BCD mode is used extensively in COBOL and in the banking industry. Also, conversions between binary and decimal numbers are really slow on processors that lack a multiply, divide or mod instruction. If the only divides and multiplies are by 10, then BCD math is quite competitive on an 8-bit processor. With the right workload, it is also competitive on some 16-bit processors.

Many older and/or embedded processors lack a fast or single-cycle multiply and divide instructions. For instance, 8080A, 8085, 8088, 8086, 80186, 80286, 80386, 68000-68020, Microchip PIC, Z80, Z8, 8052 and successors, they almost all lack really fast multiply and divide instructions. The 80386 used 9 to 41 clock cycles on multiply. The dsPIC33E is a relatively modern embedded DSP from MicroChip. It takes 18 clock cycles to do a 16-bit divide. I try hard to avoid a 32-bit divides in critical real-time code.

Some markets naturally favor monopolies. Telecommunications is a good example. 23 years after the breakup of AT&T, the phone system, internet and cable systems in the US are back to being monopolies in many areas. The lucky areas have two or three near-monopolistic competitors, and these competitors behave suspiciously like cartels.

Economics 101: Free markets only work under specific conditions. In this case, a free market requires low barriers to entry. Telecommunications has huge capital cost expenses that decline with the number of customers served. Thus, a monopoly that actively excludes competitors can maximize profits. If new entrants enter the marketplace, the monopoly can cut prices sufficiently that they can always bankrupt the new entrant, and continue to make a profit.

This is also why states have laws blocking municipalities from offering Internet. Once a municipality builds the infrastructure, the resulting system is almost guaranteed to be profitable. As such, the big telcos hire lobbyists to pass laws to prevent construction of such systems, as they will be long-term competitors against the big telcos.

Are you saying that the more advanced the Windows UI, the more power users should and will use the keyboard? The entire GUI premise is flawed if the strategy is to revert to keyboard shortcuts.

I think Microsoft's introduction of Windows 8 and the Office Ribbon have been so badly bungled that many power users have simply reverted to keyboard shortcuts. However, as a strategy, I don't think it is a good idea. Why even have the mouse when we can all go back to command line?

The sub-fields of electrical engineering are not that different. Electrical engineering is about two things:
a) Maxwell's equations.
b) Mathematical Methods to use those equations.
This can be clearly seen if you do a course in Microwave Engineering, and if the course covers Maxwell's equations, capacitance, inductance, and how they are related in transmission line and waveguide theory.

After covering Microwave engineering, it becomes obvious that a significant crossover exists between the following electrical engineering specialities:
1. High Power - Transcontinental power transmission lines follow the same rules as microwave transmission lines. It is just the geometry and wavelengths are far longer.
2. Motor Drives - Same inductance and capacitance problems, particularly when dealing with high-frequency switching power supplies driving much slower motors through cables. Ferromagnetism shows up in motor drives.
3. Power Supplies - Same as motor drives. Strong resemblance to AC/DC and DC/AC power conversion in high power electrical grid systems.
4. Circuit board design - Modern high-frequency circuit boards are all about transmission line theory.
5. IC design - Change the materials. All the theory is back again. Now you are applying Maxwell's equations at much smaller scales.
6. RF design - This is exactly what the microwave theory course is about.
7. Laser and Opto-electronic design - Maxwell's equations are back again. Frequencies, electron voltage changes, etc.

Electrical engineering is about two "simple" subjects: Maxwell's equations, and mathematical methods. Most electrical engineering projects devolve into a combination of:
a) something involving electro-magnetic theory and/or it's formal mathematical solutions, like Laplace, z-transform, Fourier Theory, and Wavelet theory, and
b) something involving Boolean logic, and/or implementations of Turing machines.

Electrical engineering is different from the rest of the engineering fields. In Electrical, there is only the four Maxwell equations, tons of mathematical abstractions, simplifications, solutions, methods and techiques, and all of the implementations and ramifications of the them. For Mechanical engineers, there is no set of unifying equations. Chemical engineers have VESPR and thermo-dynamics, but that only goes so far. Aerospace engineers have a set of CFD assumptions, but those assumptions only hold in gases, and get strange when chemical reactions and/or phase changes are involved. Civil engineers have a basic set of equations to cover the simple stuff, but the complex problems involve sophisticated mechanical engineering. In Engineering Physics, they cover the electrical stuff, plus the quantum equations (which are a mess.) Electrical engineering is the only field of engineering with only 4 equations, and tons and tons of math to simplify their solutions.

Bingo. Interlnk.exe and intersvr.exe came with DOS, and as such, are often already on the old computers hard drive. (This is really important if the floppy drive is unreliable.) They can also be copied over the serial lines with a copy command, so on the other end, you can run Windows XP or DosBox under Linux.

Once the file system is under a pseudo-modern O/S, then you can use TCP/IP networking to copy it wherever you like.

C++. All of programming in one language - except simplicity.

"When programming in C++, you have to pick the subset of the language that you want to use. Otherwise your code will be unreadable."

Canadian police are very polite. For a minor first infraction, they will apologize for interrupting your activities before announcing that if you do not stop then they may need to arrest you. This is known as a "caution." For a larger first infraction, they will apologize for interrupting your activities before delivering a search warrant and arresting you.

If the Canadian police show up with guns drawn, then you screwed up big-time.

The level of network hacking against servers and internet systems is somewhat astonishing, and not widely known outside the industry.

I did a small project where a small company wanted to monitor our equipment on a very small fleet of cars. One day, I discovered we were getting telemetry data from our cars. This created much excitement and surprise in the office, as no one was supposed to be driving any of our cars. After a bit more work, I discovered the car in question was in China. Now that was a surprise ...

We only need to measure the mass of a 9.10938291 × 10^-31 kilogram particle accurate to 1 part in 10^-37. Alternatively, we can speed the electron up to 0.999c so it weighs more, then entangle it, and then measure it's mass to 1 part in 10^-37, with less than 5 sigma of measurement error.

Either way, I should have it done by lunch time.

All that fuel to burn on Titan, and no air ...

The perfect inter-planetary refueling depot, and no oxidizer ...

What to do???

The Google vs. Oracle lawsuit made a business case for not-invented-here syndrome. I think every major platform vendor will have there own programming languages in the future. Custom APIs and programming languages stops entire classes of patent/copyright lawsuits dead. It stops developers from moving between eco-systems. It even prevents your employees from stealing top-secret software and moving to a competitors. (And if they do steal the software, it becomes really obvious when law-enforcement shows up.)

I do agree from a portability/programmer perspective, NIH programming sucks. However, the legal perspective - it's great!

Also, the funny thing with lawsuits - even if you win, you still lose.

If you think that much about the answer, then the polygraph will assume you are lying.

False positives create selection bias. A polygraph detects people that are *nervous about there lies*. It won't detect the unaware and clueless, because they do not know they did anything wrong. Most people download songs to their iPhone, and assume it is legal. The polygraph not detect people that assume they are innocent. On the other hand, some people lie all of the time. A sociopath will pass the lie-detector test because they don't believe they are lying, and one person in 25 is a sociopath.

These problems have already been encountered in the preemployment screening industry. This is one of the less biased artlicles. To quote:

One recent study found faked answers for one quarter to one half of the applicants.[44] So how can employers who want to use personality or EQ tests in their selection process mitigate against the risk of applicant faking? Counter-measures to faking include the test and retest approach to see if an individual is consistent in their answers, or asking questions that require quick responses.[45] But counter-measures to faking may result in less reliable and valid results since some tools used to detect faking do not work well.[46]

Bluntly, if your goal is to hire people that have done no wrong, then chances are that your hires have either lied to you, or are too clueless to realize their mistakes. Either way, it is really bad for the employer, especially if the employer is the FBI.

I recall working with numerical methods from about 40 years ago, and all of the calculations that required a call to sin were range reduced to the region of +/- pi/4 anyway. The reason is that the taylor series expansions for sine and cos are most accurate in the region of zero, and for values in excess of pi/4, it is more accurate to do a transformation and implement a different call.

It is likely that the serious numerical code already handle this condition inside the internal algorithms.