PageRank Channeling Part 2

[Pyrrhonist]

I've been looking into the possibilities of masking outbound links using JavaScript. Lately, I've been using this piece of code when creating an outbound link I don't want to pass PR on to.

<a href="javascript:location.href='http://www.seo-guy.com'">SEO Optimization</a>
What I'm trying to figure out is where in the iteration loop of reading a page does Googlebot calculate PR?

The possible options that I have come up with are when it follows the link or when it reads an anchor tag.

My theory is that if PR is calculated when the anchor tag is encountered (ie, the entire page is read, the number of anchor tags are counted, and that's what the PR pass-on amount is calculated with) then we would see a problem with on-page links - the ones that jump to a different point on a page - as they are created using the anchor tag as well since they would point the PR right back to the same page and create a loop... which obviously doesn't happen.

I'm really not sure of the answer to this question, and at this point am just making things more complicated for myself. If anyone has any thoughts, or ideas on how to test this, please post, and we'll see if we can prove it one way or another.

SEO Bob

[jocelyn]

[QUOTE=SEO Bob]If anyone has any thoughts, or ideas on how to test this, please post, and we'll see if we can prove it one way or another.[/QUOTE]
I've read somewhere (can't remember which Forum) that a page is stripped of the useless code to keep only the substance.
The pictures, java, flash are out the door... maybe it's the same with duplicate links and links to a text on the same page... until only spider food is left.
At this point chances are there is only anchor and links that are valid that are left.

I can't remember where, but it was pretty logical and this is why I remember this.
It saves HD space (4billion pages...) and should speed up treating...

It does not answer anything but can't hurt.

[digitalpoint]

Not an answer to the original question, but the tags, JS, etc. are not stripped off for storage or evaluation. Look at the cached version of a page in Google for example, it's all there.

- Shawn