Google Mixed Backlinks For My 2 Sites - but different IPs and whois

[alexo]

Hello,

i have 2 sites.
1-st site, it's a private forum (6 months old), which we use just for communication for our staff from all over the world.

Site is private, based on board script and as that site is just for us, i even close regustration and never promote this site. No link to this site.

2-nd site, it's acamedical educational site (8-10 years)old, but lot of links from dmoz, google dir, yahoo etc.

Few months ago PR of this site was dropped from PR5 to GREY BAR (1 week), and only after this i check PR3 (until now) and just a small amount of links if i use [link: command]

Today, i check that first site have PR5!
af first i think it's an error, than i check it via <pagerank tool> sites and ... it's really PR5.

But the most interesting is, when i try link:1-st_site

i check more than >300 links. and all this links are very known backlinks {for me} :-)

Oh my god, it's backlinks of my 2-nd site :-)

I clear my cookies, change my ISP, browser, even try via proxy site ... it's real !

P.s. both this sites don't have any realation, any backlink. They even aren't on the same server.

First site is registered by OSI (Open society Institute), second site -> by private person.

so ...

can any1 explain, how it's possible?

[whitenight]

The most simple answer is:

Google as a domain registrar is DOING THINGS IT HAS NO BUSINESS DOING.

Oh sure, I haven't looked up the exact laws regarding this yet, so this is just a MORAL argument.

But rest assured, I will be looking CLOSELY into this as it STINKS of BIG BROTHER and just plain nonsensical "not getting IT" on Goog's part.

Yes, I'm sure there's another "reasonable" explanation.
But until I hear it convincingly and logically, I'm going with the OBVIOUS one....

Goog can't help itself from associating common "ownership" sites through its
BIG BROTHER ACTIVITIES.

-----
However, I'm willing to ENTERTAIN an explanation from Goog Employees #1-3 before I go absolutely ballistic on this issue.

[alexo]

hmm ...
another interesting think in this history is, that how G find any relation between this 2 domains?

As i said, first domain is registered by OSI, have *.eu extention and is private domain.

So if you, or any SE open this domain, can see just the 1 page, with short message, that this site is private site for personal uses of OSI. Nothing else, no other links, and there ins't any link in internet to this domain. And of course there isn't any ADV(adsense) on this domain.

Second domain is on national tld. So even this 2 domains aren't registered by the same registrar.

As i said, that are on different servers.

THat's why i don't understand, how G. find any relation between this 2 domains. The only relation is that users, that some users of 2-nd site, also sometimes open 1-st site in their browsers !

Nothing else. But as i know, this what sites i open, or use it's private information, so google cann't use this information.

So .. ?

[tedster]

Google (and other companies, too) can buy traffic data from ISPs. If you think a search engine on its own collects a lot of data, think about your ISP and their record of what you do online. I'm not saying that this is definitely how Google could connect your domains, but it is one possibility that wouldn't depend on registrar, hosting or interlinking information.

[alexo]

of course, i know the ISP can collects a lot of data, but i'm sure, that this information don't go out from my ISP(national small ISP) to google.

[alexo]

i find just 1 relation between this 2 domains

I put in htaccess of the 1-st site such line

ErrorDocument 404 http://www.2-nd_domain.com

------
so can such redirection hijack backlinks and PR of the first domain?

[edited by: Robert_Charlton at 10:02 pm (utc) on Dec. 23, 2008]
[edit reason] disabled link [/edit]

[jdMorgan]

Using a canonical URL instead of a local URL-path in ErrorDocument results in a 302 redirect to the specified domain. So, you have apparently 302-Hijacked your own site.

This redirection behavior is described in the Apache ErrorDocument documentation.

Consider using local custom error pages (putting the custom error pages on the 1st site and invoking them as "ErrorDocument NNN /path-to-local-error-document.html") and then providing a text link to the second site, if you really need the visitor to get to the second site for some reason.

Even if the pages are on the same domain, it is almost always a mistake to specify a canonical URL for an ErrorDocument. The reason is that the client sees a 302-Redirect status, and not the actual server error code. This can confuse search engines, which is something we don't like to do...

Jim

[alexo]

thank you Jim, i understand it too late.

so i Hijacked my 2-nd domain.

The last question, does 2-nd domain PR drop relative to this process? or it's just independent process?

If it's relative, i can Hijack and PR10 site (f.e. google.com) and does in this case google PR dropped too?

[julinho]

Using a canonical URL instead of a local URL-path in ErrorDocument results in a 302 redirect to the specified domain. So, you have apparently 302-Hijacked your own site.

Hadn't Google taken care of the 302 hijacks a few years ago?

The only relation is that users, that some users of 2-nd site, also sometimes open 1-st site in their browsers !

If you think a search engine on its own collects a lot of data,

I think Google collects more and more data about user behavior, and the ranking algorithm depends more and more on those data (as suggested by, among others, this patent [webmasterworld.com]).

It would be interesting to know if, after fixing the .htaccess, Google sorted the domains correctly.

[alexo]

>>>It would be interesting to know if, after fixing the .htaccess, Google sorted the domains correctly.

Not, yet
the same PR and the same output [link:] as when i post this thread.