Page load times are down because pages are loading so many more tracking options and some of them are very abusive on the javascript engines. If you turn on the status line (even if you can as it is gone in some modern browsers), you will often see it saying "loading 159 out of 162" and those last ones never load. There is also something that is related to a compounding latency problem that many developers don't think about it because they don't see it when they are developing the platforms and modern tool kits help to hide it from developers too.

I guess people don't like IBM's old work on the subject that showed dropping a 3 second response to just 2 seconds resulted in substantial improved efficiency. Maybe marketing groups need to understand that a customer stuck on a slow site is a bad consumer.

When your in the business of selling random numbers, don't be surprised when someone undercuts you.

When someone calls your 1-800 number, you pay someone. That someone gives a cut of it to other parties. One of those parties may have picked your number for a reason. This can work in a way that is similar to the "False Answer Signalling" fraud that was so common years ago .

If the data fits in a database, it is not Big Data.

Google Wave was some very interesting technology but I think it was pointed in the wrong direction. It wasn't ever considered as a way for the Internet of Things to talk to each other in a reasonable way and everything Google did with it was around pushing it in front of people, not devices. There might have been a time where the terms of service didn't allow devices to communicate using it for message passing.

Language development these days seems to have stalled around a few specific areas which are separated by the family trees of the languages. There were plenty of languages from the past that excelled in one area or another that is often difficult or every inefficient with more modern architectures. Many of the 4th generation languages had some very cleaver ideas. Even new languages like go don't help with a major problem because there is no base type for money so programmers are left to their own devices and end up with floating point issues or just using cents or even worse problems because of the underlying hacks.

The limiting factor of shared fiber broadband is the packet turnaround time just like coax and radio combined with scheduling the upstream data. The *PON networks were designed for sending lots of cable TV bits one direction and being able to cope with a small percentage going the other way. There are all sorts of techniques to fix that problem and all of them fail in different ways. So far the fastest home internet isn't PON based but a dedicated point to point links to a somewhat local fiber switch that has massive amounts of upload. It would be very interesting for Google to release some documents about their different types of technology they are using in Kansas City experiment. I've heard that they are using at least 4 different types of connections.

German and English won in the engineering world because of compound words. You can invent a new device and create a name that works in letter describing it.

English wins over German because of the relative lack of gendered words. Genders can get very messed up when using compound words. As an example, if a boat is female and a trailer is male, what gender should a boat-trailer be?

I haven't flow in the US in the last year. I've been on commercial aircraft in Australia where the pilot got the wrong frequency when the controller was using "dec-ee-mal". A friend had his class do an experiment where students wrote down numbers that were being read in different styles. There were substantially more errors with the ICAO way of reading numbers than the older FAA style with the Aussie students.

In the US they use "point" which is one syllable. There is no place in aviation radio where the decimal point isn't implied which makes using "decimal" a bigger waste of radio time.

You may not have a choice. My last power bill had a connection charge that was higher than the energy consumption charge an I pay $.22 a kwh. That will be the trend in the future. In places where the grid is still locally owned, I see it being added to property taxes as the cost of batteries come down where people can go off grid.

We just put in 6 250W panels. They cost less then $190 each but installing the frame and the wiring cost more. The mPPT module happens to plug into our existing telco grade -48V DC power supply and it was only $800 but plugged into a nice $5k system. The batteries that will run one of our racks of gear for 8 hours cost $250 each for 8 of them. The silicon bits aren't a major part of the cost of going off grid now.

The USDA's budget is 100x that of the BLM. Sure one is dept of Ag and the other is Dept of Interior but I'm not sure it matters much since I think the USDA has claim to all BLM land as well.

Work recently spent about $5k for cube and it isn't printing any better than the 4 other 3d printers I've used 3 of them costs less than $1000.

You can't do this with systemd. A kill to a process group is an atomic operation in Posix so that if you do a kill -9 -1 (i.e. send a SIGKILL to init and all of its children), the kernel will not return from the "kill" syscall until it has sent the signal to all of the processes. That syscall will also prevent any other task switches until it is done so the result is no process (other than init) ever runs again even if they are in the middle of a forkbomb. A kill -1 -1 (send SIGHUP to everything via init's process group) has traditionally told all user level programs that the user logged out and all daemons that they should reload their config files.

Killing a process group (the negative process id, which is what the original commentator was talking about, not a SIGHUP) is used all the time on systems. That is how apachectl (and most other forking deamons and their control programs) tell its children to reload the config file or end in a controlled way. It is used every time a user logs out to make sure all their processes do go away. Signals are the oldest and more reliable of the IPC mechanisms and are great when the number of messages you need to send is a tiny number of options.

Over the past 3 decades, versions of the inittab syntax allowed for things in the 2nd and 3rd fields to say things could be run in the background or depend on other named states which is why the 1st field is a name.

It is amazing how many properly run systems can cope with a "kill -1 -1" to reset everything without a reboot.

Yet a 1 mm tick plastic bezel around the glass would be nearly invisible and protect the glass too. If done right, it might even make manufacturing cheaper.