On 7 September 2013 10:49, Jeremy Baron <jeremy@tuxmachine.com> wrote:

On Fri, Sep 6, 2013 at 5:31 PM, Risker <risker.wp@gmail.com> wrote:
> Yeah, I keep hearing those excuses for performance problems, Jeremy. It
> takes longer to serve up the original page here in North America on a fast
> connection - enough so that it is noticeable on a normal computer.

I don't know what that means. ("Original page"? does that mean it loads faster with a redirect than by hitting the canonical URL directly?)

Please provide enough details (steps, recipe, instructions, whatever you want to call it) so that someone else could repeat your experiment to verify your results.

Ideally we'd do that for both logged in and logged out users (and various combinations of prefs) but in the case of redirects for Shirley Temple Black and Chelsea Manning I think we mostly care about logged out users visiting the /wiki/${title} style URLs (so not people visiting &uselang= or &useskin= URLs) so let's focus on those. Which case were you testing?




Jeremy, this is not the "performance testing" list.  The paragraph you've written above is pretty well the definition of why women don't stick around wikipedia - they say something that to anyone else is obvious, but not to those who just cannot resist writing code into their responses.  You know why they call it code?  Because *most* people don't understand it.

The fact that you're entirely missing the point of this discussion by digressing into a proposal to test the speed of redirects vs canonical pages should generally be a hint that you're moving into your own comfort zone and leaving the rest of us behind. 

Risker