Timer

[CountScubula]

Hi,

I need a way to time visitors that come and log in to my site. If the timer
could keep track of the amounts of time that logged in visitors spend, then
that would be great.

Any ideas? Any freeware out there that could do this?

Jason

Ok, here is one simple aproach.

Create an <IMG SRC= "script.php?user=some_id" NAME="imgTimer"> tag on the
page that points to a php script.

then, have a piece of java reload the image, every 10 to 15 minutes.

now that script that it is calling, tracks the referer page, current time,
and updates a count of how long you are on that page. if the time between
the last time you were on the page and the current time is too great say, 2
times what the update interval is, then you must be new or back to that
page. if you are to stay on the page/site, the script should notice this.

if they close the browser without logging off, doesnt matter.

I have done this quite successfully in the past, however, depends if they
have java script enabled.


Another way I do it these days, and is how some banks do it, is everytime
you click a page, they update a counder of the time you are online, and if
you are idle more than x amount of minutes, they assume you left, or closed
the browser, and they session is usualy expired.

One site I did for a client a while back, used this scheme, so users knew
how long someone was idle.

 

[Therion Ware]

On Thu, 22 Jan 2004 20:52:53 GMT in free.christians, Dr. Jason
Gastrich ("Dr. Jason Gastrich" <[email protected]>) said, directing the
reply to free.christians

Hi,

I need a way to time visitors that come and log in to my site. If the timer
could keep track of the amounts of time that logged in visitors spend, then
that would be great.

Any ideas? Any freeware out there that could do this?

A *really* outre approach would be to run a uniquely named and looping
32x32, very low bandwidth, < 500bps, bit of streaming wmv video
embedded in each page and rely on the streaming server for the logs to
tell you what's what.

The advantages are that you actually have a live connection while the
user's on your site - as opposed to making educated guessed based on
what was last downloaded from your http server and the next http
destination, and of course you'd bypass many kinds of http cache and
get real IP info.

How you'd do this depends a bit on your server set up. Are you simply
buying space on other people's servers and re-branding, or do you run
your own? If the latter, then it's easy, if the former, then it may
work out as an expensive way to do things.

Mind you, that said, I haven't given this a go, but if all else fails,
it might be worth a try. Just pray you don't get many opera etc
users...!

Oh, want to send me another email, or several, which would help with
the problem diagnosis?


--
"Do Unto Others As You Would Have Them Do Unto You."
- Attrib: Pauline Reage.
Inexpensive VHS & other video to CD/DVD conversion?
See: <http://www.Video2CD.com>. 35.00 gets your video on DVD.
all posts to this email address are automatically deleted without being read.
** atheist poster child #1 ** #442.

 

[Nicolai P. Zwar]

They guess.

Then how come they are accurate every single time so far? In fact, they
even know whether I accurately logged off or whether the system just
kicked me out after a certain period of inactivity. So far, they have
not made a wrong guess. Good guessers, really.

They assume that if you do nothing for X minutes then you have shut
the browser, or switched off the computer, or simple forgotten about
it. At that point they decide to log you out automatically and record
that time.

Indeed, that's what they do.

For a banking site logging users out after five minutes of inactivity
is acceptable. For an distance learning course or a network monitoring
service such an action would not be acceptable.

There is no need to set the time to five minutes. You can set the
automatic log off time to any time you want. Maybe fifteen minutes, or
thirty, an hour... whatever seems suitable.

Recording the time someone logs in is easy. If they actually log out
then that is also easy to record.
But if they don't log out, if they
leave the site open, or leave the site without logging out, then any
time is a pure guess.

Well, yes, sure, but that's what I said, so it's obviously not flat out
"impossible". In fact, I specifically stated that it "depends on how you
log in and out". Obviously, this is not an HTML issue, but you can time
your visitors. How accurately, that depends on how they log in and out.
It depends on how co-operative the visitors actually are in the log in
and log out process (personally, I'm quite grateful that I can check
when and for how long I was logged into my bank account). I'm assuming
here these visitors want to be logged in and logged out in the first
place. Obviously, if they don't care to log in to begin with, there's
nothing you can do (except denying access to content).

 

[Steve Pugh]

Then how come they are accurate every single time so far?

Depends on your definition of accurate I suppose. If you walk away
from the computer and ten minutes later the server logs you out then
it records a time that's ten minutes after you stopped using the site.

Or maybe that time is actually two hours before you stop using the
site if, in the above scenario, you then come back and continue
reading the page that is still on screen, even printing it out for
your records.

There is no need to set the time to five minutes.

That was just an example.

Well, yes, sure, but that's what I said, so it's obviously not flat out
"impossible".

Yes it is. The server has no way of knowing when the user last looked
at the page in his browser, only when the page was last requested from
the server. If the server chooses to present some other time as the
time the user left the site than that time is a guess.

Steve

 

[Matt Silberstein]

In free.christians I read this message from "Nicolai P. Zwar"

If that were so, how do you explain the fact that my bank or my
cell-phone provider have log stats as to my log-ins and log-outs?

How come my bank or my cell-phone provider know when and for how long
I've been logged into my account if it is not possible?

I have to help our honorary doctor, but *if* you log in and log
out, if you don't turn away from the page to do other things for
an hour or so, if you are not looking at a cached version of the
page, then you can get some kind of vaguely useful statistics.

BTW, cell phones do not use http for call. They may put http on
top of the call, but that is different. They also are not, AFAIK,
multi-processes systems (you can't have many pages up at once).
And all they really care about is how long the *call* is. That
they can now just like the phone company (or you ISP) can tell
how long you were on if you use dial-up.

 

[Matt Silberstein]

In free.christians I read this message from Therion Ware

On Thu, 22 Jan 2004 20:52:53 GMT in free.christians, Dr. Jason
Gastrich ("Dr. Jason Gastrich" <[email protected]>) said, directing the
reply to free.christians




A *really* outre approach would be to run a uniquely named and looping
32x32, very low bandwidth, < 500bps, bit of streaming wmv video
embedded in each page and rely on the streaming server for the logs to
tell you what's what.

The advantages are that you actually have a live connection while the
user's on your site - as opposed to making educated guessed based on
what was last downloaded from your http server and the next http
destination, and of course you'd bypass many kinds of http cache and
get real IP info.

So you force the use to avoid the efficiency of a cache, force
them to use *their* bandwidth, and still don't know if they have
walked way for an hour or so.

[snip]

 

[Matt Silberstein]

In free.christians I read this message from "CountScubula"

Ok, here is one simple aproach.

Create an <IMG SRC= "script.php?user=some_id" NAME="imgTimer"> tag on the
page that points to a php script.

then, have a piece of java reload the image, every 10 to 15 minutes.

now that script that it is calling, tracks the referer page, current time,
and updates a count of how long you are on that page. if the time between
the last time you were on the page and the current time is too great say, 2
times what the update interval is, then you must be new or back to that
page. if you are to stay on the page/site, the script should notice this.

if they close the browser without logging off, doesnt matter.

So we are within 20-30 minutes or so. Works real well if the
expected time on the site is 2-5 minutes.

I have done this quite successfully in the past, however, depends if they
have java script enabled.


Another way I do it these days, and is how some banks do it, is everytime
you click a page, they update a counder of the time you are online, and if
you are idle more than x amount of minutes, they assume you left, or closed
the browser, and they session is usualy expired.

They have log-in and log-out. Do you think that people will find
sufficient value from a page that calls Galileo and Cortez
founders of the U.S. to put up with that.

 

[Nicolai P. Zwar]

Steve Pugh wrote:

Yes it is. The server has no way of knowing when the user last looked
at the page in his browser, only when the page was last requested from
the server. If the server chooses to present some other time as the
time the user left the site than that time is a guess.

Oh come on, Steve, that's an obvious given. The points you rase are
valid limitations. Obviously the plattform cannot tell how long you
actually physically _looked_ at the page any more than it can log the
color of your eyes, whether you were eating potato chips, or whether you
were listening to music in the background. Just as a time clock can only
record when you checked into your workplace and not for how long you
actually worked, or whether you actually worked at all, and how much
coffee you drank while you were at it. I'm sure that Dr. Gastrich is
aware of that fact (he said himself that "There is no way to tell if
they are watching TV or actually studying the material"), as is probably
just about anybody else who posts in this forum. Of course, you could
also just log in at the site in the background and surf to porn sites in
another browser window, hitting refresh once in a while to stay logged
in and fool the server set time out. The question wasn't about keeping
an Orwellian window into the user's workspace, but to record the time a
users spends logged in at the site. That can be done, within obvious and
reasonable limitations, some of which you mentioned. It's a far cry from
"impossible".

 

[Dr. Jason Gastrich]

On Thu, 22 Jan 2004 20:52:53 GMT in free.christians, Dr. Jason
Gastrich ("Dr. Jason Gastrich" <[email protected]>) said, directing the
reply to free.christians




A *really* outre approach would be to run a uniquely named and looping
32x32, very low bandwidth, < 500bps, bit of streaming wmv video
embedded in each page and rely on the streaming server for the logs to
tell you what's what.

The advantages are that you actually have a live connection while the
user's on your site - as opposed to making educated guessed based on
what was last downloaded from your http server and the next http
destination, and of course you'd bypass many kinds of http cache and
get real IP info.

How you'd do this depends a bit on your server set up. Are you simply
buying space on other people's servers and re-branding, or do you run
your own? If the latter, then it's easy, if the former, then it may
work out as an expensive way to do things.

Mind you, that said, I haven't given this a go, but if all else fails,
it might be worth a try. Just pray you don't get many opera etc
users...!

Oh, want to send me another email, or several, which would help with
the problem diagnosis?

Hi Therion,

Is your email accepting mail, now? I'm not sure if I have your email
address any more.

This method may help a person keep the time on their computer, but I don't
see how it would allow the site owner to know how long a person was on their
site.

JG

 

[brucie]

in post: <

[timing visitors]
It's a far cry from "impossible".

IT IS IMPOSSIBLE. the best you can do is guess and they're not even
going to be good guesses. FFS SPSM

 

[Therion Ware]

On Fri, 23 Jan 2004 20:37:12 GMT in free.christians, Dr. Jason
Gastrich ("Dr. Jason Gastrich" <[email protected]>) said, directing the
reply to free.christians

Hi Therion,

Is your email accepting mail, now? I'm not sure if I have your email
address any more.

This method may help a person keep the time on their computer, but I don't
see how it would allow the site owner to know how long a person was on their
site.

Because by checking the streaming server logs you could see how long
the stream continued for. Tot that figure up for each page they visit,
and you have the total time spent on your site. "Maybe".
--
"Do Unto Others As You Would Have Them Do Unto You."
- Attrib: Pauline Reage.
Inexpensive VHS & other video to CD/DVD conversion?
See: <http://www.Video2CD.com>. 35.00 gets your video on DVD.
all posts to this email address are automatically deleted without being read.
** atheist poster child #1 ** #442.

 

[Nicolai P. Zwar]

in post: <

[timing visitors]
It's a far cry from "impossible".

IT IS IMPOSSIBLE.

Not so.

the best you can do is guess and they're not even
going to be good guesses. FFS SPSM

Recording log in and log out times are far from being wild guesses.

 

[A Hess]

Somewhere around 1/23/04 4:26 PM, Nicolai P. Zwar typed wildly with
reckless abandon:

Recording log in and log out times are far from being wild guesses.

not if the person doesn't log out. I can't recall the last time I
actually logged out of anything. I just let it time out and I normally
leave a site long before that happens.

 

[John Burton]

Hi,

I need a way to time visitors that come and log in to my site. If the timer
could keep track of the amounts of time that logged in visitors spend, then
that would be great.

Any ideas? Any freeware out there that could do this?

Jason

http://www.sitemeter.com

I just added this to my site. As others have noted, the accuracy of the
stats you get should be considered questionable or possibly even worthless,
but you might want to try it and see what you think.

John

 

[Dr. Jason Gastrich]

http://www.sitemeter.com

I just added this to my site. As others have noted, the accuracy of
the stats you get should be considered questionable or possibly even
worthless, but you might want to try it and see what you think.

John

Thanks for your message and suggestion, John.

I've seen and used Sitemeter. Unfortunately, it would still only be an IP
tracking/timing system. We already have that function in CPanel. I'm
looking for something a little different.

JG

 

[Owen Jacobson]

Recording log in and log out times are far from being wild guesses.

Yup. You'll know exactly when the user hit the login page and when the
user hit the logout page. You still don't know anything for the
intervening time -- your user could've been off at some other website, or
had his browser closed, or not even near the computer.

 

[Dr. Jason Gastrich]

in post: <

[timing visitors]
It's a far cry from "impossible".

IT IS IMPOSSIBLE.

Not so.

the best you can do is guess and they're not even
going to be good guesses. FFS SPSM

Recording log in and log out times are far from being wild guesses.

One way, although it has some drawbacks and flaws, is to have a user give
his/her IP address as they login. Or I suppose I could get their IP from an
email they send me to register. Then, I can check in awstats in CPanel and
see how long that IP address spends on the site.

One flaw includes certain modems or ISPs that give different IP addresses.
This could mess things up.

JG

 

[rf]

Login time is exact. Logout time is a total unknown.

One way, although it has some drawbacks and flaws, is to have a user give
his/her IP address as they login.

How silly. How many "users" know their IP address? How may "users" actually
know what an IP address *is*.

Those on dial up get a different one on every connect. Do you expect them to
run IPCONFIG just to find out what it is and tell you? No, they will lie,
or go elsewhere

Or I suppose I could get their IP from an
email they send me to register.

Where do you get the IP address from an email?

Then, I can check in awstats in CPanel and
see how long that IP address spends on the site.

What about people behind a proxy. Thousands of them may have the same IP
address. What about AOL? IIRC it swaps IP addresses for each end every
request.

One flaw includes certain modems or ISPs that give different IP addresses.
This could mess things up.

Royally.

Cheers
Richard.

 

[Toby A Inkster]

Where do you get the IP address from an email?

The "Received" headers, although not with 100% reliability.

 

[Whitecrest]

not if the person doesn't log out. I can't recall the last time I
actually logged out of anything. I just let it time out and I normally
leave a site long before that happens.

If I have you logged out after 10 minutes of inactivity, I can get a
pretty close approximation, I can also distinguish betwen people that
don't log off and those that do, so I can tweak the time, and get even
more accurate.

Exact no, but very useable data.