Here is an Upstart script copied directly from the minecraft server of my 11-year-old son:

start on runlevel [2345]
stop on runlevel [^2345]

setuid attila
setgid mc
env LANG=en_US.UTF-8
chdir /home/attila/minecraft

exec java -client -Xms2500M -Xmx2510M -jar spigot.jar > /dev/null
respawn

And then:
# start minecraft

I cannot see a single line which requires documentation in order to understand it.

Upstart was already in a good shape when SystemD started. So if you consider this fragmentisation because of this, then please note that in this case it is not Ubuntu who fragments Linux distros, quite the opposite... Personally I find writing Upstart configuration files trivial. On the other hand I have never been able to write a perfect System V script. I hadn't found comprehensive documentation about it (especially about the various helper scripts and config files frequently used in init scripts). On the other hand, if you read the Upstart doc, which is a single long HTML page, at most 1 hour to read, your will know everything about it, much more than what you will ever use. I have not tried SystemD, so I have no opinion on that.

I have two, the first one displays system monitoring data. The diagrams are produced by Graphite on a real server, RPi displays them in a browser. That was not easy at first, because both Chromium and Midori are plagued with memory leaks which does not work well with Javascipt running in 24*7. My son wrote a script which reloads the tabs every hour, since that it works without issues. It only stops when there is a power outage.

I use the other one as a certificate authority, it is not connected to the internet.

Amazon does a favor with their Alexa service for the whole internet. That is the only third party global site statistics tool which provides information for free. At least I do not know any other.

Of course they should fix the vulnerability. The real issue is that the current authorization systems only give half of the necessary information, they state what information the app access, but not what it does with those information, even though that could really make a difference. Therefore people become accustomed to give horrific permissions to any app.

Looking at the book cover carefully, I believe she is actually depicted on it. Not on the forward big ship, but on the small ship going away in the background. She mentions in the book that this is the last edition for her, because nowdays she spend most of her time on a sailing ship where she only has a 30 (or 300?) baud radio connection.

I am usually a developer, but I have spent most of my time with system administration for the last half year. Whenever I have a new task in an unknown area, I always start with reading the relevant chapter from her book. I usually read a few other, unrelated sections too. As I have read this sad post, her book was actually lying next to me, open on page 938.

So 20 million Yahoo user names are revealed. Why is that interesting at all? I guess if I write a script which loops some id for a yahoo info page I get a similar list. Maybe a Google search is enough. Or do not contact external service, just guess: take all Japanese names, append one or two digits to it. Mostly these are valid names.

25 W idle is quite impressive, could you describe what parts and other measures have you used to achieve it? Now I have 3 computers online 24h in my home alone, mostly idle. Two are about 60W at idle. The third is a Raspberry Pi, which only consumes a few watts, but it has an attached LCD display too.

Regarding servers: I believe the new servers has much better energy management, so the idle - full difference may be larger than once it was. I also found that it is very hard to achieve a - seemingly - full load, the type of test program makes a big difference in power consumption, even if top displays 100% on all cores in every cases.

I agree, the wattage must be measured. As another example, I measured 80W idle and 160 W on full load on a 12 core Opteron server.

This server does not have a 3D video card. I guess you also achieved full load on a 3D card too, and that explains the large difference between the wattage in idle and full load in your case.

I mean "...I have never seen mentioned..."

AMD already has FMA3. They also published great results. Of course nobody read it, at least I have seen mentioned it in the usual generic benchmark articles people like to refer (which does not use FMA3).