Forum Moderators: open

Message Too Old, No Replies

Link Prefetch in Mozilla Browsers

What exactly is prefetched: (1) HTML or (2) HTML and all include files?

         

darraghsmyth

8:46 am on May 12, 2005 (gmt 0)

10+ Year Member



I'm referring to the Link Prefetch facility in Mozilla-based browsers (see [mozilla.org...]

If a link is instrumented on Google to be pre-fetched by a Mozilla browser (see [google.com...] which of the following will happen:

1) The browser will pre-fetch the HTML document only.

2) The browser will pre-fetch the HTML document and all it's components (images and include files for javascript, CSS etc).

3) The browser will pre-fetch the HTML and all components and then pre-run any client-side dynamic components such as onLoad functions in JavaScript.

thanks a million for your answers
darragh

tedster

4:02 am on May 13, 2005 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Welcome to the forums, darragh. That's a great question, and one that seems almost impossible to pin down, short of testing. The "detailed" information just doesn't give enough detail!

Since pre-fetching does not involve rendering but only caching, I've been assuming that html and related files, such as images, are being grabbed. If the idea is to speed up the browser, the biggest chunk of time goes to images so I just assumed they being cached. Likewise I assumed that executables are being cached but not executed unless and until the page is actually requested by the user.

But I'm just guessing -- and I've put in a decent bit of searching without finding a definitve answer. I should probably set up a test with some <link> elements and a web server - but I'm hoing that one of our members can tell us (I do get a bit lazy sometimes.)

larryhatch

10:36 am on May 13, 2005 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Darragh & tedster:

I finally found something that looks like a Google prefetch, but it goes by a different name:

From my access file: The first GET is my index.html file:

195.92.67.65 - - [12/May/2005:09:15:30 -0400] "GET / HTTP/1.1" 200 6130 "http://www.google.co.uk/search?q=ufo&hl=en&lr=&start=10&sa=N" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Preload_01_07)"

Note the Preload indicator.

Then whoever or whatever "preloaded" the following files:

195.92.67.65 - - [12/May/2005:09:15:32 -0400] "GET /copyscape.gif HTTP/1.1" 200 756 "mysite.net/" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Preload_01_07)"
195.92.67.65 - - [12/May/2005:09:15:32 -0400] "GET /GLBPAC.gif HTTP/1.1" 200 5818 "mysite.net/" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Preload_01_07)"
195.92.67.65 - - [12/May/2005:09:15:49 -0400] "GET /REGMAPS.html HTTP/1.1" 200 7301 "mysite.net/" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Preload_01_07)"

Following REGMAPS.html, and abandoning index.html, prefetch gets this image:
195.92.67.65 - - [12/May/2005:09:15:50 -0400] "GET /ITALYN.gif HTTP/1.1" 200 7857 "mysite.net/REGMAPS.html" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Preload_01_07)"

Then it jumps to BRITALL, which is a link inside REGMAPS:
195.92.67.65 - - [12/May/2005:09:16:06 -0400]
"GET /BRITALL.html HTTP/1.1" 200 3216 "mysite.net/REGMAPS.html" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Preload_01_07)"

.. and takes the corresponding .gif from BRITALL:
195.92.67.65 - - [12/May/2005:09:16:07 -0400] "GET /BRITALLM.gif HTTP/1.1" 200 18802 "mysite.net/BRITALL.html" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Preload_01_07)"

- - - -

Does this tell you anything? How about the DNS #? Note this is google.uk, not the US version.

- Larry

darraghsmyth

10:49 am on May 13, 2005 (gmt 0)

10+ Year Member



Thanks very much Larry. Looks like it does download both HTML and all include files and images, which makes sense. Hopefully it's not also running any client-side JavaScript which is my main concern.

Our client-side javascript logs hits to a server upon page download. However we obviously don't want this to happen if the page is only being downloaded for a cache. Presumably when the page is actually displayed (if the link is clicked upon), then any client-side dynamic code is run.

cheers
darragh

larryhatch

11:07 am on May 13, 2005 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Hi Darragh:

I may have gone off half-cocked here.

I Googled up Preload_01_07, and found loads of references going back months or years.
The sites were mostly Guestbooks which repeated back the entire referral string.
Most look very much like the ones I quoted.

It may be that preload and prefetch are two entirely different things. Sorry for any false alarm.

The Preloads I showed loaded most but not all images on my index page.
Out of several sub-menus, it only selected one of them, then it zeroed in on one of many pages
from that sub-menu.

Whatever preload is, it follows a path much like a natural browsing human, rather than a machine.

- Larry