I have been using thttpd for 6-8 months now with a load of about 1-2 million hits a day. When viewing sites that use the image server with firebug I notice...
78
Steve Souders
steve_souders
Oct 9, 2007 5:21 am
The response times shown in Firebug's Net Panel should be taken with a grain of salt. If you need accurate times, I recommend using a packet sniffer like...
79
djspencer1981
Oct 11, 2007 8:25 pm
Would it be possible in a future release of the YSlow add-on to have a option of which rules are checked?...
80
Steve Souders
steve_souders
Oct 11, 2007 9:06 pm
Yes, we're considering ways to maintain a default YSlow score, while allowing users to customize the scoring algorithm. -Steve Souders ... a option of which...
81
srichardjr
Oct 22, 2007 5:31 pm
Our department is conducting performance tests comparing current pages to updated pages based on the the YSLOW rules. We are using Silk Performance tools to...
82
Steve Souders
steve_souders
Oct 22, 2007 5:49 pm
This is a great question. When I speak about "backend" I mean the time for the HTML document to reach the client, and "frontend"; is everything after that. With...
83
Rahul
rah_73_1999
Oct 23, 2007 7:18 am
Hi I find the following in the cached requests tab Empty Cache Primed Cache 43.9K 1 HTML document (est) 2.5K 3 Style Sheet Files 103.7K 19 JavaScript Files ...
84
Vinayak Hegde
vinayak.hegde
Oct 23, 2007 7:34 am
... Did you check what the requests were (say using liveHTTPheaders), They might be If-modified-since requests (which are low overhead) and might not actually...
85
Rahul P. Mahindrakar
rah_73_1999
Oct 23, 2007 8:13 am
Hi Vinayak I got it now, that content is not got below. And if I want to prevent the http request totally then I need to use the Cache-Control: max-age=xxx...
86
Vinayak Hegde
vinayak.hegde
Oct 23, 2007 8:27 am
... Yeah I think that should work. Cache-control headers should be honored by both intermediate caches (such as squid) and also user-agent (browser) caches....
87
Barry Hunter
b_b_hunter
Oct 25, 2007 6:55 pm
I've seen (and understand why) many places[1] suggest far-future expires header (or max-age) on images/css etc that dont change. I use this, but by default the...
88
Steve Souders
steve_souders
Oct 25, 2007 8:32 pm
This seems risky to me. Before doing it I would want to verify that it works on A grade browsers. Have you tested it? -Steve...
89
Barry Hunter
b_b_hunter
Oct 25, 2007 8:47 pm
Thanks for the response, at least echos my feelings, I trying it now on the css/js of the site. In IE7 (watching with Fiddler) and Firefox/LiveHTTPHeaders,...
90
Steve Souders
steve_souders
Oct 25, 2007 10:17 pm
Hit "reload" to force the browser to (re-)request the resource. If there was a last-modified header, the resource should be requested with an If-Modified-Since...
93
dasfreak_au
Oct 31, 2007 2:20 am
Just curious how YSlow knows a CDN is present when running an analysis. I work for Akamai but the opinoins below are mine and aren't necessarily those of my...
94
Steve Souders
steve_souders
Oct 31, 2007 4:48 am
Hi, Graeme. You've got it about right. There's no way for YSlow to detect if a request is coming from a CDN in realtime, for the exact reason you describe -...
95
Peeyush
pkul7de
Oct 31, 2007 11:12 am
Hi all, I am having a problem in understanding what 'Response Time' in Yslow status bar exactly corresponds to. According to Yslow faq: ...
96
Steve Souders
steve_souders
Oct 31, 2007 5:09 pm
The onload event doesn't fire until the page and all the components in the page are done downloading. This doesn't include downloads that are done...
97
Ernest Mueller
mxyzplkiv
Oct 31, 2007 7:18 pm
Is it the window's .onload or the document's .onload? Window I assume. And to further answer Peeyush's question, O'Reilly says "The onLoad() handler is...
98
Peeyush KULSHRESTHA
pkul7de
Nov 1, 2007 11:29 am
Ok, that means that in crude words we can say, Response time = (time for download from network) + (time for decoding of components, eg jpeg) + (time for actual...
99
mtmetten
Nov 2, 2007 1:08 pm
Hi, In our homepage (current yslow score 95%) we decided to cut an URL request to a 1x1 pixel image and instead use the data URL scheme: http://nestoria.co.uk/...
100
Jose Noheda
jose.noheda
Nov 2, 2007 1:15 pm
AFAIK data:url is unsupported in IE...
101
Steve Souders
steve_souders
Nov 2, 2007 4:00 pm
That's correct, IE doesn't support data: URLs. I'll add a bug for YSlow to recognize them and handle them better. Thanks! -Steve ... ...
102
srichardjr
Nov 14, 2007 10:54 pm
Our company is in the midst of analyzing the impact of 8 of the 12 rules for our site. The first three: GZIP, ETags and Minify JavaScripts, demonstrated no...
103
Steve Souders
steve_souders
Nov 15, 2007 6:11 am
The rules are in approximate priority order, so you'll see a greater improvement with with Rule 1 than Rule 13 (Etags). Also, it's good to get at least 100...
104
Joe Pearson
joewxboy
Nov 15, 2007 6:21 pm
... take ... When you say that they demonstrated no improvement, what do you mean? How are you defining success, and how are you measuring performance? We...
105
Ernest Mueller
mxyzplkiv
Nov 15, 2007 6:53 pm
So with Keynote/Gomez you have to be careful that your type of measurement will show the effects of the kind of improvement you're looking at. For Keynote,...
106
srichardjr
Nov 15, 2007 10:03 pm
Ernest, Joe and Steve, Thank you for your responses. The performance tool we are using is Silk Performer. The metrics we use match the definitions of frontend...
107
Steve Souders
steve_souders
Nov 16, 2007 12:35 am
Wow, this is a great thread. Thanks Ernest and Joe for your responses. Srichardjr - I would warn against putting too much value on numbers gathered from one...
108
Steve Souders
steve_souders
Nov 16, 2007 6:03 pm
[See the full blog on Yahoo! Developer Network: http://developer.yahoo.net/blog/archives/2007/11/velocity_web_pe.html] O'Reilly just announced Velocity ...
