QUOTE(Scottie @ Oct 17 2005, 09:56 PM)
Let me rephrase- they ignore the stuff that they don't need.
More aptly put, they ignore the stuff they strip out. If your code is not structured properly, they can (and sometimes do) strip out content intended for indexing.
I can't believe they would think that a site was more relevant (or less relevant) simply because of the way it was structured- because the structure has nothing to do with whether or not the content is best suited to any particular query. When they start making judgements that CSS sites are more relevant than tabled sites, we'll see the top of the SERP's full of CSS positioned sites, right?
A few years ago, the idea that content positioned closer to the top of the page (in the source code) seemed to have merit because the search engines were indexing on a top-down model. That is, the top of the page was considered to be the most important part (this carried well beyond search engine indexing -- it applied in banner sales, where "above-the-fold" placement earned higher commissions, and where HTML design tutorials routinely advised people to place the most important content at the top).
Over the past couple of years, there has been increasing evidence (such as the contextual snippets Google offers in its results) that the search engines have been parsing the pages into segments. There have been some technical papers which explain the advantages of doing this, and clearly the software is more flexible and sophisticated than it used to be.
One of the primary reasons for parsing pages into segments, in my opinion, is to help the search engines separate navigational and advertising content from true page copy. People should not read more into that than there is. They need to figure out what the page is actually about (as opposed to what it promotes) to help determine relevance. They need to figure out what the page promotes (as opposed to what it is about) to help determine relationships. And they may be looking at how closely associated the relevance and relationships are (but that is all entirely speculative -- the patents and technical papers talk about doing these things, but we don't know what is actually being done).
The old "yes, but all things are never equal and the tabled site has x-y-z going for it" makes it nearly impossible to draw a cause and effect, but if it were an edge, I think there would be more evidence of it.
Although I use simplified table structures, I feel there is no edge (with respect to SEO design) to using tables.
Any good, clean code should be fine. However, tables were not really intended to manage page layout. I'm well aware of that. And the day may come when search engines start favoring other design structures. That day may never come.
If it does come, I'll deal with it then.