Yesterday Karl mentioned mid-session ip changes by AOL. Well to be
absolutely truthful, I wasn't aware that AOL was quite that crippled.

Anyway, in order to make things right without opening some security
holes, I need to find out about this mid-session ip change business.

Does AOL always do this, only when it's in the mood, or what? Damn
their lame asses, I might have to resort to ua identification to see
if it's AOL (or maybe ip-identification) and setting a freakin cookie,
and I really don't want to have to do that, I'd rather keep on being
cookie-free. Does AOL still require its users to use the AOL-supplied
browser, or can they use any browser they choose?

Are there any other isp's out there that are also so spastic that they
change ip's in mid-session?

--
http://www.ren-prod-inc.com/hug_soft...action=contact

hug wrote:

That doesn't quite explain it. The ip will probably change for almost every
request.

Here's a section of my access log that shows an AOL'er requesting 1 page,
followed by requests for the images on that page.


195.93.21.98 - - [15/Mar/2006:12:44:37] "GET /xxxx
195.93.21.42 - - [15/Mar/2006:12:44:37] "GET /images/
195.93.21.3 - - [15/Mar/2006:12:44:37] "GET /images/
195.93.21.36 - - [15/Mar/2006:12:44:37] "GET /images/
195.93.21.36 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.99 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.68 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.135 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.73 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.38 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.132 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.137 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.137 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.69 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.34 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.106 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.72 - - [15/Mar/2006:12:44:38] "GET /images/
195.93.21.130 - - [15/Mar/2006:12:44:38] "GET /images/




--
Brian Wakem
Email: http://homepage.ntlworld.com/b.wakem/myemail.png

Brian Wakem <no@email.com> wrote:

Thanks. Have you seen other ip's from them besides 195.193.21.xxx?
It looks like I'll need to deal with this somehow. Damn, I hate
cookies. The options I see are:

a) Let AOL users ESAD (my preference, but not feasible).
b) Identify AOL users by browser ua
c) Identify by ip

--
http://www.ren-prod-inc.com/hug_soft...action=contact

hug wrote:

A quick look shows some others in the 207.200.116.* range. Knowing AOL
there could be many others.



My preference too.



You can use standard IE with AOL I believe.



Good luck.



--
Brian Wakem
Email: http://homepage.ntlworld.com/b.wakem/myemail.png

And lo, hug didst speak in alt.www.webmaster:

There is also option d) Just don't deal with it, and save yourself from a
lot of stress.

Even cookies won't help if a user presses the back button, or is browsing
your site using multiple tabs/windows. Trying to track the paths of
everyone accessing your site is a Sisyphean task.

I compare it to taking a poll. You can ask a question of 10,000 people
within a population of one million and statistically get within 5% of the
entire population's *actual* concensus; that's analogous to normal stats
as reported by Webalizer, AWStats et al. However, to get that last 5% of
accuracy, which is what you're trying to do by path tracking, you'd need
to ask the same question to every one of the remaining 990,000 people.

You should ask yourself why all that extra effort for so little gain is
worth it.

Grey

--
The technical axiom that nothing is impossible sinisterly implies the
pitfall corollary that nothing is ridiculous.
- http://www.greywyvern.com/orca#search - Orca Search: Full-featured
spider and site-search engine

GreyWyvern <spam@greywyvern.com> wrote in
newsp.s6gmwtd6sl6xfd@news.nas.net:

EEEEEEEEEEEEEEEEEEEEEEEEEEEEEE!!!!
I just realized something.
OK, so if we say we have a small sample size (read as: less popular
site), then our margin of error in these statistics will be very high.

To overcome that margin of error, we need a larger sample size (a more
popular site). Unfortunately, the more popular the site, the more likely
it is that the site will be cached by networks & proxies, AND the more
likely it is that AOLers will be swapping IPs with each other.





--
Karl Groves
http://karlcore.com
http://chevelle.karlcore.com

Accessibility Discussion List: http://smallerurl.com/?id=6p764du

Brian Wakem wrote:

and, AFAIR, AOL users can fire up any other browser, once connected via
their version of IE, and they have to leave IE open.

That is of course, assuming that they know how to do this, which will be
an extremely small percentage of their clientele.

--
-bts
-Warning: I brake for lawn deer

Fleeing from the madness of the Somewhat jungle
Karl Groves <karl@NOSPAMkarlcore.com> stumbled into news:alt.www.webmaster
and said:

stop it - just go with option d) and blame it all on Grey.

Option d) also helps you avoid floundering for more bullsh1te when the
customer ego needs feeding.

--
William Tasso

whither a trophy?

GreyWyvern <spam@greywyvern.com> wrote:

That's certainly an option, though not a real desirable one aside from
the stress, since any mid-session ip change would make the contents of
their shopping cart disappear, which is a bad thing.

I'm not clear on what you're saying Grey, are you saying that the back
button deletes cookies? If so, that's a new one to me (not saying it
would surprise me since I'm getting used to being hit from behind).

I'll not argue that one either way.

If you were going to offer an ecommerce engine wouldn't you want the
contents of the shopping cart to stay put until the guy completes the
order? Since the shopping cart is on the server, I need to identify
the guy on the other end. Most of that is done by using unique links,
but if he copies a link and pastes it in an email to a friend, I need
to give the friend a fresh state without any specific information in
it. More thinking on my part is definitely required.

--
http://www.ren-prod-inc.com/hug_soft...action=contact

Karl Groves <karl@NOSPAMkarlcore.com> wrote:

AND the more server resources will be used if you're doing anything
nontrivial on the server.

I'll state an opinion, please feel free to tell me how wrong I am. <g>

There are two kinds of data you can collect about site usage. One
pertains to which pages are hit the most. The other pertains to the
order in which pages are navigated. Both are more useful than
nothing, but in my own work I've found the second (navigation routes)
to be much, much more useful in figuring out what's going on and
improving the site to encourage people to wander toward that "confirm
order" button. Which is to say, sample size really isn't that
important since it only takes a few page-navigation sessions to give
enough data to make some improvements (which then invalidates all
previous data and requires you to take a fresh look at things).

--
http://www.ren-prod-inc.com/hug_soft...action=contact