by pylkko » 2017-12-30 21:31
In my experience, when we are talking about a single install that is used for mostly unimportant stuff, and the user is more or less experienced with tweaking operating systems, it does not really matter what OS/distribution you are using. The only differences when moving from Debian to testing will be 1) more time will go into maintenance, and 2) security updates might be several weeks out of date compared to stable 3) more difficult to get support.
I used testing and sid for years and maybe twice I got into the situation where an update led to a non-bootable system. At least once the microcode was updated so that it required a UEFI update, and until I did that, it would not boot. Other than that, the machines were always usable. However, it was quite time consuming. For example, some software would rewrite the configuration files to default settings, so then I would have to manually re-edit them. Or at other times a program would no longer work with other packages and I had to uninstall it until it was also updated and would then reinstall it. These are really small things, but they might end up consuming more time than you would want to invest it it. If I had to guess the number one reason that power users have for going back to stable (if they ever do), I would bet it is "avoiding the hassle and time spent in tweaking the system all the time". On the other hand, if you have time to manually run several update commands daily (see above posts) like some do, then maybe you have time... hint: try automating it.
It will help if you know how to chroot a system, have some understanding of how to use the rescue shell, know how to mount a disk from a live system etc.
I don't think that testing is like Fedora, which after all, is a curated release, whereas testing is nothing, just a collection of packages.