This update was made publicly available on 1st March 2011.
- Added the ability to make a given test informational only for all websites you test. This setting affects all websites you test while it is enabled, there is no per-site option (yet). Go to Account > Tests and adjust the drop down options shown. Note this optional is not available for Summary Scores or existing Informational only tests, where it would have no effect.
- Added the ability to regenerate all existing screenshots for a website. Click on Site Settings> Edit this screenshot and Update all screenshots for this website?
- Completely rewrote Speed test for performance, scalability. The new test now splits tasks into partitions, which reduces the resources required to run the test substantially. Many parts of this test have been better optimized during the rewrite.
- The Design variations test has been made considerably faster and more stable, particularly for large websites (>3,000 pages).
- The Speed test now shows the content and transfer sizes for all files. When compression is used, these values will be different.
- The Speed test now understands and refers to general forms of HTTP compression, not just GZIP. Extra documentation has been added for all recommendations.
- In the event that a report partly fails, none of the results are shown. Instead an error is shown and no credits are used.
- Added support for Google Armenia.
- Fixed bug where exporting to Word would ignore the verbosity level (i.e. “brief” or “detailed”).
- Fixed bug where installations with very high numbers of reports (at least 20,000) could experience some unexpected, random overlap between two reports.
- Fixed bug where progress meter wouldn’t advance for sites which skipped the spider.
- Fixed bug where duplicate images could be listed in the images table of the Image usage test (harmless, but confusing and slowed down testing some large sites).
- Fixed bug where the spider would fail to fork into additional processes for some old websites.
- Improved reliability of tests which download and calculate image sizes, where the images were oversized or corrupt.
- Fixed flaw in Broken Links test, which could cause it to skip over some links under unusual load conditions.
- Fixed issue with Speed test, where some large files would have their compression over-estimated.
- The Design variations test now excludes pages which are just framesets.
- Sites upgraded from v3 now migrate language settings more intelligently.
- Fixed some minor typos.
- Reduced weight of Popularity test for the Overall summary score by 28%.
- Reduced penalty for Broken Links test on Overall summary score by 50%.
- Made the Broken Links test more granular – sites will generally tend to score slightly higher when low volumes of broken links occur.
- The Spiderability test no longer awards a 0/10 score for sites with too few pages (i.e. it no longer assumes small sites are meant to be larger). This function is now largely redundant given the recent addition of the Amount of content test, and will result in higher scores for many small sites.
- The Speed test now expects websites to perform better than previously. For a perfect 10, average pages should take half a second or less to load (previously 1 second) at the target bandwidth.
- The website search no longer persists when returning afresh to the Websites tab.