Web 100 Double checking

[Aravind Narayanan]

Hi,

After double checking the Web 100 entries that are marked in red, please remove the red highlighting if all the following conditions are satisfied:

1) Delta YSlow <= 10

2) The time does not differ by more than 20 %

3) Delta Requests do not differ by more than 20%

4) The total size does not differ by more than 20%.

Thanks,

--

Aravind Narayanan | arav...@cs.stanford.edu

TA | CS193H

[Minh Nguyen]

What if the previous reviewer had used a different page? I was the
second reviewer for MapQuest; the first probably used the beta version
at http://www.mapquest.com/beta/ , but I used the original one at the
URL listed on the spreadsheet.

It also looks like someone had accidentally touched one of my reviews,
because my load times for Washington Mutual were shifted to the left,
leading to a YSlow delta in the thousands.

Will these issues affect my grade for the project?

> <mailto:arav...@cs.stanford.edu>
> TA | CS193H
--
Minh Nguyen <m...@zoomtown.com>
AIM: trycom2000; Jabber: m...@1ec5.org; Blog: http://notes.1ec5.org/

[Steve Souders]

The first grade for Web 100 is already done and won't change based on
these recent findings about discrepencies. This assignment's grade will
be based on accuracy - for cases where two rows have a wide difference,
Aravind or I will double-check and "ding" someone who put in erroneous data.

I thought you were going to ask, "What if I'm right and the other person
is wrong? How can I meet the conditions?" You can't. Just double-check
your work. If you're right and the other person is wrong, they'll get
dinged, not you.

And finally, there have been a few discussions about YSlow results
varying. My guess is there's a bug in YSlow w/ Firebug 1.2+. It seems
like the most accurate and consistent results occur when Firebug's Net
Panel is enabled.

Therefore, do your tests with Firebug's Net Panel enabled.

-Steve

[Aaron Sarnoff]

I've gone though my sites, and most of my numbers correlate, but there are a few discrepencies due to the fact that the site has hanged (more images, flash object added, etc).  Is there something in particular that I should do in these cases since the data was valid, but if it is checked again, it does not pass the litmus proposed.

[Gustav Rydstedt]

I have another problem.
www.4shared.com fluctuates some 200% in image size from load to load. It
makes the numbers very inconsistent.

> >> <mailto:arav...@cs.stanford.edu <mailto:arav...@cs.stanford.edu>>
> >> TA | CS193H
> >>
>
>
>
>
> >

[Aravind Narayanan]

If you have this problem, just leave a comment in the last column explaining why you think the numbers fluctuate so much. Please be sure that this is the case before leaving such a comment though :)

Thanks,

--

Aravind Narayanan | a.cyb...@gmail.com

[Vibhor Nanavati]

Just to make sure that I understand the assignment completely.

I double check the rows that are highlighted against my name and report new numbers.

Do I not have to wait for the other reviewer to finish double checking before unhighlighting? And in case the deltas are under the limits you mentioned, both rows should get unhighlighted, right?

--
VIBHOR NANAVATI <vib...@stanford.edu>

[Aravind Narayanan]

Yes, so when you update your values, if the other person has not updated the values yet, you need not unhighlight. The second person to update in each pair of results will have to unhighlight both rows it if the deltas are under the limits.

Does this make sense?

Thanks

--

Aravind Narayanan | a.cyb...@gmail.com

[Vibhor Nanavati]

Yes. Thanks for the clarification.

--
VIBHOR NANAVATI <vib...@stanford.edu>

[Gustav Rydstedt]

Ok. here is another thing that has popped up several times:
what should the total size column really be?
Should it be the total of JS+CSS+Images+HTML, or the total for the
entire site(including huge flash files)?

-G

[Aravind Narayanan]

It should just be the sum of those components, as calculated in YSlow.

Also the images column should contain the size of both CSS images and normal images.

Thanks

--

Aravind Narayanan | a.cyb...@gmail.com

[Vibhor Nanavati]

Isn't the total-size column set to be the sum of those previous columns? For me, I never cross verified that column with what YSlow reports.

--
VIBHOR NANAVATI <vib...@stanford.edu>

[Aravind Narayanan]

They are set to automatically sum the previous size columns. I'm assuming that they will (roughly) agree with YSlow :)

Thanks

--

Aravind Narayanan | a.cyb...@gmail.com

[Tse-Wen Tom Wang]

They might not agree at all with YSlow's stats though, because YSlow's
stats report compressed sizes, while we are supposed to record
uncompressed sizes of HTML, CSS, and JavaScript. If the site has small
images and compresses the aforementioned file types, then the result
could vary widely.

Tom

[Jeff Pickhardt]

What about Flash? I know tribalfusion.com uses Flash, and I included
that in the total size. But there wasn't a column for Flash...

[Blair Heuer]

I noticed when checking a few of my rows, that the other row's values
for sizes seemed close to the values on the Stats panel in YSlow,
which shows the total value of files of that type, using the
compressed value where applicable. I thought we had to go to the
Components panel and add up the values, using the uncompressed when
the file was compressed, for HTML, JS, and CSS because we want the
uncompressed total size.

Is my understanding of the right way to do this correct? This may be
the cause of a lot of discrepancies in file sizes. Something for
people to look out for when double checking.

Thanks,
Blair

On Nov 9, 12:46 pm, "Jeff Pickhardt" <pickha...@gmail.com> wrote:
> What about Flash?  I know tribalfusion.com uses Flash, and I included
> that in the total size.  But there wasn't a column for Flash...
>
> On Sun, Nov 9, 2008 at 9:50 AM, Tse-Wen Tom Wang <tse...@gmail.com> wrote:
>
>
>
> > They might not agree at all with YSlow's stats though, because YSlow's
> > stats report compressed sizes, while we are supposed to record
> > uncompressed sizes of HTML, CSS, and JavaScript. If the site has small
> > images and compresses the aforementioned file types, then the result
> > could vary widely.
>
> > Tom
>

> > On Sun, Nov 9, 2008 at 9:30 AM, Aravind Narayanan <a.cyber...@gmail.com> wrote:
> >> They are set to automatically sum the previous size columns. I'm assuming
> >> that they will (roughly) agree with YSlow :)
> >> Thanks
> >> --

> >> Aravind Narayanan | a.cyber...@gmail.com

> >> On Sun, Nov 9, 2008 at 09:29, Vibhor Nanavati <nvib...@gmail.com> wrote:
>
> >>> Isn't the total-size column set to be the sum of those previous columns?
> >>> For me, I never cross verified that column with what YSlow reports.
>

> >>> On Sun, Nov 9, 2008 at 9:25 AM, Aravind Narayanan <a.cyber...@gmail.com>

> >>> wrote:
>
> >>>> It should just be the sum of those components, as calculated in YSlow.
> >>>> Also the images column should contain the size of both CSS images and
> >>>> normal images.
> >>>> Thanks
> >>>> --

> >>>> Aravind Narayanan | a.cyber...@gmail.com
> >>>> On Sun, Nov 9, 2008 at 09:26, Gustav Rydstedt <gustav.rydst...@gmail.com>

[Steve Souders]

Yes, it was supposed to be UNcompressed. Please add a note when you suspect this is the cause of discrepencies.

-Steve

[Rupesh Mishra]

Steve mentioned in lecture 15 slide 5, that % Delta requests difference should be <10%.

 I see that value as 20% in the email sent by Aravind. 

1) Delta YSlow <= 10

2) The time does not differ by more than 20 %

3) Delta Requests do not differ by more than 20%

4) The total size does not differ by more than 20%.

Which one should we take.

Thanks,
Rupesh
Robert Orben  - "Never raise your hand to your children - it leaves your midsection unprotected."

2008/11/7 Aravind Narayanan <a.cyb...@gmail.com>

[Steve Souders]

20% is correct - my bad