Open Bug 1042604 Opened 6 years ago Updated 5 years ago
Measure how many searches, for each engine, result in a final URL that can be parsed
Every time a new search is started from one of the available locations in the browser, we start loading a known URL from the search engine, providing the search terms. The engine may then redirect to a different URL, so that the final location displayed in the address bar is different. We would like to know, in a distinct way for each search engine, how many times the final URL displayed in the address bar can be parsed back to the original search terms using the function defined in bug 1040721, and how many times this is not possible. We may want to detect the third case where the URL is recognized as a search but the terms do not match the original, which might be normal because of corrections made by the engine, or indicate an encoding issue. Search engines that use POST requests or redirects to a different path are currently expected to always fall into the "not possible" case for now, but this should still be measured. This has the requirement of distinguishing each engine, and has a similarity with the current total search count detection. Given this, we should determine whether we should use Firefox Health Report or Telemetry with a defined histogram for each default engine.
Blake, I think this bug is related to the work you're doing in bug 1042604, though a bit more challenging.
Priority: -- → P4
You need to log in before you can comment on or make changes to this bug.