OSDev.org

In fairness, I think the cloud, as people define it, is not quite a cloud. Most people define cloud computing as "processing done anywhere but here on my desk." and cloud storage as "my data, somewhere else in the world."

The problem with these definitions, is the fact that this largely means that the "cloud" is in fact many little clouds, each privately owned. There is nothing wrong with that in and of itself, except for the fact that it brings its own set of issues. First, and foremost, (and often most spoken) is the fact that you're handing the reigns over to one sgned just to stand-out or because of compability-bagage. Primarily I dislike most abstractions that others have done, and think I could have done them so much better myself. So why work with other's bad abstractions when I can work with my own?
pecific entity. If they then decide to cut you off, or they go bust, or close the product line, woe betide you if you don't have any backups. This is also an issue for cloud computing, as seen in the recent Adobe fiasco.

Furthermore, the cloud as it stands requires a lot of manual work, and administration. Want a cloud backup solution? Fine, here you go, thats $lots, please. Alternatively, lots of little accounts, each of them free, or $not-very-much, which not only gives you more space for less (if done correctly), but also protects you from the problem above, with one corporate entity having all of the control. It is an issue, however, in that there are no tools to manage this. There is basically nothing to sync folders between cloud storage solutions, nothing to make several cloud "drives" as one (which basically means you have to manually split data as well as sync it), and NOTHING of this sort for cloud computing.

Then there's the problem of uplink speed, however I won't go there.

Thus, to fix the "cloud", we need to switch to the cloud. We need to, instead of having lots of little, private clouds, as you might have (say) a VPN, instead, one big, public Cloud, like we have the Internet. Fully distributed, delocalized control, automatically synchronising, and much easier to use. Still a few issues, but at least we've ironed out the big ones.

I dislike "the cloud" because it requires a nearly constant internet connection, and basically unlimited data. Maybe in large cities that's not a problem, but I'm not there, and in about two weeks, I'll probably be going a week or two between chances to be online.

I don't see why I need to pa monthly for data storage that's [for me] unreliable and unusable, which requires another large monthly fee to access, when I can buy a hard drive and store it myself for cheaper than one month of cloud storage, and I can access that anywhere without an internet connection.

I have done some "cloud computing" (well, I don't know if php and mySQL really counts as such), but that was to do a quiz. I certainly am not analysing anything online, rather do SQL-exports to my local computer and then analyse the results locally. Then I let the analysis program produce HTML-files that I uploaded. I suppose that's not real cloud-computing, but it is as far as I will go on that. I don't need to maintain anything, as both PHP and mySQL are maintained by the web hotel I use. They even maintain SVN for me, which I use for version control.

The best way to do the cloud would be to keep your own one in your house, but as no house is safe (tornados, floods, fire, etc.) you'd also want to share space with relatives/friends in a network of clouds so that there are always going to be backups of everything. I certainly don't want a lot of my data to be held on machines owned by companies which want to mine it in an attempt to make money out of me. Unfortunately, each home cloud would need to be in well defended against burglars, so there are no easy answers to this.

I do it for the powerful feeling that all the code running on a machine is my own.

Also just for fun and educational purposes.