Who is bookmarking my site?

[Bookmarks]

The statistics monitoring software we are using does not monitor who is
visiting us from bookmarks. To me, that would be very informative data to
have. First of all, it would tell us "who" thinks our website is important
enough to include us as book marks. But, more importantly, I can run
statistical algorithms with that information to try to guess at who are
bookmarked/repeat/new visitors to the website.

I think there should be ways to do some server and/or client side
programming to grab this information.

How can you glean this info from the weblogs, for users you can/should not
force through a registration rigor?

How could cookies help with that?

otf

 

[Chris Smith]

The statistics monitoring software we are using does not monitor who is
visiting us from bookmarks. To me, that would be very informative data to
have. First of all, it would tell us "who" thinks our website is important
enough to include us as book marks. But, more importantly, I can run
statistical algorithms with that information to try to guess at who are
bookmarked/repeat/new visitors to the website.

I think there should be ways to do some server and/or client side
programming to grab this information.

You can't solve this problem completely on the server, and you can't
solve it completely on the client without privileged code (signed
applets, replacing the web browser, or some such thing).

What you can do is pay attention to the Referer header. If it's not
present, that could give you a hint that the user either typed in the
URL by hand or bookmarked the site; or, of course, they could just be
using a browser that doesn't send Referer header.

--
www.designacourse.com
The Easiest Way To Train Anyone... Anywhere.

Chris Smith - Lead Software Developer/Technical Trainer
MindIQ Corporation

 

[Roedy Green]

The statistics monitoring software we are using does not monitor who is
visiting us from bookmarks. To me, that would be very informative data to
have. First of all, it would tell us "who" thinks our website is important
enough to include us as book marks. But, more importantly, I can run
statistical algorithms with that information to try to guess at who are
bookmarked/repeat/new visitors to the website.

try a search on google with link:mindprod.com

 

[Sanjay]

The statistics monitoring software we are using does not monitor who is
visiting us from bookmarks. To me, that would be very informative data to
have. First of all, it would tell us "who" thinks our website is important
enough to include us as book marks. But, more importantly, I can run
statistical algorithms with that information to try to guess at who are
bookmarked/repeat/new visitors to the website.

As Chris said bookmark visitors could just be a guess on non-referral
visitors.
You can take a look at some web analytics products, google recently
announced their web analytics product which is free.

I think there should be ways to do some server and/or client side
programming to grab this information.

How can you glean this info from the weblogs, for users you can/should not
force through a registration rigor?

Most softwares these days have client code to capture information.

How could cookies help with that?

Of course using cookies you can find out repeat visitor's/new visitors
to your web-site.

otf

Well, I am into this web analytics business too

 

[Evertjan.]

Bookmarks wrote on 16 jan 2006 in comp.lang.java.javascript:

The statistics monitoring software we are using does not monitor who
is visiting us from bookmarks. To me, that would be very informative
data to have.

It is not yours to have.

 

[Alex]

Easy.
For url http://abc.xyz always forward it to (for example)
http://abc.xyz/abc.html
User can bookmark only what is in the browser - so it would be
http://abc.xyz/abc.html

The rest is obvious.
All http://abc.xyz are new requests.
All http://abc.xyz/abc.html are bookmarked requests.

P.S. That's not java question.

 

[Oliver Wong]

The statistics monitoring software we are using does not monitor who is
visiting us from bookmarks. To me, that would be very informative data to
have. First of all, it would tell us "who" thinks our website is important
enough to include us as book marks. But, more importantly, I can run
statistical algorithms with that information to try to guess at who are
bookmarked/repeat/new visitors to the website.

I think there should be ways to do some server and/or client side
programming to grab this information.

How can you glean this info from the weblogs, for users you can/should not
force through a registration rigor?

How could cookies help with that?

It's unclear what information you want. The name, home address and phone
numbers of everyone visiting your website? A list of IP addresses? Something
else?

- Oliver

 

[onetitfemme]

// __

It's unclear what information you want. The name, home address and phone
numbers of everyone visiting your website? A list of IP addresses? Something
else?

I need their IP addresses (and HostNames, which could be queried from
the back end, but it would be nice if you could get it through some JS
or java client code)

// __

P.S. That's not java question.

I know, the thing is that I could see there would be need for some
back end programming/customization and pushing to the client, and I use
Tomcat as web server

otf

 

[Sanjay]

As Chris said bookmark visitors could just be a guess on non-referral
visitors.
You can take a look at some web analytics products, google recently
announced their web analytics product which is free.

Normally if the visit of the user starts from non-public URL's then 95%
chances are there that those are bookmarked.
Clubbed with repeat visitors I guess you can narrow down your
estimation.

 

[Oliver Wong]

// __


I need their IP addresses (and HostNames, which could be queried from
the back end, but it would be nice if you could get it through some JS
or java client code)

So to answer your original question, which was,

<quote>
How can you glean this info from the weblogs, for users you can/should not
force through a registration rigor?
</quote>

is as follows:

Most web servers will log the IP addresses of every request[*]. So you
should be able to just open the log file, and scan through it for the
information you need (where "scan through it" could be performed either by
human eyes or by a program). For HostNames, the mapping between an IP
address and a hostname can vary with time, so if you really need that info,
it's best to get your server to determine the hostname the moment the
request is made.

If you want that info to be available to a client to your website (e.g.
via a JavaApplet), you have to give your client read permission on your
logs.

- Oliver

[*] The Apache HTTP server, for example, is capapble of logging the IP
address and hostname of every request made to it.

 

[onetitfemme]

Oliver Wong wrote (and otf answered):

Most web servers will log the [*] The Apache HTTP server, for example, is capapble of

logging the IP address and hostname of every request made to it.of
every request[*]. ...

[*] The Apache HTTP server, for example, is capapble of logging the IP address and

hostname of every request made to it.
otf: the IP address that log files show is the one of their
gateway/proxy server
interfacing the internet, not the one inside their networks (which is
what you need
to internally tell apart one machine from the other and know better
about their
"bookmarks")

For HostNames, the mapping between an IP address and a hostname can vary with time,

so if you really need that info, it's best to get your server to
determine the hostname
the moment the request is made.
otf: that, of course, is taxing. So some caching strat and a close by
DNS server would
be really helpful

Thanks
otf

 

[Oliver Wong]

otf: the IP address that log files show is the one of their
gateway/proxy server
interfacing the internet, not the one inside their networks (which is
what you need
to internally tell apart one machine from the other and know better
about their
"bookmarks")

It's probably impossible to get the "internal" IP address. There is no
reason for their gateway/proxy/router to give you that information, and some
may outright refuse to do so as part of a security measure.

- Oliver

 

[nanditha.g]

It's probably impossible to get the "internal" IP address. There is no
reason for their gateway/proxy/router to give you that information, and some
may outright refuse to do so as part of a security measure.

- Oliver

I was just wondering if this exercise is to find if the users are
repeat / new visitor, then a persistent cookie would do. (Of course you
need to consider the fact that a user can remove the cookies or reject
it)

 

[Oliver Wong]

I was just wondering if this exercise is to find if the users are
repeat / new visitor

Indeed. See http://www.catb.org/~esr/faqs/smart-questions.html#goal

<quote>
Describe the goal, not the step

If you are trying to find out how to do something (as opposed to reporting a
bug), begin by describing the goal. Only then describe the particular step
towards it that you are blocked on.

Often, people who need technical help have a high-level goal in mind and get
stuck on what they think is one particular path towards the goal. They come
for help with the step, but don't realize that the path is wrong. It can
take substantial effort to get past this.
</quote>

- Oliver

 

[jcsnippets.atspace.com]

Easy.
For url http://abc.xyz always forward it to (for example)
http://abc.xyz/abc.html
User can bookmark only what is in the browser - so it would be
http://abc.xyz/abc.html

The rest is obvious.
All http://abc.xyz are new requests.
All http://abc.xyz/abc.html are bookmarked requests.

I don't think it's quite that easy - what if I remember the address and type
it in myself? The request would not originate from my bookmarks, but as a
new request from my browser...

 

[Alex]

Question was to recognize bookmarks and typing.
And answer was easy.
It was not question "was it new request or not". In this case cookies
would help.
But then you can ask - "what if I delete cookies?" Or - what if I use
different browser from the same machine... and so on and so on...
And, again, this is not Java question.
Alex.

 

[onetitfemme]

It was not question "was it new request or not". In this case cookies would help.
But then you can ask - "what if I delete cookies?"

otf: Well, yes "what if people delete cookies? (Which they should
anyway ;-))

Or - what if I use different browser from the same machine ...

otf: Well, this is why I was trying to find a way (even if not 100%
reliable) to get the network IP address of the computer. Say, I know
(request headers tell you) that the user accepts java and you are able
to send a signed applet to the user crafted in a way that you may run
via the run time something like the command line statement via exec
"ifconfig -a"
. . .

And, again, this is not Java question.

otf: No, it is not. But I remember these guys at servlets.com
initially developed something called "BrowserHack" that was able (this
of course is marketting bs) to detect 120 different features from the
computer accessing the site. Inclusing the monitor geometry and depth
....
That must have been heuristics to a large extent, but probably it is
as far as it gets

otf

 

[Oliver Wong]

You have a very strange reply style. Not nescessarily bad; just
remarking on its uniqueness.

otf: Well, this is why I was trying to find a way (even if not 100%
reliable) to get the network IP address of the computer. Say, I know
(request headers tell you) that the user accepts java and you are able
to send a signed applet to the user crafted in a way that you may run
via the run time something like the command line statement via exec
"ifconfig -a"

The applet would have to be signed, and then the JRE would ask the user
if they are willing to trust this applet. Otherwise, you cannot run commands
on the machine via exec.

And if you have 2 layers of NATs (like I have at my home), this would
fail anyway.

. . .
otf: No, it is not. But I remember these guys at servlets.com
initially developed something called "BrowserHack" that was able (this
of course is marketting bs) to detect 120 different features from the
computer accessing the site. Inclusing the monitor geometry and depth
...
That must have been heuristics to a large extent, but probably it is
as far as it gets

That was probably done via JavaScript, not Java. In JavaScript, it is
possible, for example, to request a pop-up window to be maximized, and then
to get the size of that window. From that, you can determine what the
monitor geometry is.

- Oliver