux.
The sooner we get to the point no one gives a damn, or has to for that matter, about the underlying OS their software runs on the better as far as I'm concerned.
Applications with web-based front-ends have a fair share of advantages over native ones, even if you don't use them to collaborate with others through the Internet, and even if you run them on your own machine(s). Notes (including task / todo / buy-list management), personal finance management etc. would be good candidates, as it is useful to be able to access to these from tablets and/or phones besides your desktop/laptop, and having a web-based interface is still easier than having entirely different interfaces for each device type. (You will need somewhat different designs for different screen sizes, but it is still easier to use and develop than having completely different front-ends for each). They are also inherently multi-platform (for the client-side at least, but most web-apps are also developed in a portable manner anyway). Some of these services also happen to be managing information that one may not want to be hosted at a "cloud provider", so the ability to self-host is also important.
Of course, there are still many cases where a native application is better, but I expect that there will be more and more web-based ones coming. Edited 2011-11-06 10:11 UTC
Consider that GTK3 has a web based rendering back end available. It's entirely possible that in the future web based apps will be able to use the native widgets on whatever platform they choose.
I certainly agree that, as a platform, the browser has its advantages.
For example, I find it MUCH easier to craft custom GUI elements in CSS/HTML/JS than in raw GTK+ drawing code, even with PyGTK and it does make it much easier to write a single UI that adjusts to all form factors.
However, I'm also the kind of guy who uses Gentoo Linux because it's the easiest way I've found so far to mix-and-match exactly the versions I want of various apps, lock them in if I don't like the newer versions, and apply my own patches if need be.
For apps where the browser DOES provide the simplest solution which are also fully-open source (like TiddlyWiki), I've never had an issue with using web apps.
Here is a real life testcase of the challenges that a developer face developing a commercial high performance (low latency native) Linux application.
http://www.ardour.org/node/4643#comment-26973 Edited 2011-11-06 13:28 UTC
I love how this article was posted right after the "Support Linux by Not Writing Linux-Only Software" news item was posted in the News section. It puts together two good points of view. I have to agree however that cross platform development is better.
One example given in the article, "Support Linux by Not Writing Linux-Only Software" talks about a user frustrated that AppX does not exist in Linux like it did in Windows. Now that user just wants to go back to AppX in Windows or Mac OS X because that is what they are familiar with even though the Linux variant may be 10 times better! Sure education could solve this problem, but computers have become such a commodity appliance that people don't want to 'learn computers', they just want to use it.
While I think the Linux ecosystem already has some killer apps in general, I think the only way for general users to actually use Linux is for developers to utilize cross platform development. It would ease the transition for new users in a new OS, which let's face it, can be tough for some. Maybe if more popular cross platform software existed in Linux, people would actually INSTALL Linux instead of run it on a live CD or running it in a VM where it can easily be turned off and never turned on again.
What I would love to see however is a better cross development platform. Java is too bulky. Mono is a great alternative, but is also very biased. Flash wouldn't be so bad if it weren't such a performance hog. The web is awesome (HTML 5, etc.) in that just about anyone can code an app, but it's naturally sandboxed (although that's changing too). But that's another soapbox to get on for another time.
Regardless, I am not sure if Linux-Only software is the way to go...although I must admit my little nerdy heart desires for people to move to Linux because of apps...not because they're running away from OSX/Windows.
The problem with your HIB analogy is that they've started acting like douchebags with these single game "bundles" These las 2 bundles could have easily been just a single bundle. Because of this though I like many others have dropped our payouts significantly while many more fully decided not to buy the HIB because it only had the one game at release and ignoring that other games where added over time.
I still pitched in $20, though I'd usually throw in $50, but all of that $20 went to the EFF, nothing to the HIB or the devs. None to Child's play either, Net Neutrality is far more important.
The need to run software anywhere is critical.
What happens to a specialist programme that uses hardware, storage media or file formats that no longer exist?
I recently met a scientist with that exact problem. He wrote an extremely important Fortran programme for pasture management in the 1970s. The only copy of the code still available was a paper printout (about 2000 pages). No one else really understood his code so he had to come out of retirement to work on it.
Had the paper copy been lost years of work would have wasted.
I've recently been working on a major application push into the SaaS world. The app will be available for private enterprises or via a subscription over the internet.. The reason is very simple.. no loner do we need to maintain 2 or 3 different versions but rather a single version that works across all standards based browsers.. for many business apps that's the way forward, there are some places where apps in the cloud may not be appropriate (i.e. anything that requires connectivity when there is none lol)