i should be working (but i'm fixing my computer).

Not having any major crashes or annoyances, I had been contemplating a post about the shift to web-based applications and the sort of "dead-end" that fancy desktops represent.

Of course, desktops are not all that much of a "dead-end," given that we have been hearing since web-tv, circa 1998, that desktops were obsolete; that we would each have a "web applicance" fed by java applets or some such.

...

I hated the idea. But, basically, what I hated was the thought of giving up "control" to some centralized entity that would no-doubt raise the price of cutting-edge computing to somewhere I can't go.

That was then, this is now. A lot of people are capitulating to things like gmail and google apps. That is, indeed the tug-of-war. The "web-appliance" is moving toward reality.

The "why?" has several answers, at least for me. First is mobility. If I can access all my data flawlessly on my smart phone or on someone else's computer, then I can travel a lot lighter (and don't have the risks of theft and damage). But second is simply one of efficiency. It takes a fair amount of time to maintain the six computers running in my home. I am continually reinventing the wheel, because there are just too many setup parameters to remember and too many minor variances to account for.

And lastly is the need for "backup." Although it is not a flawless solution, professionally maintained servers are less likely to suffer catastrophic data loss than my own haphazard approach. Plus, if all the data is centrally located, I can maintain one local backup and be "good to go" whatever happens. More or less.

But -- absolutely not, I am not putting my data into the hands of a commercial provider. First, I can't customize and "do it my way." But secondly, it's not secure; it will inevitably be more expensive and less flexible than DIY; and lastly, there are serious privacy issues.

So, in that context, I came across some posts in Linux-land that were dancing on the edge of the same issues. Ie, where do we go from here? (I guess I use "we" rather loosely, given my rather non-existent contributions ...)

This post opened a dialog about the sort of static (and aging) state of Gnome. It's somewhat ironic, and perhaps a good thing, that Gnome and KDE seemingly have opposite issues. Gnome is stable and unaggressive in feature development; KDE's "cutting edge" evolution is running a tad ahead of stability.

But, it's down about post #14, by Yannick Defais, where we get some discussion of the centralization issues that surround "server-client" models. And I have to agree.

Nevertheless, I think Yannick overlooks the relatively low expense of "owning" one's own website. True, it's a recurring cost, but relatively minor in the overall scheme of things. And here is where I hope that development will go.

It may be that maintaining one's own website is too difficult for most users. But, that, too, has become increasingly easier. And, perhaps with the correct open source tools, it could become reasonably transparent.

So, for instance, one does not need to use gmail (or equivalent) to have web-based email. There are a number of open source packages, such as Squirrel Mail, Round Cube and the like.

With sshfs connections, we also have the possibility to make our webservers function exactly as though they were directly connected to our local machines. So, in my implementation, my documents are directly encrypted locally, then archived both locally and on my webserver. This is script driven, largely transparent, and achieved with existing open source tools.

I could also, for instance, put my entire music collection on my website. The only reason I do not do this is that I'm afraid I would make some error and accidentally make it "public", which would have really scary legal ramifications.

Now the issue here becomes "how can I most quickly bring a new computer up to speed?" For instance, do I put everything on a bootable "live CD?" Are there enough computers out there that will boot from a flash drive, to simply bring a "live flash drive" with me? Or boot from a CD and then link to a flash drive?

Well, the tools somewhat exist, but getting them up-to-speed has so far been more cumbersome than just dragging along my laptop. Though with the nuances of airport security, one prefers to bring less and less hardware (and data) along.

And yet, locally executing programs still seem preferable to me than web-based ones. I can use Squirrel Mail to access my email, but Kmail is faster. (Thunderbird is also faster -- but Thunderbird does not implement IMAP as well. I run filters on my server to sort mail into folders. Kmail and Squirrel Mail will highlight the folders with new mail, but Thunderbird will not do so until I actually access each individual folder.)

Similarly, I can use Google Reader (though I'm not fond of allowing Google to track everything I read) -- but Akregator is faster and more convenient.

And then again, not so convenient when I switch computers.

So, I lean toward packages that are O/S neutral. Mozilla is doing a good job with these, such as Firefox, Thunderbird and Songbird. The drawback, of course, is that after installing these, they still must be "customized" for each computer. What I would like to do is click on an icon in my browser (or on my desktop, or whatever new idea materializes) and EITHER execute Firefox as already installed on my computer; or if not installed, d/l and install it from MY server, with MY customizations. And thereafter, maintain the customizations in sync on future launches.

So, what I really hope to see is a system that transparently obtains the necessary software from MY server; runs everything directly from MY server; and lastly, maintains a local copy and capability to still function if the internet disappears.

All of this is completely possible with existing tools -- but not in a time-effective, transparent manner.

Lastly, we really need transparent PIM synchronization across multiple devices.

Many of these projects are "in the works". There are pieces of them built into the major desktop distros. There are also entire web-based projects solving parts of the puzzle. But so far, I have not stumbled upon a project with the relentlessness and vision of, say KDE and Gnome. Desktop programmers tend to focus on personal computers and the other projects focus on web-based applications. The integration is not occurring very rapidly -- perhaps because not enough of us maintain our own websites.

So, in any event, this is just my "sales pitch" to programmers. Save the Free World! Build hybrid client-server apps that we can run from our own websites, so that the "centralization" predict by Yannick in the comment referenced above does not come to pass.

Because otherwise, Google (or the like) is going to rule the world...

Thank you for listening and have a pleasant tomorrow.