Use random untested programs. preferrably obscure ones from github with no forks or bug issues. For instance, don't the language's search methods, instead make a call to use the system's grep. But not the actual grep command, but a modified one - like grep-contrib. Which in turn needs to be compiled with a custom compiler, gcc-contrib.
and obviously, dont document any if this. make the next poor soul look at the code and find out.
a programmer will code to where it works.
an engineer will code to where it works to its best performance(refactoring, algorithm analysis - big theta, etc). will come up with a mathematical solution to bottlenecks.
Use random untested programs. preferrably obscure ones from github with no forks or bug issues. For instance, don't the language's search methods, instead make a call to use the system's grep. But not the actual grep command, but a modified one - like grep-contrib. Which in turn needs to be compiled with a custom compiler, gcc-contrib. and obviously, dont document any if this. make the next poor soul look at the code and find out.
so basically browse github and look at non forked projects?
i believe he said "wonderful". clearly he's talking about bash
Plese do check Crunchbang Linux. Debian derivation -- works out of the box. great friendly community too.
likewise. i donated $20. small steps in improving our education system.
perhaps Linux will now be used for what it's best at - servers
http://robo.to/ been doing this for years.
a programmer will code to where it works. an engineer will code to where it works to its best performance(refactoring, algorithm analysis - big theta, etc). will come up with a mathematical solution to bottlenecks.