Replies

If you would, when you post your findings please include specific info. The site, your connection speed, general system specs of your machine, and the type of network you are on. Are you using Firefox?

I have yet to see negatives with these changes personally, or from the feedback so far in this thread. I'm very interested in, and not predispositioned to dispute your claims.

Sure - I timed it myself last night after hunting around to find a page that would break the pipelining. 1.5 mbit DSL line, 2.8 GHz P4 with 768 MB memory running Windows 2000 SP4.

Keep in mind, I'm not saying there's no improvement ever, only that there are some cases when it doesn't improve things, some cases where it can make things slower, and that for most cases where it does improve things, the improvement is likely to not be earth-shattering for most people. On balance, it's probably worth enabling for most people, but it's the kind of thing that's worth a bit of testing to see if it helps in some particular case.

Okay, here's how we can test it. Presumably, you've already got pipelining enabled, so grab your stopwatch and get ready. The page I used is Google's translation page - it's nice, because it's got lots of little flag gifs, and even on a fast connection it'll take a bit to load, plus it's not some totally obscure page that nobody ever sees. First, let's clear your browser cache by going to "Tools" -> "Options" -> "Privacy" and hitting "Clear" next to "Cache". Once you've done that, start timing with your watch as soon as you click on the link, and as soon as the status bar at the bottom of the browser says "Done", stop the watch.

http://www.google.com/language_tools?hl=en

Got it? Now, disable http pipelining and http proxy pipelining in about:config - return to the default settings, basically. Clear your cache again and close and restart your browser. Come back to this page here on FR, grab your stopwatch, and time the page load again. You should notice a fairly significant difference in load times, although which one is faster may surprise you ;)