"chris" wrote

| I was under
| the impression that a cable connection is actually
| part of a party line, with hundreds of other customers
| sharing the same IP address.
|
| Not sure how cable companies do it, but ADSL and fibre connections all
have
| an IP allocated to them from the pool that their service provider owns. I
| imagine cable is not that different.
|

I wonder. There's been talk for some time of IP4
addresses being short. And as Char noted, in the
early days of cable, neighbors used to find each
other in their Network Neighborhood. It seems that
it would be cheaper and allow for more expansion
if cable companies can share IP addresses across
a group. So it seems farfetched that everyone
online could have their own IP address. But I don't
know the details of how it works. I thought maybe
someone else might.

| Many websites don't even
| have their own IP address.
|
| False. All web domains have an IP address.
| Again by definition.

They have an IP address by definition, but not
necessarily a dedicated IP address, which is what
I'm wondering about.

If you look at webhosting options you'll see that
a dedicated IP is sometimes an option. Probably the
cheapo servers like Dreamhost don't even offer it.
That limits how many customers they can have.
And IP4 addresses have already run out.

Shared IP means aaa.com, bbb.com and
ccc.com can all have the same IP address, which
points to their server. A requested page would
then be determined from the GET. So Dreamhost
doesn't have to dedicate either a device or an IP
to an individual customer. They just put each domain
in a separate folder on one machine and figure it
out as the GETS come in.

Given all that, it made me wonder whether a site
can really track visitors, realistically, by IP. In fact,
it's not unusual in my own web logs to see commercial
GETs coming from numerous, similar IPs, even for one
page and it's related images. And it's common (I don't
know why) to see things like an IP that resolves
to Brazil in terms of geolocation load a webpage,
followed by an IP from Europe that downloads a linked
file. Yet both show the same company in a hostname
resolution.

I guess the test would be whether deleting cookies
when the browser closes, and blocking local storage
"supercookies", allows one to get more articles.
And it could also be possible to track people with
web bugs, in which case deleting cache would also
be necessary. All of those methods seems more
realistic than IP blocking, but I just don't know
enough about it to be sure.

Increasingly I've seen what might be called
"passive aggressive" blocking. Pages designed to
malfunction if you don't allow them to run code,
show ads, etc. Forbes.com, for example, started
making their page all javascript. It's effectively a
software program. Similarly, someone sent a link to
a shopping page at "Google Express" the other day.
(First I've heard of that site). With script blocked
the page was blank. Looking at the source code I
saw that Google had embedded even the page
text into javascript. Little or no HTML. They're
deliberately showing a blank page if script is
disabled. "Yes, we want you to buy stuff here,
but not if we can't spy on you!"

In a similar vein, npr.org (of all people, a non-
profit news organization!) periodically shows me
a page with two choices: 1)Enable script and agree
to allow us to spy on you or 2) View our homepage
as nothing but a list of links.

And many sites now do things like putting a big,
gray rectangle on top of the content, which gets
removed by script. So without script I have to
disable CSS in order to read the page.

Yet I can always read articles at NYT as long as
I allow cookies. Though I've never tried reaching
their daily limit and deleting cookies. I'm usually
only going there to see one thing that's linked from
elsewhere. Interestingly, NYT was one of the first
to create a "paywall" and they claim their online
subscriptions are very successful, especially among
young people.