I deleted skype and one of my gtalk accounts, then re-added them, and they are correctly and openly written @ ~/.rtcom-accounts/accounts.cfg

I see no change in that behavior between PR1.0 and PR1.1. Haven't tried backing it up, tho, but I guess the result would be the same.

It has nothing to do with high expectations, I'm perfectly aware that an average Joe barely knows what a computer is, let alone how and where it stores files. However, as many people have noted, providing a base64 encoding or something equally trivial would not give anyone anymore security - what's the difference if the file containing passwords instead of `password` have `cGFzc3dvcmQ=`? It's the same f. thing!

If you know where to find the file, you've probably found out that on one of the following ways:
1) You are a tech-savy and you know where some application stores its files. In that case, you already know how it stores it, and how to decode possibly encoded passwords.
2) You found it on the internet (for example on this thread). If the files were encoded using base64 (or something as trivial as b64) instead of plain text, this thread would already have a step-by-step instructions on how to deobfuscate those passwords, so you'd still get the passwords with one additional step.
3) You were browsing through someone's device long enough and checking each file and suddenly you came across a file that stored accounts data. If passwords were plain text, you'd know them immediately, if they weren't, chances are that with simple google search for that file you'll find a thread/blog/whatever that explains how to extract the passwords.

In all three possible cases, passwords are not any more safe stored with a trivial, reversible encoding, than in a plain text. But knowing that your passwords are not safely stored is actually better thing than having a false sense of security - this way you won't be giving your device to anyone that easily and you'll know the risks involved.

That is a perfectly good solution for paranoids. And that should be set as an enhancement requirement. The whole argument here is that stored passwords in trivial encoding are not any more safe than those in plain text. If someone wants real encryption, that's a perfectly valid request, but they should be prepared to give up on the convenience of password-less auto login.

And how would that be any different than, as Jaffa already noted, having a slightly different copy/paste command if the file was base64 encoded: