Patrick Beeson

The Roanoke Times/roanoke.com has received a fair bit of press since the release of their database initiative called DataSphere, including a review I published a few days after its launch.

And in a move I hope will continue with future product releases, roanoke.com Online Editor John Jackson has posted a lengthy technical explainer about the DataSphere's programming language choice and use of search engine optimization, or SEO, techniques.

In other words, Jackson talks geek on a newspaper Web site blog.

While I'm sad to read about the continued use of ASP.NET and the atrocious markup it spews on pages, I'm glad that steps are being taken to reduce the clutter that I battled during my tenure in Roanoke as Web Standards Developer.

I'm no programmer -- that would be my brother Eric, who uses ASP.NET on a daily basis -- but I've never see another development language that shows such a disregard for Web standards out-of-the-box.

Django, a Python framework used to build the Ellington CMS used by Scripps' sites, is a godsend to designers charged with ensuring clean, semantic markup. Its template system is meant to express presentation, not program logic.

Fortunately, ASP.NET can be made to work with Web standards as Microsoft documents.

It's also good to hear that the DataSphere is receiving consideration for SEO juice, though I'm always skeptical of SEO plans that don't take into consideration every aspect of a site from the text and links to the markup. Even HTML element id and class names should be worked into a game plan for better rankings.

Using a well-crafted description is a good play. But unfortunately it's contained in a meaningless "span" element with the class name "lblDescription." The database title is also a "span," with a class of "lblName."

Both of these appear to be standard-issue class names generated by ASP.NET rendering the markup to the page.

This could be done much better by using "p" elements for the description, and an "h2" element for the database name. There's also a fair amount of non-semantic "div" elements, one using the id "div-description-info." Google knows it's a "div," that doesn't belong in the id!

Other additions to the DataSphere pages such as related links do indeed help with SEO.

A good idea to further this concept would be to create sections or topic pages for names such as "Frank Beamer." When a user clicks the linked name, they would be taken to a page listing content related to Beamer and his position as coach of the Virginia Tech Hokies. This could also be done within stories and multimedia elsewhere on roanoke.com.

Could you imagine if every reporter and copy editor were assigned to work hand-in-hand with SEO experts like DataSphere Data Delivery Editor Matt Chittum (this rocks, BTW)? This would help tremendously in terms of finding newspaper Web site content online.

This entry isn't meant to take issue with John's blog entry about DataSphere's Web strategy. Rather, I'm simply trying to continue the discussion on the topics mentioned since I specialize in Web standards and many aspects of SEO.