Published at LXer:

On the occasion of the presentation of Google's Chrome OS operating system, Richard Stallman, founder of the GNU project and the Free Software Foundation (FSF), has reiterated his criticism of cloud computing. As Stallman told the Guardian newspaper, people are being taught "careless computing"; they are simply not as careful when their documents and other information are stored on third-party servers instead of on devices under their own control. He says there are also legal risks, for instance when users in the US lose access control to their data in the cloud because authorities have stepped in.

Read More...

The concept of the "cloud" still bothers me. There is no help if the cloud dissipates.

Exactly, I wouldn't use that either.

I'm never using any cloud BS. I want my data to remain on my computer, and I want my power to do anything I want.

NO way, nothing leaves the boundaries of my LAN.

All my base are belong to ME!!!

And how do you feel, H_TeXMeX_H and sycamorex, about an "Ubuntu One Private Repository" as a viable alternative to Chrome OS, as I had already posted a link to a blueprint for?

I wouldn't want any of my data on anyone elses server, even if it is only a program setting. Bad enough that I use Yahoo for my email.

Edit: By the way, your idea wouldn't work for me. I think it would only work if I only use one type of computer, but I don't. And I don't see the point in setting up my workstation exactly like my file server or my laptop. And even if I would only use one type of system, I would install it one time and then back it up to an installable live-CD/DVD. Would be easier and I don't have to have internet at install time.

If I understand you correctly, it's only for installed packages. If so, wouldn't it be easier/safer to create a 'queue' of installed packages locally and next time you install Ubuntu provide it to the installer? Saves bandwith and time. Why have the same .debs in two places (ubuntu official repository and your private one)? A text file with a list/queue/tags is much more efficient. BTW, something similar can be done in slackware.

And what if those installed packages are from PPAs? Then you need to re-source the PPAs to install the software, don't you... Unless those packages are backed up on another repo, like in this case.

Just save your sources.list?

In this case, if you don't want to build them again, I'd create such a repo locally (eg. on my NAS) so that packages built for my specification with the flags I want are available for me at any time. Again, saves bandwidth. Besides, the argument that someone else could benefit from them or I could benefit from someone else's packages is not convenient for me. With the exception of Slackware maintainers' private repos of slackware packages, I don't trust other people's private repositories.

Same here, all my computers differ in what is installed and what is not. I also prefer installing packages manually, myself, not automatically, unless I write the script myself.