I'm using it on my personal site, http://www.iangreenleaf.com, but it
hardly counts as "production" :)

It's a very nice library, and small enough that you can read and understand
the full source in a couple of hours. There are a few small fiddly UI
things that may or may not be a problem for you - I ran into problems with
Firefox leaving little dotted borders around my links, but was able to work
around it. Pushstate isn't supported in a few important platforms (not only
IE, but iOS4), so you'll want to have graceful degradation. But no big
gotchas that I know of.

It definitely *feels* snappier - I think just not having the momentary
blank page makes a huge difference in perception. I'd be curious to hear
what you come up with in benchmarks, though. I don't know if there's
actually a big difference in absolute response time, or if it's just how it
feels to the user. Probably the only way to get a useful value would be to
use a headless browser like Selenium, and time how long it takes after a
click to display a piece of text or something.

Ian

On Mon, Apr 9, 2012 at 4:26 PM, Ylan Segal <[email protected]> wrote:

> I have been playing around with PJAX after reading a mention of it in a
> description of how Basecamp got faster:
>
>
> https://37signals.com/svn/posts/3112-how-basecamp-next-got-to-be-so-damn-fast-without-using-much-client-side-ui
> http://railscasts.com/episodes/294-playing-with-pjax
> https://github.com/defunkt/jquery-pjax
>
> The author can describe what it does better than me:
>
> "pjax loads HTML from your server into the current page without a full
> reload. It's ajax with real permalinks, page titles, and a working back
> button that fully degrades."
>
> After playing around for it for a bit, it definitely looks like this could
> be used to make websites snappier. Is anyone using this in production? Any
> know issues / gotchas to be aware of?
>
> Before I go any further, I would like to benchmark the site with and
> without PJAX, to have a sense of potential savings in page load time.
> Anyone have an idea how I could benchmark this? The javascript on the page
> has to be executed for pjax to kick-in.
>
> After some benchmarking, I think I would a/b test this on a production
> site to see if faster laod times increases conversions, which seems to be
> the consensus, but never hurts to measure for yourself, right?.
>
> Thanks for any pointers,
>
> --
> Ylan
>
> --
> SD Ruby mailing list
> [email protected]
> http://groups.google.com/group/sdruby

-- 
SD Ruby mailing list
[email protected]
http://groups.google.com/group/sdruby

Reply via email to