Age Expiry fails?

[methodz]

Hello,

Im using the following in my httpd.conf (apache 2.x):

ExpiresActive On
ExpiresByType image/gif "access plus 45 minutes"
ExpiresByType image/png "access plus 45 minutes"
ExpiresByType image/jpeg "access plus 45 minutes"
ExpiresByType text/css "access plus 45 minutes"
ExpiresByType text/javascript "access plus 45 minutes"
ExpiresByType application/x-javascript "access plus 45 minutes"
ExpiresByType application/x-shockwave-flash "access plus 45 minutes"

When i do the test i failed in the following as below:

Enable browser caching of static assets:
FAILED (45.0 minutes) - http://*****/finder/userfiles/images/banners/mbanner-jeepjam.jpg
FAILED (45.0 minutes) - http://*****/finder/userfiles/images/banners/mbanner_afg1.jpg
FAILED (45.0 minutes) - http://*****/finder/userfiles/images/banners/mbanner_iphone1.jpg
FAILED (45.0 minutes) - http://*****/finder/userfiles/images/banners/mbanner_leaders.jpg....etc

Why does it fail although it does recognize the 45 minutes age expiry?

Thanks!

[pmeenan]

Sorry, it doesn't fail to recognize it, it just doesn't think 45 minutes is terribly useful :-) Anything under 24 hours will fail and if it is between 24 hours and 30 days it will still get a warning.

[p83822]

(01-19-2011 11:01 PM)pmeenan Wrote:  Sorry, it doesn't fail to recognize it, it just doesn't think 45 minutes is terribly useful :-) Anything under 24 hours will fail and if it is between 24 hours and 30 days it will still get a warning.

Been waiting for someone to raise this for a while as I've always been confused by this one...

Given we reckon that a good proportion of users hit our sites with a clear cache (http://www.yuiblog.com/blog/2007/01/04/p...ch-part-2/)
then surely the big win is ensuring that the browser cache kicks in for at least the duration of the browsing. Long (>30 day caching) is irrelevant for half the users. But relevant for the other half, of course.

Long is good, but the grading doesn't fit the facts as I see them.

[pmeenan]

FWIW, the in-session cache logic for browsers is significantly different from the logic that is used when they come back in a new browser session so trying to tune the cache for their session doesn't really buy you much.

Even in the Yahoo study, 40-60% of users (and 80% of page views) were a cached experience and the performance improvements are enormous. The main point of the post is to not ignore the first-time experience and assume that your users have everything cached, but the same is true the other way - if you could make the experience 80-90% faster for 40-60% of the users it certainly deserves your attention :-)

Unless you're willing to accept your site being broken for 45 minutes at a time, you need to do the same work to allow for a 45-minute expiration as you would for a long expiration so why not?