+1  A: 

I never heard or seen anything that indicates it matters. All you're likely to do is complicate your development. The "highly respected" site is getting good ranking because it's popular, that's all.

It's pretty well a given that search engines store the full path at some point, it's unlikely they wouldn't perform this conversion during the crawl process to remove duplicates.

I don't really follow your logic anyway. You know good structure, relevant content and popularity are the key to ranking so what makes you think you'll gain anything by spending even a minute on random optimisations like this?

SpliFF
A: 

Taken from the Search Engine Optimisation FAQ at the SitePoint Forums:

Should I use relative links or absolute links?

Absolute links. It is recommended by Google as it is possible for crawlers to miss some relative links.

If I can find the link that Google states this I'll update this post.

EDIT: This might be what the post is referring to, but I've stated my reasons as to why this might be correct in the comments.

http://www.google.com/support/webmasters/bin/answer.py?hl=en&answer=35156

EnderMB
I find that hard to believe. Really hard to believe. If a browser can follow it and it isn't javascript there's no reason anything but the most simplistic crawler would miss it (and simplistic crawlers are usually SPAM harvesters anyway).
SpliFF
To be honest I agree with you, but at least with absolute links you're 100% sure that the link will find its place. With a relative link there's always that tiny chance that the link may lead somewhere it shouldn't. I'd be shocked if Google couldn't compensate for such a problem though.Regardless, as you've said in your answer, such a change won't have any noticeable affect. It's not known if Google will save the previous path should an error occur or whether they'd want to assume that. It's another great mystery of Google.
EnderMB
That SitePoint link was written nearly six years ago - Google's algorithm has changed dramatically since then, so I wouldn't put a whole lot of faith in its validity.
Greg Hluska
The FAQ there was last updated less than two years ago, and it's no big news that Google's algorithm changes almost monthly to continuously improve. However, it's basic logic that absolute links will almost ALWAYS be crawled correctly whereas human or CMS error could force relative links to lead nowhere far more. You don't need to know anything about Google to know that.
EnderMB
A: 

I highly doubt Google will be missing any relative links. Apparently the latest version of their Crawler will even execute some javascript. Don't bother with absolute links, instead, great a good sitemap and submit it to google through webmaster tools. Yahoo and Microsoft also allow you to submit your sitemap so it might be worthwhile to look into that too - google it.

Radu
A: 

I don't think there is an answer to this question, but I will weigh in anyways. Personally, I think that using absolute URLs is the best. The web is full of crappy content scrapers. Many of the people who wrote these scrapers forget to change the original URLs (in absolute links) before they post the content onto their own page. So, in that regard, absolute URLs can turn into a really dodgy way to get a couple extra links.

If I follow that, it seems logical that absolute links would also be a great indicator of duplicate content caused by content scrapers.

A couple of years ago, I did some research into what happens to a page's search rankings when you dramatically change content/navigation (ie - in the case of a dramatic re-design). At that point, I found that having absolute URLs seemed to spook Google a little less. But, there were some problems with my research:

a) The 'absolute URL bonus' was barely quantifiable (an average of less than two positions of difference) b) The 'absolute URL bonus' only lasted a few weeks before Google settled down and started treating both pages the same c) The research is two years old and the Google algorithm has changed dramatically in that time

When I add a and b together, I'm left with a very unsettled feeling. Google gets a little weird from time to time, so the bonus may have been a fluke that I attributed to absolute URLs. Good old experimental bias.....Either way though, the difference was so slight and lasted for such a short time that I don't think it is worth spending a whole lot of extra time making absolutes!

Best of luck with your site

Greg

Greg Hluska