Wednesday, September 03, 2008
Google Chrome, SEO & Cuil - quick stuff
Hi all - it's been quite a while, hasn't it? Since my last post, we completed the enormous job - but just the first step in an ongoing, iterative process - of refreshing our website. We had over 3500 pages on the site. The web team also met regularly to audit the site & designate ownership of pages & segments of the site. In the midst of the activity, the staging area's root began to experience problems and special steps had to be taken to get control of that situation. Basically, there are tensions when we try to collaborate - tensions that can only be surmounted by standards, rules, procedures, and communication. When version control and the staging area's structure became an issue and the voluntarily-adhered to "rules" were not enough to keep the site in order, we had to shut down the broader access that several others (besides the webmaster) had to the root of the staging area. That's a whole issue worthy of further discussion, but not the point of today's entry.
Today, I wanted to mention SEO. SEO, or search engine optimization, is a whole web field unto itself. I can't begin to touch it here. Web content authors here at my institution haven't been taught to think about meta information and SEO, so understandably, there were also many pages that simply lacked meta description tags. And some pages that even lacked meaningful titles. When applying the web refresh conversion template to our pages, it knocked out meta tags - when there were any - re: descriptions.
Perhaps more problematically, the year's worth of expansion of the site and the recent site auditing had made our sitemap.xml file obsolete. But our site was too big for most free sitemap.xml generation programs to encompass. The one that I did use for free that handled our entire site took over a month to run! So SEO is a problem that we're going to continue to wrestle with in the coming months. But adding meta description tags and meaningful titles to all newly created pages and revisiting pages that are being freshly updated to ensure that they have this information is part of my current web publishing m.o.
Sadly, I tried out CUIL - the new search engine from some former Googlers - and discovered that it does not pull up our home page when you type in a search for our library by its proper name or by its URL. I've already put in requests 2 x this summer to have CUIL crawl our site, but it doesn't seem to be representing us well in its search results. There appears to be little documentation and no other contact information beyond the form that you can use to request that your site be crawled. I wonder if anyone else has had these difficulties and dealt with them successfully?
Today, I began to work with the new Google Chrome browser. I was trying out an online exhibit that we're putting up in conjunction with the airing of a History Detectives episode that was filmed at our location. I learned - the hard way - that Chrome is built on the same technology as Safari3 and that - unlike Firefox 3 and IE5+, it may render your pages in odd ways. I had a dynamic web template for the online exhibit that had 2 include files for the header and footer and and editable center of the document, where the content would be placed. When previewing the pages that used this setup in IE and the latest Firefox, they rendered fine. I ran the W3C validator - fine. But a user of an earlier Firefox and myself, using Chrome, both found that the pages would show up with no content in that crucial center - just seemingly blank space. The header and footer were fine, the center, not so much. Even the source code looked ok.
The solution? Well, it turns out that the online exhibit's content was setup to display using a layout table. This layout table appeared in the source code after the header include file and before the footer include file. The header and footer include files have divs in which the meat of those files lives. So I figured that maybe it had to do with having a div for the header, the closer div tag for the header, then just the table tags, then the opener div tag for the footer, so I enclosed the layout table in its own div. This still didn't fix the issue. Then I spotted the tiniest hint of the table off to the right of the screen, most of it hidden (yes, in my css, if the table were to go out of bounds of the container area, it would be hidden). I decided to add in a clearing div after the header include statement in the dynamic web template, which would then add itself into all of the online exhibit pages automatically. This worked. It was the clearing div tag that brought everything back. Without that tag, the page appeared to be blank between the header and footer when what really happened was that the table was treated as overflow content and kept hidden.