Javascript links now being followed or cached?

[Marcia]

I'm not quite sure how to describe what I've just seen, but in a site's left navigation there's a 3-level nested menu, with the one visible link appearing to be Javascript, linking to the next level "flyout menu", followed by the third sub-level.

If you look at the Google text cache, all those links are listed as indexed - including the ones "visible" on the site pages that appear to be Javascript when mousing over and looking at the source code.

Are JS links now being followed or fetched when crawled, or could that be some kind of hybrid code? I don't ever use JS, so I'm unaware of its coding possibilities; but it's always been assumed Google doesn't follow JS links and this looks like they have.

[tedster]

Are you sure that javascript is generating the links? Many times the javascript is simply switching the visibility of the div that contains the links, but the link itself is right there in the source code. It's just not visible until there's a mouseover.

Not that I would put it past Google to start finding some links that were previously obscured by javascript - but I've yet to see it.

[Marcia]

I can click on the links on all levels of that menu in a regular browser that's Javascript enabled. This is what you see mousing over for the first two levels, and the third level is a regular hyperlink - - every which way tried, you see this for the first two:

javascript:void()

1) Browser with JS enabled - all clickable
2) Browser with JS disabled - first two levels not clickable (the 3rd is)
3) Google translate with JS disabled: first two not clickable (the 3rd is)
4) Google translate with JS enabled: first two not clickable (the 3rd is)

But all three show up in the text cache as though they were "normal" <a href=> links, which is what got me wondering. And then, just the title of this 2006 paper I ran into a while back is what I'd call provocative:

PageRank without hyperlinks: Structural re-ranking using links induced by language models [arxiv.org]

And a patent application (US 2005/0149851 in case it disappears):

Generating hyperlinks and anchor text in HTML and non-HTML documents [appft1.uspto.gov]

They can also identify what different portions of pages are (i.e. navigation, footer, body text), so it's logical that there could be a connection made between a page and its inlinks, especially being on the same domain.

The time I saw a non-linked URL show up was back when Google was still showing backlinks and I caught some guestbook spam (it was someone's pharm site, around 2002-2003) that showed up as link: that wasn't a link in the guestbook.

It's interesting to speculate on; there have been questions in posts here for a while on whether or not non-linked URLs can get counted, and these give some food for thought.

[edited by: Marcia at 6:17 am (utc) on Oct. 13, 2008]

[Marcia]

CORRECTION:

I just tried it again, and only the third level sub-menu links are clickable, not the first two, even though the cursor make them look like clickable links. That makes it more mysterious to me why they all show up cached with the anchors underlined the way regular links generally are.

[edited by: Marcia at 6:40 am (utc) on Oct. 13, 2008]