Programed redirect, can google read it?

need to mask external links

[jitesh]

I need to mask some of the external links from my sites. There are some practical difficulties in removing them. Therefore I need to redirect them through programs.

Now the point is can google read and follow this type of links [ http://www.mydomain.com/cgi-bin/http://www.externalsite.com ]. Here the external site is visible. The another way to do is through this type of URL's [ http://www.mydomain.com/rd/NUMBER ] where the number correspond to a specific site.

Which will be the best way to hide those links from search engines?

Thanks

[korkus2000]

Using the number will be the best way. I have heard many say that google will follow any thing that has http:// in it. Not only links but in the text also. I think google will follow both links though. I have also heard people say there is some PR givin off to the external page using these links.

[Yidaki]

I have heard many say that google will follow any thing that has http:// in it. Not only links but in the text also. I think google will follow both links though. I have also heard people say there is some PR giving off to the external page using these links.

Just strip the http:// part off the url and put it back again the moment you redirect using Location: $redirectTo. Additionally i use a subdirectory for the redirect sript and disallow it through robots.txt. I have the feeling that sometimes Disallow: /jumpto.php doesn't work. Call me paranoid but i feel safer using a url like /ressources/jumpto.php along with Disallow: /ressources/. Can't hurt and googlebot doesn't resolve any of these links - definitely!

Ok, here's the script jumpto.php - works perfectly for me:

<?php
$redirectTo = $HTTP_GET_VARS["where"];
if ( $redirectTo == '' )
{
// backdrop in case $redirectTo is empty
$redirectTo = 'www.example.com';
}
print Header("Location: [$redirectTo");...]
// optional (else it's a 302 temporary status):
// print Header("HTTP/1.0 301 Moved Permanently", true)
?>

Now your links should look like this:
http://www.example.com/ressources/jumpto.php?where=www.externalsite.com

... and your robots.txt should contain this:
User-Agent: *
Disallow: /ressources/

I use this technique to avoid any potential bad neighborhood troubles and controll the flow of pr through my sites.

A positive side effect of the above solution is that your log stats wil be much better analyzable since you can search and sort outgoing clicks by domain instead of dealing with database numbers. And it's pretty easy to code, too, when you add new links to your site!

Btw: i'm pretty sure it doesn't but ... does the fact that google doesn't resolve the click script really mean that no pr transfer is involved and that googlebot doesn't record the portion after the "=", GoogleGuy? You're not collecting even plain text url matches that have neither a http:// part nor a href tag in front of it, or!? Googlebot is pretty well known as a link eating monster ... >:)

Added:
After thinking a bit i came to the conclusion that you could additionally encode any www. occurence within a url using a hard coded flag. ie: replace www. with #w# and back ... that'll be even more secure. ;)

[jitesh]

Thanks for the insight Yidaki. That was highly appreciable.

Let's assume that this type of links will not be followed by google bot...but still IMO they are links, so the very point is if they will pass PR or not. The PR formula divides page rank of each page by the no of links present in the page. The question is whether only resolved links are taken into consideration or just any links count.