Haddock Directory A directory of 5,000 URLs on more than 300 pages, all of which are generated by Frontier

Phil Gyford describes the project:
The Site
The Haddock Directory is a Yahoo-style catalogue of websites which are posted each day to the London-based Haddock mailing list. Since going online in 1997 it's grown to over 5,000 URLs on more than 300 pages, all of which are generated by Frontier. The site's innovative (if space-intensive) heirarchical navigation system is also output by Frontier, a feat which would be impossible by hand. Aside from these static html pages the their is a search engine which is written in Perl, using a text file created by Frontier as its source.

Backstage
Each of the thousands of URLs is stored in a table along with its site's title, comment, date and the categories in which it should appear. A second table holds the information for the category heirarchy -- its structure, page titles, meta tag keywords -- to which Frontier is ideally suited.

First a script works through the URL table and copies details for each entry to the pages in which it should appear. Then each page is sorted either alphabetically or by date. Next, the navigation for the top part of each page is created by a script which works through the category table and generates the appropriate html. As each page is rendered this navigation and the right-hand column is included.

When it comes to adding new entries each day, a Frontier script grabs the URL and title of the page from Netscape, along with a comment from the clipboard, and creates a new entry in the URL table. The only part of the process Frontier can't manage is deciding which categories each new entry should be assigned to. A script then uploads a new front page with the last few days' entries and a channel for my.userland.com.