One million inbound links!

I’ve been spending quite a bit of time on ardamis.com lately, giving it a new look, working at improving the site’s navigation, cultivating some inbound links, and posting more regularly. It’s rewarding to see that the effort is paying off.

Over the last few days, I’ve been concentrating on reducing page load times by sending the proper headers and compressing files.

I’ll give it some time and see how performance improves.

According to Google’s Webmaster Tools’ performance overview, with Super Cache running, a single minified CSS, a single minified JavaScript, etc. but no compression or header tweaks:
On average, pages in your site take 2.8 seconds to load (updated on Jun 28, 2010). This is faster than 53% of sites.

The chart illustrating page load times is pretty much all over the place, but at no time has the site dipped into the 20th percentile, indicating a ‘fast’ site. I’m trying to change that.

To change all the individual posts’ comments feed links to rel=”nofollow”, open ‘\wp-includes\feed-functions.php’ and add rel=”nofollow” to line 84 (in WordPress version 2.0.6), as so:

echo "$link_text";

One could use the robots.txt file to disallow Googlebot from all /feed/ directories, but this would also restrict it from the general site’s feed and the all-inclusive /comments/feed/, and I’d like the both of these feeds to continue to be spidered. Another, minor consequence of using robots.txt to restrict Googlebot is that Google Sitemaps will warn you of “URLs restricted by robots.txt”.

To deny all spiders from any /feed/ directory, add the following to your robots.txt file:

User-agent:*
Disallow: /feed/

To deny just Googlebot from any /feed/ directory, use:

User-agent: Googlebot
Disallow: /feed/

For whatever reason, the whole-site comments feed at http://www.ardamis.com/comments/feed/ does not appear among my indexed pages, while the nearly empty individual post feeds are indexed. Also, the general site feed at http://www.ardamis.com/feed/ is in the Supplemental Index. It’s a mystery to me why.

The first few solutions that I found involved adding a line to each non-WordPress page. This does indeed allow the page to incorporate WordPress tags, theme elements and styles, but there is a serious drawback to this method because of the way WP manages a web site.

When you click on the link to a WP page, or enter it into the address bar, you aren’t actually going to a file that resides at that address. Instead, WP uses that address as an instruction to pull various database entries and form an index.php page that resides in the WP installation directory. For example, while the address for this page appears to be http://www.ardamis.com/2006/07/10/wordpress-googlebot-404-error/ , the actual page is at http://www.ardamis.com/index.php.

WordPress assumes that it is responsible for every page on the site, and that there are no pages on the site that it doesn’t know about. If you try to browse to a page that WP doesn’t know about, it sends you a 404 error page instead. This is what you want it to do, so long as you don’t create any pages outside of WordPress.

But a problem arises when you do want to create a page that WP doesn’t know about. When you visit the page, WP checks the database and finds that it doesn’t exist, so it puts a 404 error in the http header, but the page does exist, so Apache sends the page with the 404 error. This behavior seemed to cause some problems with some versions of IE but none with Firefox. It did, however, result in a 404 header being given to Googlebot, so that non-WordPress pages would incorrectly show up in Google Sitemaps as Not Found.

To get around this problem and send the correct http header code: HTTP Status Code: HTTP/1.1 200 OK, I needed to require a different file, wp-config.php, and then select specific functions for use in the page. results in a page that can use all of the desired tags and theme elements and also sends the correct header code: HTTP Status Code: HTTP/1.1 200 OK

The following code results in a page that can use all of the tags and theme elements (you may need to adjust the path to wp-config.php):

init();
$wp->parse_request();
$wp->query_posts();
$wp->register_globals();
?>




*** Heading Goes Here ***

*** Content in Paragraph Tags Goes Here ***

Testing the method

Using wp-blog-header.php as the include, I created a GoogleBot/WordPress 404 Testing Page as the index.php file in the /testing/ folder. I added the url http://www.ardamis.com/testing/ to my Google xml sitemap, and waited for the sitemap to be downloaded. Sure enough, a few days later Google Sitemaps was listing the /testing/ url among the Not Found errors.

The next step was to remove what I suspected was the culprit, the include of the WordPress header, wp-blog-header.php, and see if Googlebot could again access the page. A few days after removing the include, and after the sitemap was downloaded, the Not Found error disappeared. I’m interpreting this as Googlebot once again successfully crawling the page.

The third step was to use the above code, including wp-config.php and then testing the HTTP Request and Response Header. The header looks ok, and Googlebot likes it. It looks like this does the trick.

Google has released a beta version of Google Talk, a text and voice instant messenger application. Still in its infancy, the program is very light on features, offering text and voice messaging and little else. Frankly, I hope it remains this way, rather than adding dozens of non-messaging features ala MSN Messenger.

Google Talk is compatible with any Jabber-based IM, meaning that you can communicate (via text only at this time) with Mac and Linux friends who use Apple’s iChat, GAIM or Psi.

Currently, voice chat is possible only if both parties are using Google Chat. Reportedly, this function works well even on a 56k connection.

Your contact list is populated by your Gmail address book. Google Talk requires a Gmail username and password to login. Inviting a contact from your Gmail address book who does not have a Gmail account will result in that person being offered one of your Gmail invites.

The program is free of any ads, at the moment. Here’s hoping it will stay that way. It does not currently encrypt text or voice messages, but Google claims that the release version will.

You are prompted to uninstall Gmail Notifier during installation, as that functionality is replaced by Google Talk.

I’ve been using Google Maps for some time now, as a replacement to Yahoo! Maps and the godforsaken MapQuest for plotting out jaunts around Chicago. Completely functional, though still in beta, it’s just about the coolest cartography thing on the web, challenged only by this map of Switzerland. I’m watching to see how long it takes before my home town of 20,000 gets the full satellite zoom.

If this wasn’t cool enough, a new Google beta project called Google RideFinder lets you see the distribution of cabs around your city. Google RideFinder currently tracks the movements of cabs in twelve cities. It’s interesting and somewhat troublesome to see where the cabs are concentrated and what areas they seem to be avoiding.