exceptional-performance : Messages : 1688-1719 of 2065

#1688

From: Sergey Chernyshev <sergey.chernyshev@...>
Date: Sun Oct 3, 2010 11:56 pm
Subject: Web Performance Automation, Joshua Bixby @ New York Web Performance Meetup

sergeycherny...

Send Email

If you're in New York or around, save the date! It's going to be a great session, come join us!

On October 13th, Joshua Bixby (@JoshuaBixby), President of Strangeloop Networks and author of Web Performance Today blog will talk about Web Performance Automation.

�� RSVP here:�http://www.meetup.com/Web-Performance-NY/calendar/14393683/

Abstract:
Depending on who you talk to, performance automation can range from the holy grail of performance tuning to a set of programming hacks that will bring down your website. In reality, it is neither of these things.

Joshua will talk about:�

Different approaches to automation�
Pros and cons of each�
What kinds of performance best practices can or should be automated�
What best practices are better left to hands-on developers�
How �automatic� is performance automation, actually?�
A survey of the automation market today

Agenda:�
6:00 - Arrive to the event, meet other members�
6:15 - Introduction to Web Performance�
6:30 - Web Performance Automation, Joshua Bixby�
8:00 - Q&A�
8:30 - Open Discussion, Networking

Time Inc.�offered their Watercooler meeting to host NY Web Performance meetup.
Host contact:�Alla Gringaus, Web Technology Fellow�
Questions:�646 391 9671

Strangeloop Networks�is sponsoring�pizza and sodas!

Location:
Rockefeller Center Time & Life Bldg

1271 Ave of the Americas (entrance 50th St and 6th Ave)

2nd Floor, room #1

New York, NY 10020

Nearest Transit:�
50th St - 8th Ave (C, E)�
50th St - Ave of the Americas (F, V, B, D)�
49th St - 7th Ave (N, R, W)

Get directions on Google Maps:�http://bit.ly/c5rTgs
Street View for the Entrance:�http://bit.ly/bpxpOW

You'll need to bring a photo ID for security downstairs.

Don't miss it, see you there on Wednesday, October 13th at�6:00.

Thank you,

� � � �� Sergey

--
Sergey Chernyshev
http://www.sergeychernyshev.com/

http://www.sergeychernyshev.com/blog/

http://www.showslow.com/

Reply | Messages in this Topic (3)

#1689

From: "Screen Rant" <vholtreman@...>
Date: Tue Oct 5, 2010 11:16 pm
Subject: Site crashing for months - a site performance recommendation was the cause

vholtreman

Send Email

I'd been pulling my hair out since early May of this year trying to figure out
why my Wordpress site would randomly crash the Firefox 3.X browser. I thought
for sure it was some rogue ad from a shady advertiser via remnant display. Then
I thought maybe there was some plugin conflict of some sort.

Literally months of troubleshooting - disabling entire ad networks, uninstalling
Flash, harassing my sales rep to track down the problem since it must be an ad,
tearing apart my templates, etc.

In the end it turned out to be the addition of this just above my opening <body>
tag:

<?php flush(); ?>

I had added that to my template based on the Best Practices document, but had
forgotten to put two and two together. I finally went back through my emails to
see when I started noticing Firefox crashes coming out of the blue, and noticed
that I added the above to my site a couple of days prior.

I spoke to a developer after the fact and he remarked "Oh yeah, flush has always
been unstable."

So thanks so much for a recommendation that caused me five months of angst and
annoyed countless visitors to my site. Thank goodness I finally found the
problem and corrected it. You might want to put some sort of caveat in your
document.

On a side note, my site did not crash in Firefox 4 Beta, if that's any help to
anyone.

Sincerely,

Vic

Reply | Messages in this Topic (4)

#1690

From: "Patrick Meenan" <PatMeenan@...>
Date: Wed Oct 6, 2010 1:01 am
Subject: RE: Site crashing for months - a site performance recommendation was the cause

pmeenan

Send Email

You must have been triggering some edge case with Firefox because there is no way that should even be possible (much less make any sense). If I had to put money on it I would actually bet that one of your Firefox plugins doesn't know how to deal with chunked encoding or partial streams of data. flush() just starts sending the part of the document that is available while the rest of the html is generated. For it to actually work the server needs to switch to chunked encoding but there's nothing special about it, it's part of the spec and every browser I have seen knows how to properly handle it (and by definition any app that talks TCP has to know how to deal with partial data and accumulate it until the correct amount of data has arrived.

Flush has always been difficult to get working correctly but that is on the server side and is more of an Apache problem than something specific to flush itself. I've never heard it referred to as "unstable". Firefox (particularly by the time it got to 3.x) has been tested out the wazoo for chunked encoding and flushed documents and if there was a browser problem with it then every search engine would also crash the browser.

-Pat

From: exceptional-performance@yahoogroups.com [mailto:exceptional-performance@yahoogroups.com] On Behalf Of Screen Rant
Sent: Tuesday, October 05, 2010 7:17 PM
To: exceptional-performance@yahoogroups.com
Subject: [exceptional-performance] Site crashing for months - a site performance recommendation was the cause

I'd been pulling my hair out since early May of this year trying to figure out why my Wordpress site would randomly crash the Firefox 3.X browser. I thought for sure it was some rogue ad from a shady advertiser via remnant display. Then I thought maybe there was some plugin conflict of some sort.

Literally months of troubleshooting - disabling entire ad networks, uninstalling Flash, harassing my sales rep to track down the problem since it must be an ad, tearing apart my templates, etc.

In the end it turned out to be the addition of this just above my opening <body> tag:

<?php flush(); ?>

I had added that to my template based on the Best Practices document, but had forgotten to put two and two together. I finally went back through my emails to see when I started noticing Firefox crashes coming out of the blue, and noticed that I added the above to my site a couple of days prior.

I spoke to a developer after the fact and he remarked "Oh yeah, flush has always been unstable."

So thanks so much for a recommendation that caused me five months of angst and annoyed countless visitors to my site. Thank goodness I finally found the problem and corrected it. You might want to put some sort of caveat in your document.

On a side note, my site did not crash in Firefox 4 Beta, if that's any help to anyone.

Sincerely,

Vic

Reply | Messages in this Topic (4)

#1691

From: Ryan Doherty <ryan.doherty@...>
Date: Wed Oct 6, 2010 5:14 pm
Subject: Re: Site crashing for months - a site performance recommendation was the cause

ryan.doherty

Send Email

I agree with Patrick, and if it's a reproduceable crash (without plugins) please file a bug: http://bugzilla.mozilla.org . You can get a link to your crash report by typing 'about:crashes' into the URL bar in Firefox.

Ryan Doherty

On Tue, Oct 5, 2010 at 6:01 PM, Patrick Meenan <PatMeenan@...> wrote:

�

You must have been triggering some edge case with Firefox because there is no way that should even be possible (much less make any sense).� If I had to put money on it I would actually bet that one of your Firefox plugins doesn't know how to deal with chunked encoding or partial streams of data.� flush() just starts sending the part of the document that is available while the rest of the html is generated.� For it to actually work the server needs to switch to chunked encoding but there's nothing special about it, it's part of the spec and every browser I have seen knows how to properly handle it (and by definition any app that talks TCP has to know how to deal with partial data and accumulate it until the correct amount of data has arrived.

Flush has always been difficult to get working correctly but that is on the server side and is more of an Apache problem than something specific to flush itself.� I've never heard it referred to as "unstable".� Firefox (particularly by the time it got to 3.x) has been tested out the wazoo for chunked encoding and flushed documents and if there was a browser problem with it then every search engine would also crash the browser.

-Pat
�
From: exceptional-performance@yahoogroups.com [mailto:exceptional-performance@yahoogroups.com] On Behalf Of Screen Rant
Sent: Tuesday, October 05, 2010 7:17 PM
To: exceptional-performance@yahoogroups.com
Subject: [exceptional-performance] Site crashing for months - a site performance recommendation was the cause

�
�
I'd been pulling my hair out since early May of this year trying to figure out why my Wordpress site would randomly crash the Firefox 3.X browser. I thought for sure it was some rogue ad from a shady advertiser via remnant display. Then I thought maybe there was some plugin conflict of some sort.

Literally months of troubleshooting - disabling entire ad networks, uninstalling Flash, harassing my sales rep to track down the problem since it must be an ad, tearing apart my templates, etc.

In the end it turned out to be the addition of this just above my opening <body> tag:

<?php flush(); ?>

I had added that to my template based on the Best Practices document, but had forgotten to put two and two together. I finally went back through my emails to see when I started noticing Firefox crashes coming out of the blue, and noticed that I added the above to my site a couple of days prior.

I spoke to a developer after the fact and he remarked "Oh yeah, flush has always been unstable."

So thanks so much for a recommendation that caused me five months of angst and annoyed countless visitors to my site. Thank goodness I finally found the problem and corrected it. You might want to put some sort of caveat in your document.

On a side note, my site did not crash in Firefox 4 Beta, if that's any help to anyone.

Sincerely,

Vic

Reply | Messages in this Topic (4)

#1692

From: Marcel Duran <contact@...>
Date: Wed Oct 6, 2010 5:42 pm
Subject: Re: Site crashing for months - a site performance recommendation was the cause

marcelduran

Send Email

The PHP manual describes some browser caveats with flush() (http://us.php.net/flush) if your server is using plain vanilla php, none of these caveats mention browser crashing.

I think it's not fair blaming the performance best practices, otherwise as Patrick said, any search engine would be crashing browsers on every single query. I suggest you investigate a bit deeper and file a bug for mozilla or whatever plugin might be causing the issue.

Marcel

On Wed, Oct 6, 2010 at 10:14 AM, Ryan Doherty <ryan.doherty@...> wrote:

�

I agree with Patrick, and if it's a reproduceable crash (without plugins) please file a bug: http://bugzilla.mozilla.org . You can get a link to your crash report by typing 'about:crashes' into the URL bar in Firefox.

Ryan Doherty

On Tue, Oct 5, 2010 at 6:01 PM, Patrick Meenan <PatMeenan@...> wrote:

�

You must have been triggering some edge case with Firefox because there is no way that should even be possible (much less make any sense).� If I had to put money on it I would actually bet that one of your Firefox plugins doesn't know how to deal with chunked encoding or partial streams of data.� flush() just starts sending the part of the document that is available while the rest of the html is generated.� For it to actually work the server needs to switch to chunked encoding but there's nothing special about it, it's part of the spec and every browser I have seen knows how to properly handle it (and by definition any app that talks TCP has to know how to deal with partial data and accumulate it until the correct amount of data has arrived.

Flush has always been difficult to get working correctly but that is on the server side and is more of an Apache problem than something specific to flush itself.� I've never heard it referred to as "unstable".� Firefox (particularly by the time it got to 3.x) has been tested out the wazoo for chunked encoding and flushed documents and if there was a browser problem with it then every search engine would also crash the browser.

-Pat
�
From: exceptional-performance@yahoogroups.com [mailto:exceptional-performance@yahoogroups.com] On Behalf Of Screen Rant
Sent: Tuesday, October 05, 2010 7:17 PM

To: exceptional-performance@yahoogroups.com
Subject: [exceptional-performance] Site crashing for months - a site performance recommendation was the cause

�
�
I'd been pulling my hair out since early May of this year trying to figure out why my Wordpress site would randomly crash the Firefox 3.X browser. I thought for sure it was some rogue ad from a shady advertiser via remnant display. Then I thought maybe there was some plugin conflict of some sort.

Literally months of troubleshooting - disabling entire ad networks, uninstalling Flash, harassing my sales rep to track down the problem since it must be an ad, tearing apart my templates, etc.

In the end it turned out to be the addition of this just above my opening <body> tag:

<?php flush(); ?>

I had added that to my template based on the Best Practices document, but had forgotten to put two and two together. I finally went back through my emails to see when I started noticing Firefox crashes coming out of the blue, and noticed that I added the above to my site a couple of days prior.

I spoke to a developer after the fact and he remarked "Oh yeah, flush has always been unstable."

So thanks so much for a recommendation that caused me five months of angst and annoyed countless visitors to my site. Thank goodness I finally found the problem and corrected it. You might want to put some sort of caveat in your document.

On a side note, my site did not crash in Firefox 4 Beta, if that's any help to anyone.

Sincerely,

Vic

--
Marcel Duran

Reply | Messages in this Topic (4)

#1693

From: "Steve Robillard" <s.robillard@...>
Date: Fri Oct 8, 2010 1:18 pm
Subject: RE: Web Performance Automation, Joshua Bixby @ New York Web Performance Meetup

n1taz67

Send Email

Sergey,

Any chance this will be recorded. I would love to see what else in the deployment process I can automate besides minimizing CSS, JS and optimizing images.

Thanks,

Steve

From: exceptional-performance@yahoogroups.com [mailto:exceptional-performance@yahoogroups.com] On Behalf Of Sergey Chernyshev
Sent: Sunday, October 03, 2010 7:56 PM
To: exceptional-performance@yahoogroups.com
Subject: [exceptional-performance] Web Performance Automation, Joshua Bixby @ New York Web Performance Meetup

If you're in New York or around, save the date! It's going to be a great session, come join us!

On October 13th, Joshua Bixby (@JoshuaBixby), President of Strangeloop Networks and author of Web Performance Today blog will talk about Web Performance Automation.

RSVP here: http://www.meetup.com/Web-Performance-NY/calendar/14393683/

Joshua will talk about:

Different approaches to automation
Pros and cons of each
What kinds of performance best practices can or should be automated
What best practices are better left to hands-on developers
How “automatic” is performance automation, actually?
A survey of the automation market today

Agenda:
6:00 - Arrive to the event, meet other members
6:15 - Introduction to Web Performance
6:30 - Web Performance Automation, Joshua Bixby
8:00 - Q&A
8:30 - Open Discussion, Networking

Time Inc. offered their Watercooler meeting to host NY Web Performance meetup.
Host contact: Alla Gringaus, Web Technology Fellow
Questions: 646 391 9671

Strangeloop Networks is sponsoring pizza and sodas!

Location:
Rockefeller Center Time & Life Bldg

1271 Ave of the Americas (entrance 50th St and 6th Ave)

2nd Floor, room #1

New York, NY 10020

Nearest Transit:
50th St - 8th Ave (C, E)
50th St - Ave of the Americas (F, V, B, D)
49th St - 7th Ave (N, R, W)

Get directions on Google Maps: http://bit.ly/c5rTgs
Street View for the Entrance: http://bit.ly/bpxpOW

You'll need to bring a photo ID for security downstairs.

Don't miss it, see you there on Wednesday, October 13th at 6:00.

Thank you,

Sergey

--
Sergey Chernyshev
http://www.sergeychernyshev.com/

http://www.sergeychernyshev.com/blog/

http://www.showslow.com/

Reply | Messages in this Topic (3)

#1694

From: Sergey Chernyshev <sergey.chernyshev@...>
Date: Fri Oct 8, 2010 4:13 pm
Subject: Re: Web Performance Automation, Joshua Bixby @ New York Web Performance Meetup

sergeycherny...

Send Email

I'll try to record it, but can't promise, unfortunately.

� � � �� Sergey

On Fri, Oct 8, 2010 at 9:18 AM, Steve Robillard <s.robillard@...> wrote:

�

Sergey,
�
Any chance this will be recorded. I would love to see what else in the deployment process I can automate besides minimizing CSS, JS and optimizing images.

�
Thanks,
Steve

�
From: exceptional-performance@yahoogroups.com [mailto:exceptional-performance@yahoogroups.com] On Behalf Of Sergey Chernyshev
Sent: Sunday, October 03, 2010 7:56 PM

To: exceptional-performance@yahoogroups.com
Subject: [exceptional-performance] Web Performance Automation, Joshua Bixby @ New York Web Performance Meetup

�
�
If you're in New York or around, save the date! It's going to be a great session, come join us!

�
�
On October 13th, Joshua Bixby (@JoshuaBixby), President of Strangeloop Networks and author of Web Performance Today blog will talk about Web Performance Automation.

�� RSVP here:�http://www.meetup.com/Web-Performance-NY/calendar/14393683/

Abstract:
Depending on who you talk to, performance automation can range from the holy grail of performance tuning to a set of programming hacks that will bring down your website. In reality, it is neither of these things.

�
Joshua will talk about:�
Different approaches to automation�
Pros and cons of each�
What kinds of performance best practices can or should be automated�
What best practices are better left to hands-on developers�
How �automatic� is performance automation, actually?�

A survey of the automation market today
Agenda:�
6:00 - Arrive to the event, meet other members�
6:15 - Introduction to Web Performance�
6:30 - Web Performance Automation, Joshua Bixby�
8:00 - Q&A�
8:30 - Open Discussion, Networking
�
Time Inc.�offered their Watercooler meeting to host NY Web Performance meetup.
Host contact:�Alla Gringaus, Web Technology Fellow�
Questions:�646 391 9671

Strangeloop Networks�is sponsoring�pizza and sodas!
Location:
Rockefeller Center Time & Life Bldg
1271 Ave of the Americas (entrance 50th St and 6th Ave)
2nd Floor, room #1
New York, NY 10020
Nearest Transit:�
50th St - 8th Ave (C, E)�
50th St - Ave of the Americas (F, V, B, D)�
49th St - 7th Ave (N, R, W)

Get directions on Google Maps:�http://bit.ly/c5rTgs
Street View for the Entrance:�http://bit.ly/bpxpOW
�
You'll need to bring a photo ID for security downstairs.

Don't miss it, see you there on Wednesday, October 13th at�6:00.

Thank you,

� � � �� Sergey

--
Sergey Chernyshev
http://www.sergeychernyshev.com/

http://www.sergeychernyshev.com/blog/
http://www.showslow.com/

�
�

Reply | Messages in this Topic (3)

#1695

From: Philip Tellis <philip@...>
Date: Tue Oct 12, 2010 3:46 pm
Subject: need opinions on measuring first byte time

philiptellis

Send Email

We're trying to come up with a standardised way of measuring first-byte
time in boomerang.  It would be cool if people here could add their
opinion here:
http://developer.yahoo.net/forum/?showtopic=7296

Philip

Reply | Messages in this Topic (1)

#1697

From: "bettyytso" <bettyytso@...>
Date: Tue Oct 19, 2010 6:47 am
Subject: YSlow 2.1.0 available - offers major bug fixes

bettyytso

Send Email

YSlow 2.1.0 is now available (courtesy of Marcel Duran).  Detailed release notes
please refer to YDN blog:
    
http://developer.yahoo.com/blogs/ydn/posts/2010/10/yslow-2-1-0-release-offers-ma\
jor-bug-fixes/

Download it here:
     https://addons.mozilla.org/en-US/firefox/addon/5369/

Twitter:
     http://search.twitter.com/search?q=yslow

Questions and feedbacks welcome.

Thanks,
-betty

Reply | Messages in this Topic (5)

#1698

From: Sergey Chernyshev <sergey.chernyshev@...>
Date: Wed Oct 20, 2010 4:46 am
Subject: Re: YSlow 2.1.0 available - offers major bug fixes

sergeycherny...

Send Email

Great news. I've also added the new rule to ShowSlow and released 0.11:

http://www.showslow.com/blog/showslow-0-11-added-yslow-2-1-0-avoid-empty-image-src-rule/

Is there any documentation about the list of beaconing attributes?

Any other way for me to keep up to date with data you send?

Thank you,

� � � �� Sergey

On Tue, Oct 19, 2010 at 2:47 AM, bettyytso <bettyytso@...> wrote:

�

YSlow 2.1.0 is now available (courtesy of Marcel Duran). Detailed release notes please refer to YDN blog:
http://developer.yahoo.com/blogs/ydn/posts/2010/10/yslow-2-1-0-release-offers-major-bug-fixes/

Download it here:
https://addons.mozilla.org/en-US/firefox/addon/5369/

Twitter:
http://search.twitter.com/search?q=yslow

Questions and feedbacks welcome.

Thanks,
-betty

Reply | Messages in this Topic (5)

#1699

From: Marcel Duran <contact@...>
Date: Wed Oct 20, 2010 6:48 pm
Subject: Re: YSlow 2.1.0 available - offers major bug fixes

marcelduran

Send Email

Hi Sergey,

We're glad you like the new version and already integrated with ShowSlow, this is awesome.

You can find the list of beaconing attributes here:�http://developer.yahoo.com/yslow/help/beacons.html#version2

The beaconing function process all rules and send them all when using�beaconInfo =�grade, it means any new rule will be sent as well. In this case "yemptysrc" is the new rule on YSlow 2.1.0.

Since YSlow is extensible anyone can create new rules and/or rulesets and these can potentially be beaconed without prior knowledge.

Any new rule officially added to YSlow (on ruleset id = ydefault, parameter i on beacon) will be publicly announced on YDN and on some well know performance related lists, like the "Avoid empty src or href" beconed as "yemptysrc" which is documented at�http://developer.yahoo.com/performance/rules.html#emptysrc

Do you have any suggestion to help us keep up in sync?

I noticed however you've added this new rule to ShowSlow as "Avoid Empty Image src". Actually this rule checks for src or href on tags that fetch remote content, not only bound to images:

- <img src="">

- <script src="">

- <link rel="stylesheet" href="">

Since we're just sending the rule id like ycdn, yemptysrc, etc, it's up to the client to map the id to a name, specially for other languages, does it help if we send the official rule name in the beacon?

Best,

Marcel

On Tue, Oct 19, 2010 at 9:46 PM, Sergey Chernyshev <sergey.chernyshev@...> wrote:

�

Great news. I've also added the new rule to ShowSlow and released 0.11:
http://www.showslow.com/blog/showslow-0-11-added-yslow-2-1-0-avoid-empty-image-src-rule/

Is there any documentation about the list of beaconing attributes?
Any other way for me to keep up to date with data you send?

Thank you,

� � � �� Sergey

On Tue, Oct 19, 2010 at 2:47 AM, bettyytso <bettyytso@...> wrote:

�

YSlow 2.1.0 is now available (courtesy of Marcel Duran). Detailed release notes please refer to YDN blog:
http://developer.yahoo.com/blogs/ydn/posts/2010/10/yslow-2-1-0-release-offers-major-bug-fixes/

Download it here:
https://addons.mozilla.org/en-US/firefox/addon/5369/

Twitter:
http://search.twitter.com/search?q=yslow

Questions and feedbacks welcome.

Thanks,
-betty

--
Marcel Duran

Reply | Messages in this Topic (5)

#1700

From: Adrian Yee <adrian@...>
Date: Wed Oct 20, 2010 9:07 pm
Subject: Re: YSlow 2.1.0 available - offers major bug fixes

adrian@...

Send Email

On 10/20/10 11:48, Marcel Duran wrote:
> Since we're just sending the rule id like ycdn, yemptysrc, etc, it's up
> to the client to map the id to a name, specially for other languages,
> does it help if we send the official rule name in the beacon?

I guess it would be useful in some cases, but as you mention, for
multi-language set ups, it wouldn't be.  FYI, GTmetrix wouldn't use the
name passed in via the beacon.

Adrian

Reply | Messages in this Topic (5)

#1701

From: "paul_gifford" <paul_gifford@...>
Date: Wed Oct 20, 2010 9:29 pm
Subject: YSlow and JSLint - ignoring files

paul_gifford

Send Email

Hi,

   Is there a way to have JSLint ignore certain JavaScript files?  I have a page
that uses a lot of YUI 2 components and my JSLint error report is filled with
issues with YUI code, not mine.  I'd like to configure YSlow to ignore the YUI
library and just look at my code.

Thanks!

Paul

Reply | Messages in this Topic (2)

#1702

From: Marcel Duran <contact@...>
Date: Wed Oct 20, 2010 11:36 pm
Subject: Re: YSlow and JSLint - ignoring files

marcelduran

Send Email

It's not currently possible, but I'll make sure to add your suggestion to YSlow wish-list, maybe in next versions.

BTW we're all ears for suggestions to improve YSlow, feel free to share your thoughts.

Marcel�

On Wed, Oct 20, 2010 at 2:29 PM, paul_gifford <paul_gifford@...> wrote:

�

Hi,

Is there a way to have JSLint ignore certain JavaScript files? I have a page that uses a lot of YUI 2 components and my JSLint error report is filled with issues with YUI code, not mine. I'd like to configure YSlow to ignore the YUI library and just look at my code.

Thanks!

Paul

--
Marcel Duran

Reply | Messages in this Topic (2)

#1703

From: "abhiramabhat" <abhiramabhat@...>
Date: Thu Oct 21, 2010 12:48 pm
Subject: A command line interface to smush.it

abhiramabhat

Send Email

Hello,

I have written a command line tool which takes in a directory of images and
smushes all the images in that and it's sub directories, downloads the smushed
images and replaces them in their respective directories. Read further in my
blog post -
http://abhirama.wordpress.com/2010/10/20/a-command-line-tool-to-run-smush-it/

Reply | Messages in this Topic (1)

#1704

From: "jared_rosoff" <jrosoff@...>
Date: Thu Oct 21, 2010 4:41 pm
Subject: Measuring web performance using Google Analytics

jared_rosoff

Send Email

Performers,

I've been working on a way to track web performance using google analytics. I build a proof of concept javascript tag that uses custom variables in GA and allows us to segment metrics like views, bounce rate and tie on site by the performance experienced by real users.

Check out our first results here : http://blog.yottaa.com/2010/10/how-to-measure-page-load-time-with-google-analytics/

We're working on some next steps to measure performance across more than just landing pages and to integrate better data from web timings API.

Would love to hear feedback, tho.

Enjoy.

--j

Reply | Messages in this Topic (7)

#1705

From: Steve Souders <steve@...>
Date: Thu Oct 21, 2010 4:59 pm
Subject: Re: Measuring web performance using Google Analytics

steve_souders

Send Email

I love work like this. Hats off to you Jared. Although I'll point out something like this already exists in GA called TimeTracker:
    http://code.google.com/apis/analytics/docs/tracking/eventTrackerWrappers.html

I feel the "export to Excel" step is not workable - too much data is lost. I love moving forward, but in this situation I would suggest holding off on dumping load time measurements into GA. I'd advocate following these paths:
    - evangelize support for time measurements in GA. (I'm doing that and anyone else who uses GA can do that too through your GA contacts or contact me and I'll forward your emails.)
    - Use a JS library for measuring page load times (like Episodes or Jared's code). Then we need someone to create a open source project that's a data warehouse for latency measurements with associated reporting tools and charts. (It'd be great to either do this on App Engine or standard LAMP stack for easy adoption.)

-Steve


On 10/21/2010 9:41 AM, jared_rosoff wrote:

Performers,

I've been working on a way to track web performance using google analytics. I build a proof of concept javascript tag that uses custom variables in GA and allows us to segment metrics like views, bounce rate and tie on site by the performance experienced by real users.

Check out our first results here : http://blog.yottaa.com/2010/10/how-to-measure-page-load-time-with-google-analytics/

We're working on some next steps to measure performance across more than just landing pages and to integrate better data from web timings API.

Would love to hear feedback, tho.

Enjoy.

--j

Reply | Messages in this Topic (7)

#1706

From: "jared_rosoff" <jrosoff@...>
Date: Thu Oct 21, 2010 6:25 pm
Subject: Re: Measuring web performance using Google Analytics

jared_rosoff

Send Email

Thanks for the encouragement Steve!

I am aware of TimeTracker and have used it. It's a nice drop-in tool. But it
suffers from some problems that I see.

1) The biggest problem with Event Tracker is that it keeps performance data as a
silo of information separate from your web analytics data. By utilizing Custom
Variables instead of events, your performance data is integrated with your
analytics data and allows you to segment your analytics data by performance. For
example, using event tracking, it is not possible to see how performance impacts
bounce rate or time on site. If we (the web performers of the world) want to
influence investment in performance, we need to be able to show the business
impact of performance which requires displaying the business benefit of improved
performance (more ad impressions, higher conversion rate, increased engagement).
We should be able to slice the data along many axis (by page, by session, by
time of day, by geography, by browser). All of these are unavailable with event
tracker, but can be done when we use custom variables. So while the charting of
event tracking may be superior in GA, custom variables gives us much richer
ability to drill into data. This is the big promise of custom variables -- it
allows us to segment our business intelligence by the performance dimension and
justify continued investment.

2) Event Tracking adds an additional pixel tag to the page in addition to the
trackPageview call. This is not as important as the previous point, but it does
negatively impact performance which is ironic for a tool meant to help us
improve performance.

I agree with your assessment that dropping the data into excel is a huge
downside. GA does not deal with this data well and could go a long way to better
support this use case. I'm not sure what data you think we lose in this export,
tho. Can you elaborate?

Where I disagree with you is in the statement that we should hold off dropping
our data into google analytics. Based on the segmentation capabilities of GA and
it's ability to integrate this data with business intelligence data, this feels
like the right way to go. The approach of building a separate repository
eliminates our ability to do this segmentation and requires recreation of a high
scale data collection and reporting engine that already exists.

The alternative I'd like to suggest is that we _do_ collect this data into
google analytics (as custom segments), but build a separate application /
service that leverages the Google Analytics API to access the data and visualize
it in more meaningful ways. This would preserve the link to business
intelligence and ability to segment that has already been built into GA and
eliminate the need to drop into Excel.

As you may know, we have already integrated our product at Yottaa with google
analytics so you can see correlations between external measurement data and your
business intelligence data. Would you and the rest of the community be
interested in a free and open visualization tool from Yottaa that eliminated the
need for excel and provided better charting and reporting capability over this
data? I feel that this is a better step incrementally towards our goal rather
than putting the breaks on and tossing out a mature and widely adopted
technology (GA). (here i'm putting my product manager hat on and trying to get
some feedback on our product roadmap -- so please do chime in if you think this
is a tool you'd use)

Thoughts?
--j

(shameless plug: for those of you not familiar with Yottaa, we provide a free
web performance analysis and monitoring tool at http://www.yottaa.com)

--- In exceptional-performance@yahoogroups.com, Steve Souders <steve@...> wrote:
>
> I love work like this. Hats off to you Jared. Although I'll point out
> something like this already exists in GA called TimeTracker:
>
> http://code.google.com/apis/analytics/docs/tracking/eventTrackerWrappers.html
>
> I feel the "export to Excel" step is not workable - too much data is
> lost. I love moving forward, but in this situation I would suggest
> holding off on dumping load time measurements into GA. I'd advocate
> following these paths:
>      - evangelize support for time measurements in GA. (I'm doing that
> and anyone else who uses GA can do that too through your GA contacts or
> contact me <http://stevesouders.com/contact.php> and I'll forward your
> emails.)
>      - Use a JS library for measuring page load times (like Episodes
> <http://stevesouders.com/episodes/> or Jared's code). Then we need
> someone to create a open source project that's a data warehouse for
> latency measurements with associated reporting tools and charts. (It'd
> be great to either do this on App Engine or standard LAMP stack for easy
> adoption.)
>
> -Steve
>
>
> On 10/21/2010 9:41 AM, jared_rosoff wrote:
> > Performers,
> >
> >  I've been working on a way to track web performance using google
> > analytics. I build a proof of concept javascript tag that uses custom
> > variables in GA and allows us to segment metrics like views, bounce
> > rate and tie on site by the performance experienced by real users.
> >  Check out our first results here :
> >
http://blog.yottaa.com/2010/10/how-to-measure-page-load-time-with-google-analyti\
cs/
> >
> >
> >  We're working on some next steps to measure performance across more
> > than just landing pages and to integrate better data from web timings
> > API.
> >
> >  Would love to hear feedback, tho.
> >  Enjoy.
> >  --j
> >
>

Reply | Messages in this Topic (7)

#1707

From: Sergey Chernyshev <sergey.chernyshev@...>
Date: Fri Oct 22, 2010 2:11 am
Subject: Re: Re: Measuring web performance using Google Analytics

sergeycherny...

Send Email

I have one problem with GA - there is no way to insert metrics into it that are not related to specific page loads. I was thinking of inserting data about page quality (e.g. YSlow / PageSpeed scores), but it seems that there is no insertion API ;(

� � � �� Sergey

On Thu, Oct 21, 2010 at 2:25 PM, jared_rosoff <jrosoff@...> wrote:

�

Thanks for the encouragement Steve!

I am aware of TimeTracker and have used it. It's a nice drop-in tool. But it suffers from some problems that I see.

1) The biggest problem with Event Tracker is that it keeps performance data as a silo of information separate from your web analytics data. By utilizing Custom Variables instead of events, your performance data is integrated with your analytics data and allows you to segment your analytics data by performance. For example, using event tracking, it is not possible to see how performance impacts bounce rate or time on site. If we (the web performers of the world) want to influence investment in performance, we need to be able to show the business impact of performance which requires displaying the business benefit of improved performance (more ad impressions, higher conversion rate, increased engagement). We should be able to slice the data along many axis (by page, by session, by time of day, by geography, by browser). All of these are unavailable with event tracker, but can be done when we use custom variables. So while the charting of event tracking may be superior in GA, custom variables gives us much richer ability to drill into data. This is the big promise of custom variables -- it allows us to segment our business intelligence by the performance dimension and justify continued investment.

2) Event Tracking adds an additional pixel tag to the page in addition to the trackPageview call. This is not as important as the previous point, but it does negatively impact performance which is ironic for a tool meant to help us improve performance.

I agree with your assessment that dropping the data into excel is a huge downside. GA does not deal with this data well and could go a long way to better support this use case. I'm not sure what data you think we lose in this export, tho. Can you elaborate?

Where I disagree with you is in the statement that we should hold off dropping our data into google analytics. Based on the segmentation capabilities of GA and it's ability to integrate this data with business intelligence data, this feels like the right way to go. The approach of building a separate repository eliminates our ability to do this segmentation and requires recreation of a high scale data collection and reporting engine that already exists.

The alternative I'd like to suggest is that we _do_ collect this data into google analytics (as custom segments), but build a separate application / service that leverages the Google Analytics API to access the data and visualize it in more meaningful ways. This would preserve the link to business intelligence and ability to segment that has already been built into GA and eliminate the need to drop into Excel.

As you may know, we have already integrated our product at Yottaa with google analytics so you can see correlations between external measurement data and your business intelligence data. Would you and the rest of the community be interested in a free and open visualization tool from Yottaa that eliminated the need for excel and provided better charting and reporting capability over this data? I feel that this is a better step incrementally towards our goal rather than putting the breaks on and tossing out a mature and widely adopted technology (GA). (here i'm putting my product manager hat on and trying to get some feedback on our product roadmap -- so please do chime in if you think this is a tool you'd use)

Thoughts?
--j

(shameless plug: for those of you not familiar with Yottaa, we provide a free web performance analysis and monitoring tool at http://www.yottaa.com)

--- In exceptional-performance@yahoogroups.com, Steve Souders <steve@...> wrote:
>
> I love work like this. Hats off to you Jared. Although I'll point out
> something like this already exists in GA called TimeTracker:
>
> http://code.google.com/apis/analytics/docs/tracking/eventTrackerWrappers.html
>
> I feel the "export to Excel" step is not workable - too much data is
> lost. I love moving forward, but in this situation I would suggest
> holding off on dumping load time measurements into GA. I'd advocate
> following these paths:
> - evangelize support for time measurements in GA. (I'm doing that
> and anyone else who uses GA can do that too through your GA contacts or
> contact me <http://stevesouders.com/contact.php> and I'll forward your
> emails.)
> - Use a JS library for measuring page load times (like Episodes
> <http://stevesouders.com/episodes/> or Jared's code). Then we need
> someone to create a open source project that's a data warehouse for
> latency measurements with associated reporting tools and charts. (It'd
> be great to either do this on App Engine or standard LAMP stack for easy
> adoption.)
>
> -Steve
>
>
> On 10/21/2010 9:41 AM, jared_rosoff wrote:
> > Performers,
> >
> > I've been working on a way to track web performance using google
> > analytics. I build a proof of concept javascript tag that uses custom
> > variables in GA and allows us to segment metrics like views, bounce
> > rate and tie on site by the performance experienced by real users.
> > Check out our first results here :
> > http://blog.yottaa.com/2010/10/how-to-measure-page-load-time-with-google-analytics/
> >
> >
> > We're working on some next steps to measure performance across more
> > than just landing pages and to integrate better data from web timings
> > API.
> >
> > Would love to hear feedback, tho.
> > Enjoy.
> > --j
> >
>

Reply | Messages in this Topic (7)

#1708

From: Sergey Chernyshev <sergey.chernyshev@...>
Date: Fri Oct 22, 2010 2:19 am
Subject: Re: YSlow 2.1.0 available - offers major bug fixes

sergeycherny...

Send Email

Duran,

Actually if you check this page:�http://developer.yahoo.com/yslow/help/beacons.html#version2�- it doesn't mention yemptysrc ;)

Ideally, it would be nice to have a documentation page that always reflects the current state and I can subscribe to, e.g. wiki page, for example.

I used "Avoid empty src or href" as the name for the rule because that's exactly how your documentation describes it ;)�http://developer.yahoo.com/performance/rules.html#emptysrc

It's understandable that YSlow might send more rules if custom sets are defined, but it'd be harder to maintain in ShowSlow and is not supported yet. If there is interest in this functionality, I'll be happy to add custom rules to YSlow beacon configuration.

Sending meta-data about the beaconing software along with the data seems wasteful - just keep the documentation updated and that'd be good enough.

Thank you,

� � � �� Sergey

--
Sergey Chernyshev
http://www.sergeychernyshev.com/

http://www.meetup.com/Web-Performance-NY/

http://www.showslow.com/

On Wed, Oct 20, 2010 at 2:48 PM, Marcel Duran <contact@...> wrote:

�

Hi Sergey,

We're glad you like the new version and already integrated with ShowSlow, this is awesome.

You can find the list of beaconing attributes here:�http://developer.yahoo.com/yslow/help/beacons.html#version2

The beaconing function process all rules and send them all when using�beaconInfo =�grade, it means any new rule will be sent as well. In this case "yemptysrc" is the new rule on YSlow 2.1.0.

Since YSlow is extensible anyone can create new rules and/or rulesets and these can potentially be beaconed without prior knowledge.
Any new rule officially added to YSlow (on ruleset id = ydefault, parameter i on beacon) will be publicly announced on YDN and on some well know performance related lists, like the "Avoid empty src or href" beconed as "yemptysrc" which is documented at�http://developer.yahoo.com/performance/rules.html#emptysrc

Do you have any suggestion to help us keep up in sync?

I noticed however you've added this new rule to ShowSlow as "Avoid Empty Image src". Actually this rule checks for src or href on tags that fetch remote content, not only bound to images:

- <img src="">
- <script src="">
- <link rel="stylesheet" href="">

Since we're just sending the rule id like ycdn, yemptysrc, etc, it's up to the client to map the id to a name, specially for other languages, does it help if we send the official rule name in the beacon?

Best,

Marcel

On Tue, Oct 19, 2010 at 9:46 PM, Sergey Chernyshev <sergey.chernyshev@...> wrote:

�

Great news. I've also added the new rule to ShowSlow and released 0.11:
http://www.showslow.com/blog/showslow-0-11-added-yslow-2-1-0-avoid-empty-image-src-rule/

Is there any documentation about the list of beaconing attributes?
Any other way for me to keep up to date with data you send?

Thank you,

� � � �� Sergey

On Tue, Oct 19, 2010 at 2:47 AM, bettyytso <bettyytso@...> wrote:

�

YSlow 2.1.0 is now available (courtesy of Marcel Duran). Detailed release notes please refer to YDN blog:
http://developer.yahoo.com/blogs/ydn/posts/2010/10/yslow-2-1-0-release-offers-major-bug-fixes/

Download it here:
https://addons.mozilla.org/en-US/firefox/addon/5369/

Twitter:
http://search.twitter.com/search?q=yslow

Questions and feedbacks welcome.

Thanks,
-betty

--
Marcel Duran

Reply | Messages in this Topic (5)

#1710

From: Marcel Duran <contact@...>
Date: Fri Oct 22, 2010 9:20 am
Subject: Re: Measuring web performance using Google Analytics

marcelduran

Send Email

Good job Jared!

You could however automate the excel export process using GA Data Export API (http://code.google.com/intl/en/apis/analytics/docs/gdata/home.html) or perhaps you don't even need to export to excel, just retrieve the data you need using this API then build the charts using any charts API.

In order to track any page in a website not only the landing page, you could set the custom variable name using something like:

_gaq.push(["_setCustomVar",1,'pageTime_'�+ encodeURI(location.pathname), load_time,2]);

Best,

Marcel

On Thu, Oct 21, 2010 at 9:41 AM, jared_rosoff <jrosoff@...> wrote:

�

Performers,�

�I've been working on a way to track web performance using google analytics. I build a proof of concept javascript tag that uses custom variables in GA and allows us to segment metrics like views, bounce rate and tie on site by the performance experienced by real users.�

�
�Check out our first results here : http://blog.yottaa.com/2010/10/how-to-measure-page-load-time-with-google-analytics/�

�We're working on some next steps to measure performance across more than just landing pages and to integrate better data from web timings API.�

�Would love to hear feedback, tho.�

�
�Enjoy.�
�--j�
�

--
Marcel Duran

Reply | Messages in this Topic (7)

#1711

From: "stephen_thair" <stephen.thair@...>
Date: Fri Oct 22, 2010 12:45 pm
Subject: Re: Measuring web performance using Google Analytics

stephen_thair

Send Email

@Souders - perhaps its possible to build on top of Piwik for the reporting?

Piwik uses a standard approach of javascript tags
(http://piwik.org/docs/javascript-tracking/) which post the required data
collected from the client browser back to the Piwik web server application for
storage in a database.

The performance data that we'd want to collect is not part of the "standard"
Piwik dataset and hence would require custom development.But it's not like we
are short of options with Episodes or Boomerang for extracting the timing data.

The simplest way would encode performance data in JSON format into the custom
data attribute that Piwik provides i.e. setCustomData( object|string).

A custom Piwik "plugin"
(http://dev.piwik.org/trac/wiki/Plugins/HowToWritePlugin) would then decode that
performance data and store it into the Piwik database (custom tables would need
to be defined in order to store the data).

It is work noting that like most analytics packages Piwik operates a "two tier"
data storage model where data is periodically "archived" (what Site Confidence
would called "rolled up") into a summary, normalised data structure for longer
term storage as opposed to the simpler "raw format" which is just a flat table
on the data returned from the Piwik tag (but is faster to write to the
database).

The risk is, however, the potential size of the datasets required.
Using the Piwik database schema as a guide
(http://dev.piwik.org/trac/wiki/DatabaseSchema) the "statistics logging table"
schema indicates that it might take ~ 900 Bytes per page hit (allowing 256 chars
for the referrer URL).

For a medium to large site like Totaljobs.com this would be about 45M page
impressions/month, which would give 45,000,000 * 900 / 1000 (for KB) / 1000 (for
MB) / 1000 (GB) / 30 (days) = 1.35 GB / day for the "raw data".

The amount of storage required for the "archived" data would be dependent on a
number of factors, primarily the archiving period (hourly, daily etc), the
distribution of your site over time etc. Note that at present Piwik does not
remove any data from the "raw" data table when it "archives" � the "archived"
summary data is purely used to speed up reporting by pre-calculating values for
summary time periods.

Despite this, the biggest concern is performance under load for high volume
websites, both in terms of the performance to record the page impression but
more importantly the reporting performance.

One of the key factors that will influence performance will be how "real time"
customers require the data � the more "real time" the larger datasets that need
to be dynamically calculated (as opposed to pre-calculated by the "archiving"
process) and if you reduce the archiving process the more "pre-calculated data"
that needs to be stored.

Most major analytics providers only update the reporting data warehouse every
hour, and normally then only for a subset of the complete reporting database. In
the case of the free Google Analytics the data is normally 24 hours behind.
For "operational" use e.g. "if my website slowing down right now under load"
then the preference is for "real time"... for more "strategic" analytics an hour
or even a days delay might be acceptable.

It may be possible to partition the data in such a way that a subset of URL's
have "real time" data and the others only hourly, but that would need further
investigation.

-cheers

Steve
http://twitter.com/theopsmgr

Reply | Messages in this Topic (7)

#1712

From: Andy Davies <dajdavies@...>
Date: Fri Oct 22, 2010 3:03 pm
Subject: Re: Re: Measuring web performance using Google Analytics

dajdavies

Send Email

One advantage of going for a GA based approach is the breadth of use that GA has.

I'm sure I've seen a GA based approach the split the pages into fast / medium / slow somewhere (but of course can't find it now!)

Andy

On 22 October 2010 13:45, stephen_thair <stephen.thair@...> wrote:

@Souders - perhaps its possible to build on top of Piwik for the reporting?

Piwik uses a standard approach of javascript tags (http://piwik.org/docs/javascript-tracking/) which post the required data collected from the client browser back to the Piwik web server application for storage in a database.

The performance data that we'd want to collect is not part of the "standard" Piwik dataset and hence would require custom development.But it's not like we are short of options with Episodes or Boomerang for extracting the timing data.

The simplest way would encode performance data in JSON format into the custom data attribute that Piwik provides i.e. setCustomData( object|string).

A custom Piwik "plugin" (http://dev.piwik.org/trac/wiki/Plugins/HowToWritePlugin) would then decode that performance data and store it into the Piwik database (custom tables would need to be defined in order to store the data).

It is work noting that like most analytics packages Piwik operates a "two tier" data storage model where data is periodically "archived" (what Site Confidence would called "rolled up") into a summary, normalised data structure for longer term storage as opposed to the simpler "raw format" which is just a flat table on the data returned from the Piwik tag (but is faster to write to the database).

The risk is, however, the potential size of the datasets required.
Using the Piwik database schema as a guide (http://dev.piwik.org/trac/wiki/DatabaseSchema) the "statistics logging table" schema indicates that it might take ~ 900 Bytes per page hit (allowing 256 chars for the referrer URL).

For a medium to large site like Totaljobs.com this would be about 45M page impressions/month, which would give 45,000,000 * 900 / 1000 (for KB) / 1000 (for MB) / 1000 (GB) / 30 (days) = 1.35 GB / day for the "raw data".

The amount of storage required for the "archived" data would be dependent on a number of factors, primarily the archiving period (hourly, daily etc), the distribution of your site over time etc. Note that at present Piwik does not remove any data from the "raw" data table when it "archives" � the "archived" summary data is purely used to speed up reporting by pre-calculating values for summary time periods.

Despite this, the biggest concern is performance under load for high volume websites, both in terms of the performance to record the page impression but more importantly the reporting performance.

One of the key factors that will influence performance will be how "real time" customers require the data � the more "real time" the larger datasets that need to be dynamically calculated (as opposed to pre-calculated by the "archiving" process) and if you reduce the archiving process the more "pre-calculated data" that needs to be stored.

Most major analytics providers only update the reporting data warehouse every hour, and normally then only for a subset of the complete reporting database. In the case of the free Google Analytics the data is normally 24 hours behind.
For "operational" use e.g. "if my website slowing down right now under load" then the preference is for "real time"... for more "strategic" analytics an hour or even a days delay might be acceptable.

It may be possible to partition the data in such a way that a subset of URL's have "real time" data and the others only hourly, but that would need further investigation.

-cheers

Steve
http://twitter.com/theopsmgr

------------------------------------

--
http://developer.yahoo.com/performance/

Yahoo! Groups Links

<*> To visit your group on the web, go to:

� �http://groups.yahoo.com/group/exceptional-performance/

<*> Your email settings:
� �Individual Email | Traditional

<*> To change settings online go to:
� �http://groups.yahoo.com/group/exceptional-performance/join
� �(Yahoo! ID required)

<*> To change settings via email:
� �exceptional-performance-digest@yahoogroups.com
� �exceptional-performance-fullfeatured@yahoogroups.com

<*> To unsubscribe from this group, send an email to:
� �exceptional-performance-unsubscribe@yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
� �http://docs.yahoo.com/info/terms/

Reply | Messages in this Topic (7)

#1713

From: "stephen_thair" <stephen.thair@...>
Date: Fri Oct 22, 2010 10:35 pm
Subject: LDN Web Performance Group Created!

stephen_thair

Send Email

Hello all,

I have create a London Web Performance Group on meetup.com!

http://www.meetup.com/London-Web-Performance-Group/

I haven't had time to customise it or tweak it yet... but we already have 17
people signed up, so that's a good start!

I hope to arrange an event this side of Xmas, otherwise we'll kick off properly
in late January once everyone is back from holidays etc.

Spread the word!

cheers,

Steve
@TheOpsMgr

Reply | Messages in this Topic (1)

#1714

From: "JoeS" <shawjoej@...>
Date: Mon Oct 25, 2010 8:01 pm
Subject: Newbie Question: Can't Download Smushed Images

jshaw1953

Send Email

Hi,

I can upload images to http://www.smushit.com/ysmush.it/.

I can see the images getting smushed.

But, nothing happens when I click "download smushed images" on the Results page.

Am I missing something?

Advice requested.



Joe

Reply | Messages in this Topic (14)

#1715

From: Antonia Kwok <antoniakwok@...>
Date: Tue Oct 26, 2010 1:08 am
Subject: Re: Newbie Question: Can't Download Smushed Images

antoniakwok

Send Email

Could it be that all your images are already optimized? Does the result tab on the top mentioned any saving from smushing the images?

Sent from my iPhone

On Oct 25, 2010, at 1:01 PM, "JoeS" <shawjoej@...> wrote:

Hi,

I can upload images to http://www.smushit.com/ysmush.it/.

I can see the images getting smushed.

But, nothing happens when I click "download smushed images" on the Results page.

Am I missing something?

Advice requested.

Joe

Reply | Messages in this Topic (14)

#1716

From: "Joe" <shawjoej@...>
Date: Tue Oct 26, 2010 1:15 pm
Subject: RE: Can't Download Smushed Images

jshaw1953

Send Email

Hi,

Antonia asked about actual smushing. In fact, I can observe the smushing statistics, but the download button doesn’t work.

I get the problem with IE8 and with Firefox. I can’t figure out why the download button just “sits there.”

Joe

From: exceptional-performance@yahoogroups.com [mailto:exceptional-performance@yahoogroups.com] On Behalf Of Antonia Kwok
Sent: Monday, October 25, 2010 9:08 PM
To: exceptional-performance@yahoogroups.com
Cc: exceptional-performance@yahoogroups.com
Subject: Re: [exceptional-performance] Newbie Question: Can't Download Smushed Images

Could it be that all your images are already optimized? Does the result tab on the top mentioned any saving from smushing the images?

Sent from my iPhone

On Oct 25, 2010, at 1:01 PM, "JoeS" <shawjoej@...> wrote:

Hi,

I can upload images to http://www.smushit.com/ysmush.it/.

I can see the images getting smushed.

But, nothing happens when I click "download smushed images" on the Results page.

Am I missing something?

Advice requested.

Joe

Reply | Messages in this Topic (1)

#1717

From: "Jacqueline" <emailwinter@...>
Date: Wed Oct 27, 2010 2:06 am
Subject: Rule 2 // Yslow on Firebug/Firefox for Mac

jacquelinehu...

Send Email

In regards to the paragraph below, there is no way to add a CDN to firefox that
i know of. Please advise. Thank you.


How do I add my CDN hostname to YSlow for Rule 2?

Rule 2 says to use a content delivery network (CDN). The score for this rule is
computed by checking the hostname of each component against the list of known
CDNs. Unfortunately, the list of "known CDNs" are the ones used by Yahoo!. Most
likely these are not relevant to your web site, except for potentially
yui.yahooapis.com. If you want an accurate score for your web site, you can add
your CDN hostnames to YSlow using Firefox's preferences. Here are the steps to
follow:

    1. Go to about:config in Firefox. You'll see the current list of preferences.
    2. Right-click in the window and choose New and String to create a new string
preference.
    3. Enter extensions.yslow.cdnHostnames for the preference name.
    4. For the string value, enter the hostname of your CDN, for example,
mycdn.com.
    5. Restart Firefox.
       Do not use quotes. If you have multiple CDN hostnames, separate them with
commas.

If you specify CDN hostnames in your preferences, they'll be shown under the
details for Rule 2 in the Performance view.

Reply | Messages in this Topic (2)

#1718

From: Marcel Duran <contact@...>
Date: Wed Oct 27, 2010 5:23 am
Subject: Re: Rule 2 // Yslow on Firebug/Firefox for Mac

marcelduran

Send Email

Hi Jacqueline,

I didn't get your question.�Adding CDN hostnames to YSlow FF configs only makes rule 2 to consider these CDNs when running YSlow on your pages.

After following the steps to add new CDNs to your YSlow FF configuration you'll notice a message on rule 2 message panel (right):

"Using these CDN hostnames from your preferences: foobar.com"

I've just tested with the same configuration you mentioned and it works as expected in YSlow 2.1.0

Thanks,

Marcel

On Tue, Oct 26, 2010 at 7:06 PM, Jacqueline <emailwinter@...> wrote:

�

In regards to the paragraph below, there is no way to add a CDN to firefox that i know of. Please advise. Thank you.

How do I add my CDN hostname to YSlow for Rule 2?

Rule 2 says to use a content delivery network (CDN). The score for this rule is computed by checking the hostname of each component against the list of known CDNs. Unfortunately, the list of "known CDNs" are the ones used by Yahoo!. Most likely these are not relevant to your web site, except for potentially yui.yahooapis.com. If you want an accurate score for your web site, you can add your CDN hostnames to YSlow using Firefox's preferences. Here are the steps to follow:

1. Go to about:config in Firefox. You'll see the current list of preferences.
2. Right-click in the window and choose New and String to create a new string preference.
3. Enter extensions.yslow.cdnHostnames for the preference name.
4. For the string value, enter the hostname of your CDN, for example, mycdn.com.
5. Restart Firefox.
Do not use quotes. If you have multiple CDN hostnames, separate them with commas.

If you specify CDN hostnames in your preferences, they'll be shown under the details for Rule 2 in the Performance view.

--
Marcel Duran

Reply | Messages in this Topic (2)

#1719

From: "quentin_johansen" <tchai8@...>
Date: Thu Oct 28, 2010 3:29 pm
Subject: Strange... Compression enabled via .htaccess, but not visible in ySlow

quentin_joha...

Send Email

Hello,
I have enabled compression (and other components) with .htaccess.
If I check compression on [gidnetwork.com...] [seoconsultants.com...] and
[whatsmyip.org...] i get confirmation that components are being zippped, but if
I check ySlow, they are NOT... This includes the other components too like ET
tags and expires headers...
Strangely, I've also noticed that the images/media are loading SLOWER (by a
second or two). Please take a look at the below - any assistance to help resolve
these issues would be very much appreciated... Thanks/Andrew

<IfModule mod_rewrite.c>
RewriteEngine On
RewriteBase /
RewriteCond %{REQUEST_FILENAME} !-d
RewriteCond %{REQUEST_FILENAME} !-f
RewriteRule ^(.*)$ index.php?url=$1 [QSA,L]
</IfModule>


<ifModule mod_gzip.c>
mod_gzip_on Yes
mod_gzip_dechunk Yes
mod_gzip_item_include file \.(html?|txt|css|js|php|pl)$
mod_gzip_item_include handler ^cgi-script$
mod_gzip_item_include mime ^text/.*
mod_gzip_item_include mime ^application/x-javascript.*
mod_gzip_item_exclude mime ^image/.*
mod_gzip_item_exclude rspheader ^Content-Encoding:.*gzip.*
</ifModule>


# BEGIN COMPRESSION AND CACHING
<IfModule mod_deflate.c>
# Enable compression
AddOutputFilterByType DEFLATE text/css text/javascript application/x-javascript
text/html text/plain text/xml image/x-icon
<IfModule mod_setenvif.c>
BrowserMatch ^Mozilla/4 gzip-only-text/html
BrowserMatch ^Mozilla/4\.0[678] no-gzip
BrowserMatch \bMSIE !no-gzip !gzip-only-text/html
BrowserMatch \bMSI[E] !no-gzip !gzip-only-text/html
</IfModule>
<IfModule mod_headers.c>
# Make sure proxies deliver correct content
Header append Vary User-Agent env=!dont-vary
# Ensure proxies deliver compressed content correctly
Header append Vary Accept-Encoding
</IfModule>
</IfModule>
<IfModule mod_headers.c>
# No ETags, No Pragma
Header unset Pragma
Header unset ETag
# Default cache time to 1 year (31536000 sec)
Header set Cache-Control "max-age=31536000, public, must-revalidate"
</IfModule>
# No ETags
FileETag none
# CACHE SETTINGS (mod_expires)
<IfModule mod_expires.c>
# Turn on Expires
ExpiresActive On
# set default to "access plus 1 year"
ExpiresDefault A31536000
# html - "modification plus 1 hour"
ExpiresByType text/html M3600
# css and JavaScript - "modification plus 6 weeks"
ExpiresByType text/css M3628800
ExpiresByType text/javascript M3628800
ExpiresByType application/x-javascript M3628800
</IfModule>
# No cache for php-files
<FilesMatch "\.(php)$">
<IfModule mod_expires.c>
ExpiresActive Off
</IfModule>
<IfModule mod_headers.c>
Header set Cache-Control "private, no-cache, no-store, proxy-revalidate,
no-transform"
</IfModule>
</FilesMatch>

Reply | Messages in this Topic (9)

Yahoo! Groups

Breaking News Visit Yahoo! News for the latest.

The Yahoo! Groups Product Blog

Check it out!

Group Information

Yahoo! Groups Tips

Did you know...

Messages