Automatic favorites import using browser history

TailRank launched a bunch of new features today focused on improving the meme tracker's ability to recommend new sources and hot posts. Users can now have their favorite weblogs automatically recognized without the need for relatively confusing terms such as OPML, RSS, or blogrolls. Favorite sites are imported using a user's browser history. I think this technology will be common across many sites in the next 6 months, and here's how you do it.

The user clicks on an "Auto Configure" button that initiates the magic. Clicking the button loads a list of links that are potentially sites visited by the user. In TailRank's case the top X blogs are loaded into an iFrame hidden from view using CSS height and width values of zero.

You can then cycle through the list of links using JavaScript, identifying the links with an anchor pseudo-class of visited. Pass each visited link back to the server and you now have a list of URLs stored in a user's browser history.

It's quick and easy for the user with a low barrier to entry.

Tags:

Posted on March 2, 2006 at 2:35PM

Updated at 4:26PM

Permalink

Comments [1]

By Niall Kennedy, a web technologist from

San Francisco, California, United States

1 Comments

Martin Dittus on March 3, 2006 at 2:17 AM wrote: #

Man what a great idea -- I initially stumbled over "the top X blogs are loaded into an iFrame", but when I realized that all you are doing is creating the _links_ to those pages I was sold. You can easily check hundreds of URLs that way, if the user is willing to spend the couple of seconds this takes.

Potential problems I see:

non-unique URLs (with or without 'www'?) -- this e.g. is a problem when you use del.icio.us to watch URLs; some blogs/sites are inconsistent in their usage of URLs, and of course people occasionally type URLs themselves, so it might be that a user has been to http://example.com/ but not http://www.example.com/
People who are using aggregators might not have visited a site's homepage for a long time, but they might occasionally visit article pages.
Both problems can be approached by using links that not simply point to the site's homepage, but e.g. to an image file that is used on every page (e.g. the logo), or maybe even to a stylesheet.

But of course that's harder to automate, and introduces other problems.

Add a comment

Some comments may be placed in a moderation queue to ensure topical relevancy. You may contact author Niall Kennedy directly if you prefer to comment privately.

Commentary on "Automatic favorites import using browser history":

Subscribe to new comments

Martin Dittus on March 3, 2006 at 2:17 AM wrote: #
Man what a great idea -- I initially stumbled over "the top X blogs are loaded into an iFrame", but when I realized that all you are doing is creating the _links_ to those pages I was sold. You can easily check hundreds of URLs that way, if the user is willing to spend the couple of seconds this takes.

Potential problems I see:
- non-unique URLs (with or without 'www'?) -- this e.g. is a problem when you use del.icio.us to watch URLs; some blogs/sites are inconsistent in their usage of URLs, and of course people occasionally type URLs themselves, so it might be that a user has been to http://example.com/ but not http://www.example.com/
- People who are using aggregators might not have visited a site's homepage for a long time, but they might occasionally visit article pages.
- Both problems can be approached by using links that not simply point to the site's homepage, but e.g. to an image file that is used on every page (e.g. the logo), or maybe even to a stylesheet.
But of course that's harder to automate, and introduces other problems.

Automatic favorites import using browser history