SEO seems to be a subject surrounded by both myths and varying opinions. Nevertheless, I would like to know if you have an opinion on how important it is to have semantically correct XHTML, that is, well-formed and valid XHTML, with regards to SEO.
- Well formed (x)HTML means there's less of a chance that the crawlers will incorrectly parse your documents. Poorly written (x)HTML leaves the possibility open that the crawlers either cannot interpret, or misinterpret your page's content causing the page to not be indexed or to be indexed incorrectly (to your detriment).
- The clues given by (x)HTML is a very large part of ranking algorithms. It's a big reason why flash isn't ever going to be as SE-friendly as(x)HTML. It simply lacks the ability to give strong clues to the search engines about the page's content (notice I didn't say Flash couldn't be crawled or indexed). For example,
<abbr title="International Business Machines >IBM</abbr>tells the search engines that IBM is short for International Business Machines. This helps them associate the two. (Not coincidentally this is also good for users).
- Page speed matters. Poorly writtn (x)HTML means it will take longer for your pages to render. If your pages take too long you run the risk of having your pages penalized (you need to be in the bottom 1% for this to happen and poor (x)HTML won't do this on its own but when combined with other factors it certainly doesn't help you in this respect).
- If your (x)HTML is properly written you can take advantage of microformats. But if you can't get your (x)HTML to format properly you certainly aren't going to get microformats to work for you.
If you want your XHTML to be properly parsed and "well understood" by the search engine crawler/indexer, then yes...it is important.
The major search engines should be able to handle minor issues but those types of mistakes may or may not hurt your ranking depending on which search engine you're talking about.
You should try to write semantic and valid XHTML even without regards to SEO, as it helps to keep your code clean.
As for search engines, Google uses very complex algorithms to rank your site, and validity & semantic code will definitely be amongs them. Nobody except for Google can tell you exactly how much does it matter, but it will definitely help.