[analog-help] Statistics for framed pages

Aengus analog07 at eircom.net
Mon Sep 17 18:45:58 PDT 2007


On Monday, September 17, 2007 6:03 PM [EDT],
Nerissa Murphy <NMurphy at moreland.vic.gov.au> wrote:

> Hi there
>
> We currently have a framed website. I'd like to confirm if the number
> of
> accesses to framed pages in the statistics is in addition to the
> nested
> pages. The animals.htm example below would seem to indicate it is in
> addition to the framed page. I understand it's not so clear cut, but
> any
> information or explanatory note would be appreciated.

Analog isn't aware of any particular relationship between any two objects on 
your website - it doesn't know anything at all about your website, it's only 
concerned about what is recorded in your logfiles. That's an important 
distinction, because if, for example, you have pages on your site that have 
never been visited, then they're not listed in your logfile, and Analog will 
never include them in any reports. In the same way, Analog has no way of 
knowing that a particluar page contains a FRAME statement, and wouldn't know 
what to do with the information anyway.

> Examples
>
>   708:       : /services/wasteandrecycling-fr.htm
> <http://www.moreland.vic.gov.au/services/wasteandrecycling-fr.htm>
> (framed page)
>   736:   /services/wasteandrecycling.htm
> <http://www.moreland.vic.gov.au/services/wasteandrecycling.htm>
> (nested page)

The normal sequence of events when a user requests a page that has a FRAME 
statement is the browser requests the main (framing) page, and then requests 
the page or pages that are listed in the FRAME statements. So there wil be 
at least 2 requests recorded in your servers log files (in your case, you 
actually have 3 FRAME statements, the top and left frame, as well as the 
content frame, so there should be 4 entries in the log whenever a browser 
requests the framing page). Without aything more to go on, it would be 
reasonable to interpret these two lines as 708 requests for the framing 
page, which would usually lead to 708 requests for the framed page, and an 
additional 28 requests for that page directly - every request for the 
framing page should cause a request for the framed page, but it is also 
possible to call the page directly, without it's frame.

>   537:       : /services/animals-fr.htm
> <http://www.moreland.vic.gov.au/services/animals-fr.htm>  (framed
> page)
>   333:  : /services/animals.htm
> <http://www.moreland.vic.gov.au/services/animals.htm>  (nested page)

This example doesn't fit into that explanation, because you have more 
requests for the framing page than for the framed page - that shouldn't 
normally happen. The only obvious explanation is that the framing page was 
changed at some point, and that it used to frame a different page, so that 
the framing page has been called for a longer period of time than the 
currently listed framed page.

There may be additional information in your logfiles that can provide 
greater insight. If you are logging the referrer field, you can run a report 
than only include the framing page, to see how many of the requests for the 
framed page were generated when the browser loaded the framing page first.

REFINCLUDE http://www.moreland.vic.gov.au/services/wasteandrecycling-fr.htm
should generate a Request Report that shows 708 requests for 
/services/wasteandrecycling.htm (and for the top and left frames, though 
they may already be cached in the browser, and may not be requested again).

REFINCLUDE http://www.moreland.vic.gov.au/services/animals-fr.htm
might give some indication of why that framing page was requested more often 
than the page that it is framing.

Aengus 



More information about the analog-help mailing list