www.modxcms.com or modxcms.com

Just thought I would point out that some of your site is indexed in the SE’s with the www. and some are not. I think it has a lot to do with some links (such as forum header) have the full domain in the links without the www.

I would recommend you pick one and prevent the other with a 301 redirect. You would of course need to fix the links in the site design to use one or the other.

You know, I just did some googling and reading on this topic from various sources; SEO advice, W3C, IETF, and until now I didn’t have a very thorough understanding of these issues. Now that I do, I have to thoroughly agree with you and want to take some time to start defining some best practices in this area for all MODxers. If we’re going to proclaim ourselves as the SEO CMS, we better have a strong understanding and approach to these problems, and exemplify the proper solutions with our own site. Thanks for the kick in the rear.

This is certainly not a simple topic, but content duplication, be it via the domain, or even a specific URI, is a very valid concern when implementing a site. I’m going to go through all the issues involving content duplication and other SEO points this week, so we can start turning the best practices we do define into appropriate product features. The top two issues on my list are the "to www or not to www issue", and how to balance accessibility via multiple URLs, i.e. using weblinks, symbolic links (a future feature), and other dynamic techniques used that might produce very similar content which could be be detrimental to SEO.

Additional thoughts on the subject welcome. Here’s some links I was reading on the subject...

http://donor.kb.nl/rapporten/URI.html (a bit technical)
http://www.smart-it-consulting.com/article.htm?node=166&page=129 (consolidating domains)
http://www.smart-it-consulting.com/article.htm?node=166&page=127#www-advantages-127 (advantages of www)

...I’m sure there are many more...

If you want to see what google says about the issue, see http://www.mattcutts.com/blog/seo-advice-url-canonicalization/

Matt Cutts is responsible for Quality and Webmaster Relations at google.

I have been doing SEO since 2000 and I have seen it all (or most of it). Most CMS solutions don’t really do much to help these issues out of the box. The main reason I have become instantly addicted to ModX is because of how modular it is. This means whatever it doesn’t do, I can make it do. But the idea of having ModX follow true SEO best practices is really exciting.

I have actually been very active spreading the word about ModX on webmaster / seo forums like SitePoint.com and WebmasterWorld.com. I know the more codes / webmasters / seo’s I can get involved with ModX, the faster it’s available plugins and core code will grow.

I would love it if I could help define such seo best practices. I would even go as far as offering to become a ModX SEO advocate. Sort of how you have advocates who promote to other languages.

I was already planning on creating a post with recomendations for the core code once I was done with my SEO Strict URLs plugin. There are just certain things I am doing with that plugin that really need to be done at the core to be truely effective.

There are a number of issues that need to be discussed. Not just duplicate content, but also best practices for how plugins should be formating content. For example, NewsListing currently puts blog titles on a page in <h2> tags. It should really be using <h1> tags for the title and <h2> tages for sub topics (although sub titles may be a txt editor issue). It’s tough since so much of ModX is 3rd party snippets and plugins. You can’t control them all, but we can have certain best practices that any snippet / plugin must pass to be included at part of the base install process / core code.

And we can clearly control the output of the default content/examples from the source. What people do with it after the fact is their business.

I’ve never noticed any links to www.modxcms.com, but they’re bound to happen.

I believe that human optimisation is also important. Many people automatically add www. to an address - I often have clients who get their webmail address wrong because they try and add www. to the beginning. Also, some people go the other way. I’ve even seen non www addresses advertised that don’t actually work without the www!

You’re always going to get humans who get it wrong, but these people are people who want to visit or link to your website.

Therefore both should always "work". The wrong one should be corrected silently.

Another area of confusion is the fact that sessions and cookies can be treated differently between the www. and non-www - you could be logged in one one and not the other. Also, if you’ve changed DNS, it is entirely possible that www.x will be cached one way and x cached the other way.

Plesk-type control panels make it easy to have both www. and no-www. point to the same space, so the default setup for most people is "both working and considered separate". This is bad practice but scripts and MODx can solve this and substantiate the "SEO CMS" claim.

My personal preference is for www. for top level and no-www for subdomains (eg news.bbc.co.uk). I can justify this, but I don’t think I need to. There isn’t a "to www or not to www" issue, the point is that any variant should ’work’ (indirectly) but only one indexable URI should be returned.

And yes, Jeremy’s contributions should absolutely be taken on board. It’s nice to have someone looking at MODx with a specific perspective.

H1/H2 tags are a particular favourite of mine. I’ve seen great results from changing <div class="product_name"> (or worse, <td class..>) into H1s in other apps. Of all the "semantic markup", H1 and pals have to be the most important.

Newslisting’s successor, Ditto, allows for a much greater amount of control. However, my belief is that if the blog entry is the only one on the page its title should be in an H1 tag, but if it’s a list of them, the titles should be in H2 tags. Under that, it is likely that the MODx template will output the title in H1 and Ditto will never actually output a H1 tag. To me, the page with the blog entry is more important than the listing page, which would probably have a H1 on "Recent Blog Entries" or somesuch. Am I wrong here?

Newslisting’s successor, Ditto, allows for a much greater amount of control. However, my belief is that if the blog entry is the only one on the page its title should be in an H1 tag, but if it’s a list of them, the titles should be in H2 tags. Under that, it is likely that the MODx template will output the title in H1 and Ditto will never actually output a H1 tag. To me, the page with the blog entry is more important than the listing page, which would probably have a H1 on "Recent Blog Entries" or somesuch. Am I wrong here?

Yes you are correct. H1 tags should always contain the topic of that particular page. The same goes for the title and description tages. It is what they are meant for. The search engines use the <H*> Tags how they where meant to be used. H1 is the main heading of the page. H2 is the next main heading. Then H3 and so on.

I’m trying to finish up some other things with the Strict URLs plugin tonight, but I’ll see if I can put together an outline of what I consider to be topics covered by a best practices document.