My recent foray into web2.0 living, when I left my power cord at home yesterday, was an interesting experience, and has left me thinking a good deal about computer usage in the future.
Before I get started lets just assume that by web 2.0 I mean, the move towards using tools like AJAX and Ruby-On-Rails to build quick and sturdy applications run server-side, and usually, centrally located as services. Google is like the web 2.0 company, but I mentioned Meebo, and 37Signals are great examples of the kinds of applications that are pushing these models.
For example, ideally what GoogleDocs has over traditional office software is that it's platform, and even machine independent, and it uses high quality file formats. Also, there's a whole level of stuff that Google takes care of (updating the software, maintaining the servers, and so forth,) that traditional software users have to pay attention to.
The side effect is that you have to "live" your computing life entirely in a web browser, which many of you may already be doing (I was watching theBoy do something on his computer, and I was sort of amazed at how much time he spent in a web browser and want not: I apparently live in a different world.) But I don't think this is the way to go: web browsers aren't very standard or consistent, and for the past 10 years, every browser I've ever used has consistently been the among the biggest and slowest applications I use regularly. (Clearly Photoshop wins this, no contest, and MS Office products are close, but I don't use the former very much and the latter at all any more.)
As people start to use more than one computer, internet connectivity becomes more ubiquitous, and as large swaths of the computing (and developing) public starts to use non-Windows OSes  it makes sense that the smart thing to do from a development perspective is to write Web-based programs: everyone can use them regardless of platform, they work everywhere there's a connection, and in a lot of cases they're as easy if not easier to write than their desktop equivalents. 
Also, being Web-based is good for business: it means that you're charging people, not for the rights to the intellectual property that is your software, but for the use of that software running on their server. I'm all in favor of business models that "use" IP rather than "sell" IP. There aren't effective ways to sell IP, but there are lots of effective ways to generate services.
But despite this, it's all incredibly unsatisfying. Web browsers make really bad application interfaces/layers, and writing new programs isn't always the best way to get better and more open file formats. I don't know, I think someday most of our computing won't happen on the devices we're holding in our hands, but I'm not sure that the way to acomplish this is through a web-browser.
|||So MS has most of total market, that's pretty much fact, but I figure that in certain markets, apple has a much larger share. College students, Designers, Hipsters, Ruby-on-Rails developers and so forth are all disproportionately Mac users, I'd figure. Adding this to the fact that Ubuntu is, I think doing well, and Vista continues to suck...|
|||If you're administering a web-server that people are going to run your programs on, then you can write the program and know that it's going to run on a box with the right version of Ruby, or Python, or SQLite3 and so forth.|