Right. Improved drive reliability still doesn't negate a need for backups. It's moved from NAND write wear to controller failure as the primary killer of SSD disks. On one hand this isn't very predictable, on the other measures to control the damage when it happens are available.

This. SLC offers a "promise" of greater longevity with lots of write cycling. Still when most SSD's fail they die at the controller.

A number of 840 evo drives experienced substantial performance degradation much faster than was reasonable. That said I ended up getting an 850 evo recently because with Samsung's output volume I imagine their whiner/[just works] ratio is probably rather favourable to the consumer. In the past I mostly used Intel SSD's to great life span (storage needs began exceeding requirements well before hardware failure). When buying as SSD the big concern is going to be how many is the maker putting out versus how many complaints there are. Of course with the recent resurrection of storage firmware diddling in the news it would make sense to take measures to keep NSAware away from your drives as it can't be good for longevity or performance.

This. So much this. When these regressions happen there are people behind them. The great value of a Linus or a Theo is shaming this people out the door. At least this was caught in -Current and not -Stable. This incident appears to be at least as much a social engineering attack as a code quality issue.

Mostly the guide on ribalinux.blogspot.com did it. I didn't follow it exactly because who really wants/needs Dbus. If you google around there's plenty of workable guides, just make sure they point to more recent versions of OpenBSD. Installing XFCE now is as simple as pkg_add XFCE and editing a couple files to get it to automatically fire up by default. The most time consuming part of the affair was getting a new wifi card that OpenBSD supports (~20 dollars) and reflashing my BIOS to remove the Wifi whitelist.

Actually you can do exactly this in PC-BSD, which is one of my dislikes of this desktop BSD effort. Its repository leans heavily on the linux compatibility layer instead of ported software.

LibreSSL attracted that Cook fellow to OpenBSD who has contributed substantially to a portability layer for it and he's now branching out last I hear to other OpenBSd derived tools portability efforts. It isn't much different that porting a Linux first program to a BSD except if it was born on a BSD first tytpically less porting effort is required.

PC-BSD started Lumina very recently and haven't taken it out of Beta yet. Probably best to give it a little more time.

That and it's more than two years out before they consider systemd. And yet many systemd fans point to Mint as an example of modern linux that just works.

Loving OpenBSD on a laptop. 15 minutes of futzing with configs and a beautiful desktop environment of your choice. I haven't loved this much since Solaris 8. (Oh and that Android is linux bsbs... Google doesn't use glibc for android, they use OpenBSD's libc)

But Linux Mint itself is still without systemd until at least 2017 by their own timeline and even then on their "Cinnamon" desktop spin they are still planning at least now to support multiple init systems when Linux Mint 18 comes out.