measure time on internet ..??

in company I work for, management would like to implemet sysmem that
would measure time spend on interner per employee .. in company we have
intranet and CRM which are web applications so only time spend visiting
"outside" web sites should be measured ..
I believe that MS ISA Server can do that but we are looking for another
solution ..

Any Idea ..
THX in advance

Mr. X <(E-Mail Removed)> wrote:
> in company I work for, management would like to implemet sysmem that
> would measure time spend on interner per employee .. in company we have
> intranet and CRM which are web applications so only time spend visiting
> "outside" web sites should be measured ..

It's very hard to do this well.

Consider: you grab a page from a site. You spend one hour reading it
(there's a lot of text) and then click on a link it provides to a
different site. You start reading that new page.

Now consider: you grab a page from a site. You spend 5 seconds glancing
at it and then get back to your "real" work. An hour later you go to
another site and spend another 5 seconds skimming the resulting page.

Finally, consider: you download a page from a site. You spend 5 seconds
glancing at it and then get back to your "real" work. An hour later you go
to another site and spend another 5 seconds skimming the resulting page.
What you haven't realised is that while the first page was open, it
was using a javascript applet to update information on the page every
few seconds.

In each case, how long have you spent "on the Internet", and how does
a machine determine the difference?

Chris

In article <(E-Mail Removed)>,
Chris Davies <chris-(E-Mail Removed)> writes:
> Mr. X <(E-Mail Removed)> wrote:
>> in company I work for, management would like to implemet sysmem that
>> would measure time spend on interner per employee .. in company we have
>> intranet and CRM which are web applications so only time spend visiting
>> "outside" web sites should be measured ..
>
> It's very hard to do this well.

Exactly. You explain the practical reasons, though I like to state it a
little more facetiously when asked how much time I spend on the
Internet. I don't spend any time on the Internet, it's my computer which
does, and it's connected 24 hours a day.

--
Tim Clark

Tim Clark wrote:
> In article <(E-Mail Removed)>,
> Chris Davies <chris-(E-Mail Removed)> writes:
>> Mr. X <(E-Mail Removed)> wrote:
>>> in company I work for, management would like to implemet sysmem that
>>> would measure time spend on interner per employee .. in company we have
>>> intranet and CRM which are web applications so only time spend visiting
>>> "outside" web sites should be measured ..
>> It's very hard to do this well.
>
> Exactly. You explain the practical reasons, though I like to state it a
> little more facetiously when asked how much time I spend on the
> Internet. I don't spend any time on the Internet, it's my computer which
> does, and it's connected 24 hours a day.
>
You are both right .. but management is stupid and thay dont take "no"
as an answer .. so YES .. they want to see how long time people spend
visiting external sites ... no matter if they actually surfing or they
only have their browsers open ..

Mr. X wrote:

> in company I work for, management would like to implemet sysmem that
> would measure time spend on interner per employee .. in company we have
> intranet and CRM which are web applications so only time spend visiting
> "outside" web sites should be measured ..
> I believe that MS ISA Server can do that but we are looking for another
> solution ..
>
It is not possible the way your bosses want to. Accessing a website is
a "one-shot" get-request, followed by some time reading it, then clicking
another link issuing another get - from the same page or another.
One can have a webpage open the whole day, in background, for the occasional
look, and still do work most of the time.
What would be possible is to count outside page hits, or transferred volume
per user, from the proxy. But then, there are pages doing auto-refresh
scripts every other second, even when the page just sits in the background.
--
vista policy violation: Microsoft optical mouse found penguin patterns
on mousepad. Partition scan in progress to remove offending
incompatible products. Reactivate MS software.
Linux 2.6.22.9-1mdvcustom [LinuxCounter#295241,ICQ#4918962]

Mr. X <(E-Mail Removed)> wrote:
> You are both right .. but management is stupid and thay dont take "no"
> as an answer .. so YES .. they want to see how long time people spend
> visiting external sites ... no matter if they actually surfing or they
> only have their browsers open ..

The only way to achieve that requirement is to install an applet on
every workstation that measures the time for which the user's browser
is running.

I think you need to have the requirements defined more precisely, or else
be prepared to say that you can measure something different...

Once example might be to measure the total time a user is downloading web
pages (and other items via HTTP) from the first item to the last, such
that the timer is stopped whenever the interval between items exceeds,
say, 10 minutes.

Example:

Time Object Cumulative duration
---------- ------------------------------ ----------------------
10:00:00 http://blah.../blah.html (timer starts)
10:00:01 http://blah.../image1.jpg 00:00:01
10:00:02 http://blah.../image2.jpg 00:00:02
10:01:40 http://yadda.../yadda.html 00:01:40

No further activity (for now) so the timer stops

10:38:01 http://another.../ (timer continues)
10:38:02 http://another.../image.png 00:01:41

Total duration is 00:01:41, in two periods

Assuming you've got some way of tieing log entries to specific users,
I should imagine.a perl script could parse a squid cache log and provide
this sort of report fairly easily.

Chris

On Dec 8, 6:05 am, "Mr. X" <ib1628_ma...@yahoo.com> wrote:

> You are both right .. but management is stupid and thay dont take "no"
> as an answer .. so YES .. they want to see how long time people spend
> visiting external sites ... no matter if they actually surfing or they
> only have their browsers open ..

Suppose I save a PDF file that I downloaded, close my browser, and
then read it for the next hour. This is a very common usage pattern
for me.

This is a question that is typical of a particular category of
questions. Other examples:

How can I tell how much free memory I have on my Linux machine?

How can I my program tell if the computer it's running on has an
Internet connection?

The answer is always the same -- define very precisely what it is you
want to measure, so that questions like "does X count" have a clear
answer. Then you can at least ask the right question, which is the
first step to getting the right answer.

DS

David Schwartz wrote:
> On Dec 8, 6:05 am, "Mr. X" <ib1628_ma...@yahoo.com> wrote:
>
>> You are both right .. but management is stupid and thay dont take "no"
>> as an answer .. so YES .. they want to see how long time people spend
>> visiting external sites ... no matter if they actually surfing or they
>> only have their browsers open ..

My browser is running 24x7 whether I'm using it or not. So measuring
browser run time wouldn't be an accurate gauge. But what might be
reasonably accurate would be to measure how long the browser window
has focus. That leaves out documents that invoke an external reader.
And of course sneaky people that found out the scheme could defocus
the browser while reading. More sophisticated log analysis could spot
that reasonably well though.