[ale] End to end testing?

[Michael Trausch]

... Or just profile the thing, find the bottleneck, wash, rinse, and repeat, until you're happy.

Never ask an end user for the definition of fast enough. They have no idea what is possible or reasonable and will always say "immediate response" or something similar. Instead go for as fast as possible without reprogramming and be done with it. If programming is needed report that and either expand the job or outsource. Of course, I am available for custom programming, but refuse to touch Ruby or Perl (IMHO they're both convoluted messes, and I *STILL* cannot stand the runtime performance of any ruby software have used, preferring the much more predictable, well-defined, lower stress method of direct native coding in C or C++. Scripts are okay one level into a system. Any deeper than that and forget it unless you're rich. Ruby is for the people in the Sprint commercials. ;)

Sent from my iPad

> On May 21, 2015, at 11:22 AM, leam hall <leamhall at gmail.com> wrote:
> 
> I'm leaning towards the latter at the moment. A problem might be "the website is slow". The solution might focus on "define 'fast enough'" and then map out the various site components so the end result is understood.