Site Unload?

[Ross M. Greenberg]

I want to have an onUnload that runs only when the site is left, not when
the current page is unloaded. Any ideas?

Thanks!

Ross

[Randy Webb]

Ross M. Greenberg wrote:

"Randy Webb" <Hi************ @aol.com> wrote in message
news:lP******** ************@co mcast.com...

You can take this "challenge" for as long as you want, and you will
still ultimately fail in the end. The browser has NO way to know where I
am going, unless its from a link within the page.

I thought an event was triggered when a web page is about to be accessed --
aren't custom 404's generated locally for invalid URLs, for example?

That depends on how the server of the invalid URL is setup. But it's
irrelevant.

But tell us this.

Lets say I am on page234.html on your site.
I have page352.html bookmarked.
I click the page352.html bookmark.

Did I leave the site?

If the http://<whatever> is the same, no. If different, yes.

The browser has no way of knowing that. It only knows that you left
*one* page, not how you left, just that the page is unloading. The one
exception would be if you had onclicks in all your links to internal
pages that set a variable, onunload you check that variable. Search the
archives on how its done.

But that still does not cover me leaving one page of your site to
another page of your site via a favorites or by typing in the URL.

The answer is no but your script will fail to know that.

The survey in mind seeks to ask them directly what they liked or did not
like about the site. I'm seeing more than a few users leaving the site
from
its home page. I need to find out why so I can fix what they don't like.

Then put the link on the main page. As was pointed out, if the user
wants to fill it out, they will. <sarcasm>If they don't, then popping up
a survey page will surely entice them to.</sarcasm>

But this was a means to get them to let me know *as* they leave, not when
they're still at/just arrived at the site.

Then put the link on all pages. But if people are going to fill it out,
they will do so. Attempting to popup a survey to ask what you thought of
a site is very anti-user oriented.

But there is no way for you to definitively know where I am going when I
leave one page of your site.

--
Randy
comp.lang.javas cript FAQ - http://jibbering.com/faq

[Evertjan.]

Ross M. Greenberg wrote on 15 dec 2004 in comp.lang.javas cript:

I thought an event was triggered when a web page is about to be
accessed -- aren't custom 404's generated locally for invalid URLs,
for example?

No, not if the domain is valid.

--
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[Ross M. Greenberg]

I listened to you folks, and stuck the survey via a link.

"Randy Webb" <Hi************ @aol.com> wrote in message
news:Se******** ************@co mcast.com...

Ross M. Greenberg wrote:

"Randy Webb" <Hi************ @aol.com> wrote in message
news:lP******** ************@co mcast.com...

You can take this "challenge" for as long as you want, and you will
still ultimately fail in the end. The browser has NO way to know where I
am going, unless its from a link within the page.

I thought an event was triggered when a web page is about to be accessed --

aren't custom 404's generated locally for invalid URLs, for example?

That depends on how the server of the invalid URL is setup. But it's
irrelevant.
But tell us this.

Lets say I am on page234.html on your site.
I have page352.html bookmarked.
I click the page352.html bookmark.

Did I leave the site?

If the http://<whatever> is the same, no. If different, yes.

The browser has no way of knowing that. It only knows that you left
*one* page, not how you left, just that the page is unloading. The one
exception would be if you had onclicks in all your links to internal
pages that set a variable, onunload you check that variable. Search the
archives on how its done.

But that still does not cover me leaving one page of your site to
another page of your site via a favorites or by typing in the URL.

The answer is no but your script will fail to know that.
The survey in mind seeks to ask them directly what they liked or did notlike about the site. I'm seeing more than a few users leaving the site

from

its home page. I need to find out why so I can fix what they don't like.
Then put the link on the main page. As was pointed out, if the user
wants to fill it out, they will. <sarcasm>If they don't, then popping up
a survey page will surely entice them to.</sarcasm>

But this was a means to get them to let me know *as* they leave, not when they're still at/just arrived at the site.

Then put the link on all pages. But if people are going to fill it out,
they will do so. Attempting to popup a survey to ask what you thought of
a site is very anti-user oriented.

But there is no way for you to definitively know where I am going when I
leave one page of your site.

--
Randy
comp.lang.javas cript FAQ - http://jibbering.com/faq

[Ross M. Greenberg]

Well, I stuck the survey in via a link

"Evertjan." <ex************ **@interxnl.net > wrote in message
news:Xn******** ************@19 4.109.133.29...

Ross M. Greenberg wrote on 15 dec 2004 in comp.lang.javas cript:

I thought an event was triggered when a web page is about to be
accessed -- aren't custom 404's generated locally for invalid URLs,
for example?

No, not if the domain is valid.

--
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[Evertjan.]

Ross M. Greenberg wrote on 15 dec 2004 in comp.lang.javas cript:

"Evertjan." <ex************ **@interxnl.net > wrote in message
news:Xn******** ************@19 4.109.133.29...

Ross M. Greenberg wrote on 15 dec 2004 in comp.lang.javas cript:

> I thought an event was triggered when a web page is about to be
> accessed -- aren't custom 404's generated locally for invalid URLs,
> for example?
>

No, not if the domain is valid.

Well, I stuck the survey in via a link

[please do not toppost on usenet]

I don't know what you mean by this.

Is your survey stuck in a link?

;-}
--
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[Laurent Bugnion]

Hi,

Evertjan. wrote:

Ross M. Greenberg wrote on 14 dec 2004 in comp.lang.javas cript:

I want to have an onUnload that runs only when the site is left, not when
the current page is unloaded. Any ideas?

"site is left" is not defined,
[unless you have a button saying "I solemnly declare to leave the site"]

With serverside ASP scripting you could do something ON THE SERVER when a
session ends. This is not very reliable and has no effect on clientside.

To be precise, the Session_End event (or similar on other platforms) is
very reliable. It will be called when the Session ends, but the session
often ends on a timeout, typically 20 minutes in ASP.NET, unless the
user is nice enough to press a "Log Off" button (which allows to
terminate the session earlier). The Session_End event *has* to be called
or else the Server process will explode with unroomed resources.

And yes, it has no effect on the client at all.

Laurent
--
Laurent Bugnion, GalaSoft
Software engineering: http://www.galasoft-LB.ch
Private/Malaysia: http://mypage.bluewin.ch/lbugnion
Support children in Calcutta: http://www.calcutta-espoir.ch

[Evertjan.]

Laurent Bugnion wrote on 17 dec 2004 in comp.lang.javas cript:

With serverside ASP scripting you could do something ON THE SERVER
when a session ends. This is not very reliable and has no effect on
clientside.
To be precise, the Session_End event (or similar on other platforms)
is very reliable. It will be called when the Session ends, but the
session often ends on a timeout, typically 20 minutes in ASP.NET,

I was talking ASP, not ASP.NET
unless the user is nice enough to press a "Log Off" button (which
allows to terminate the session earlier). The Session_End event *has*
to be called or else the Server process will explode with unroomed
resources.
Same, not in ASP classic.
And yes, it has no effect on the client at all.

Paradox: some clients gets mad, because of that lack of effect.

--
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[Laurent Bugnion]

Hi,

Evertjan. wrote:

Laurent Bugnion wrote on 17 dec 2004 in comp.lang.javas cript:

With serverside ASP scripting you could do something ON THE SERVER
when a session ends. This is not very reliable and has no effect on
clientside .

To be precise, the Session_End event (or similar on other platforms)
is very reliable. It will be called when the Session ends, but the
session often ends on a timeout, typically 20 minutes in ASP.NET,

I was talking ASP, not ASP.NET

unless the user is nice enough to press a "Log Off" button (which
allows to terminate the session earlier). The Session_End event *has*
to be called or else the Server process will explode with unroomed
resources.

Same, not in ASP classic.

That's surprising. I believe you, of course (I am much better in ASP.NET
than ASP), but I find the apparent lack of reliability of this critical
event a huge problem for ASP servers. How do you handle this?

Laurent
--
Laurent Bugnion, GalaSoft
Software engineering: http://www.galasoft-LB.ch
Private/Malaysia: http://mypage.bluewin.ch/lbugnion
Support children in Calcutta: http://www.calcutta-espoir.ch

[Evertjan.]

Laurent Bugnion wrote on 19 dec 2004 in comp.lang.javas cript:

unless the user is nice enough to press a "Log Off" button (which
allows to terminate the session earlier). The Session_End event *has*
to be called or else the Server process will explode with unroomed
resources.

Same, not in ASP classic.

That's surprising. I believe you, of course (I am much better in ASP.NET
than ASP), but I find the apparent lack of reliability of this critical
event a huge problem for ASP servers. How do you handle this?

Under ASP, server process will NOT explode.
[Resources are automaticly ended at the end of each page rendering.]

That should make ASP much more reliable than ASP.NET as you describe.

I suppose [I never even tasted it] the problem lies with ASP.NET.

--
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[Laurent Bugnion]

Hi,

Evertjan. wrote:

Laurent Bugnion wrote on 19 dec 2004 in comp.lang.javas cript:

unless the user is nice enough to press a "Log Off" button (which
allows to terminate the session earlier). The Session_End event *has*
to be called or else the Server process will explode with unroomed
resources .
Same, not in ASP classic.

That's surprising. I believe you, of course (I am much better in ASP.NET
than ASP), but I find the apparent lack of reliability of this critical
event a huge problem for ASP servers. How do you handle this?

Under ASP, server process will NOT explode.
[Resources are automaticly ended at the end of each page rendering.]

That should make ASP much more reliable than ASP.NET as you describe.

I suppose [I never even tasted it] the problem lies with ASP.NET.

OK, that's the case in ASP.NET too (each Page object is deleted at the
end of every call, and of course its members too), but you also have
Session objects, for example you will save there DB connections, file
contents, "global" variables, etc... IIRC, ASP also have the Application
and Session objects, doesn't it? So my question was, how do you know
when to room the resources saved in the Session object if you don't know
when the said session has timed out?

Sorry, for the more and more OT post, it just made me curious ;-)

Laurent
--
Laurent Bugnion, GalaSoft
Software engineering: http://www.galasoft-LB.ch
Private/Malaysia: http://mypage.bluewin.ch/lbugnion
Support children in Calcutta: http://www.calcutta-espoir.ch