Great tool! I used it and flunked miserably, but at least now I know what I need to do. Thanks!

Dianna, I'm glad you liked it. Remember that it's in beta, so please provide any constructive feedback or bug reports you might have.

Just for the heck of it, run this on google.com and yahoo.com – they end up with a score of 30 each. msn.com gets 65!...

I think this tool is somewhat useful, but still needs a lot of work if you want to label it as a SEO Analyzer. For starters, why not anaylze the keywords on the page? I know I'm always anaylzing my own pages for density and prominance – two major factors (in my mind) for a seach engine savvy site.

Other than that? you could take that a step further and perhaps give suggestions on keyword used and their actual popularity on the net? A simple equation might be total hits per day divided by total net saturation – which would give your users a better outlook on their chances at success for a particular keyword/phrase.

Saiko, thanks for the feedback. The tool is quite new and very much a beta. It's main focus is web standards, and it's very biased that way. After getting a lot of feedback, I added a subheader stating that more clearly. I like your ideas of beefing up the keyword area, and will work on implementing something better.

Since the original post by Philipp, we added a browsable Lynx-like viewer to each report. That way people can get a feeling how their site works semantically in the Googlebot and accessibility world.

We're working on a separate tool that will be geared more towards SERPs. Perhaps in the future we'll combine the two tools to make a more robust analyzer.

Thanks again for the feedback and checking it out.

It says I use nested tables. I don't see it. I use tables, but they aren't nested, and I use them for displaying tabular data like you're supposed to.

   http://www.windypundit.com

I really like the lynx capability; i'll probably be using that again. Its always interesting to see what the search engines see.

I also don't see the align attributes it's complaining about.

I gotta say, I'm not in love with some of the XHTML standard issues raised by the analyzer. I realize that search engines may not share my opinions, but judging by how often I see it, I'm not convinced that

   script language="whatever"

is useless.

Also, I see nothing wrong with occasional use of the style attribute. Sometimes I just want a one-off font variation and it seems silly to create a separate CSS rule for it.

Style attributes are especially useful when the XHTML is generated by a templating or scripting mechanism, as in a blogging tool or a content management system.

The templating mechanism already accomplishes the content/appearance split, and the stylesheet mechanism just complicates the context in which the code is generated. Eliminating the style attribute would remove the ability to use templating and scripting tools to change the appearance of the rendered page, except to the styles pre-defined in the stylesheet.

Finally, I miss table-based page layout. A multi-column page like my blog requires nested div elements that change around when I reorder the parts of the page. Tables are simple, direct, and easy to understand.

Mark wrote
> The templating mechanism already
> accomplishes the content/appearance split, and
> the stylesheet mechanism just
> complicates the context in which the
> code is generated.

Mark, I also sometimes use inline styles. In particular because I have experiences with a CSS becoming too complex if you keep adding "semantically correct" classes for every little thing (I ended up with pretty much just .update, .more, .note, .via, .footnote here, and try to keep it at that). However, one thing to keep in mind: even when using a templating system, only external stylesheets will achieve true cross-media... because inline styles will render on any media. E.g. what good is "color: red" on a text-to-speech system? so the code would get bloated, and what's worse, it potentially makes a system render wrongly (e.g. "color: red" with a background color provided only in the external style for media "screen" or so might cause problems). Again, I also use inline styles, but I do know that it's kinda dirty :)