42

In an old-school networked environment you would log in to a computer and have a common, network shared, home directory. This meant that it didn't matter what machine you sat at you always had your desktop and application settings etc.

Increasingly we operate on portable devices that can't guarantee access to a universal home directory. And this means lots of manual copying and tweaking of settings. I'm thinking mail account details, subscribed news feeds, bookmarks etc as well as desktop prefs and possibly even session information (open web pages etc).

What I'd like to see is a simple way of getting my KDE settings from one machine to another, and keeping everything synced. The method would need to be able to identify potential problems and provide appropriate guidance for resolving those problem (might for example include automatically downloading extra widgets or excluding certain settings where incompatiblities or non resolvable network names exist)

It would take a trivial amount of storage on a kde server to host user applications settings and desktop preferences (omitting security critical stuff of course). Perhaps this information could even be hosted by the KDE project What a value add for KDE users! (even if it could only be done via a modest subscription fee to cover hosting costs)

Wouldn't this require loads of bandwidth? That's why my college didn't use roaming profiles with Windows or Mac.

Roaming profiles requires your settings to be remotely hosted - like having part of your home directory on a file share.

What I had in mind here was more like rsyncing your settings on login. Something that only communicates changes. I would not expect a huge amount of bandwidth for that. I do notice that my ~/.kdemod4 currently consumes 1.3G, but 90% of that is nepomuk data that would not be shared. This is mostly about mail settings and news feeds and what has (or has not) been read, web browser settings and bookmarks, etc. I'm not exactly sure where the session data is stored but that would also seem likely to be a realtively trivial amount of data. Especially if cut back to only that which can be meaningfully shared.

Now I understand, it sounds like a great idea. I don't use rsync, how fast would it be to check every single file in just about every directory?