On 8.7.2014 20:45, Les Mikesell wrote: > On Tue, Jul 8, 2014 at 12:22 PM, Gilbert Sebenste > <sebenste at weather.admin.niu.edu> wrote: >> On Tue, 8 Jul 2014, Robert Moskowitz wrote: >> >>> and did the conversion for display to save another byte. Efficiency? >>> We were desperate for every byte we could squeeze out. the US Post >>> Office created a standard so that all US cities (and supposedly streets) >>> could be entered in 14 characters or less. We changed the abbreviation >>> of Nebraska from NB to NE (I remember writing that conversion program) >>> so we could more easily mix US and Canada addresses (those they would >>> not change their 6 character code to our 5 digit one). We burned CPU to >>> save storage. then rewrote key routines in assembler and hacked the >>> COBOL calls to make it all work. >>> >>> Things change. Design goals change. Systems have to change. >> Of course they do. And those were changes in efficiency that were the >> result of needed productivity improvements. Change for the sake of >> major improvement(s). Wonderful, well-designed, efficient AND necessary. >> And it obviously made things more productive for everyone! >> >> I argue that systemd neither improves efficiency, productivity or >> satisfaction...nor is it necessary. > More to the point, those 'old' efficiency hacks were from a time when > programmer time was cheaper than the computer resources. Now, the > computers should be doing the work for us instead of the other way > around. Can anyone really make the argument that we can't afford the > computer resources for transparent backwards compatibility now? > Yes. Computers have probably resources, but programmers time is expensive (or when doing it in your own time it isn't fun to do) so thats why people aren't implementing them. -vpk