How to test a broken link checker

Rusty Results Research

The test

Based on the test criteria and the weighing factors, the broken link checkers are tested with the following test sites:

  1. A Different End Of The Internet for checking:
    1. Examines links to other resources (the src attribute) - Should
    2. Examines structural links (the link tag) - Should
    3. Respects robots.txt of the-linked-to-website - Could
    4. Presents unchecked links due to the robots.txt directive at the-linked-to-website - Must
  2. httpstat.us for checking:
    1. Detects link error 404 - Must
    2. Detects all link errors in the 400 and 500 range - Should
    3. Detects redirected links - Should
  3. Dead Link City for checking:
    1. Examines links in stylesheets - Could
  4. Rusty Links for checking:
    1. Examines internal links - Must
    2. Examines relative links in combination with the base tag - Should
    3. Examines inline (anchor) links - Could
    4. Examines external links - Must
    5. Examines standard hyperlinks (the a tag) - Must
    6. Ignores structural links with the dns-prefetch attribute - Could
    7. Examines links with other schemes (at least FTP://) - Could
    8. Examines Unicode links - Should
    9. Detects blacklisted links (at least blacklisted by Google Safe Browsing) - Could
    10. Respects robots.txt - Should
    11. Allows you to test websites with up to 500 links - Must
    12. Allows you to test websites with up to 2000 links - Should
    13. Allows you to test websites with up to 5000 links - Should
    14. Allows you to test websites without a restriction on the number of links - Should

The following tests will be done by hand:

  1. Allows you to take a subdirectory as base for the test - Should
  2. Allows you to start the test with a single click (after entering the to be tested URL) - Could
  3. Allows you to use your mobile device - Could (only for online link checkers)
  4. Allows you to use a graphical user interface - Could (only for offline link checkers)
  5. Allows you to test an unlimited number of websites - Should
  6. Allows you to retest a website unlimited - Must
  7. Allows you to save the results (internal or by an CSV export) - Could
  8. Presents the results in one overview - Should
  9. Presents 200 Ok and broken links - Could
  10. Presents the webpage where an issue is found - Should
  11. Presents in detail (e.g. linenumber, source viewer) where an issue is found - Could
  12. Presents in a clear way the results when there are no issues found - Could
  13. Test at least 5 links per second - Could
  14. Test at least 10 links per second - Could
  15. Test at least 15 links per second - Could
  16. Test at least 20 links per second - Could
  17. Test at least 25 links per second - Could

Each criterium will get the label PASSED or FAILED, depending on the test results.

The score will be calculated by counting the PASSED Shoulds and Coulds. Where a Should is 5 points and a Could 1 point.

The last step is to translate the total score to a percentage of the maximum score.

Note that the speed in links per second will be measured while testing Rusty Links.

As extra the results of testing Dead Link City will also be mentioned.

Note that due to the feedback to Dead Link City their robots.txt file updated. Broken link checkers tested with the old robots.txt directive are scored as nn/75. Broken link checkers tested with the updated robots.txt directive are scored as nn/74.