In the current suggesty (http://en-suggesty.wmflabs.org/suggest.html) as of 2016-01-28, searching for Obama brings up 10 Obama-related pages (yay), but "Obama" is number 9. An exact match should always be at the top of the list, period.
Another example is searching for Orca, which doesn't show the Orca page in the top 10.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | • Deskana | T121616 EPIC: Resolve technical or product blockers to enable us to do a more full rollout of the completion suggester | |||
| Resolved | • Deskana | T125112 Completion suggester should favor exact matches |
After some brief discussion, @EBernhardson confirmed that this problem is actually limited to our test server; in production, an exact match would always be favoured. So the intent of this task is resolved, and I'll mark it as such.
I think this raises a different point; there is another task in here for "Make sure that there are no substantial user-facing differences in products between test environments and production", but such a task would be so high level and require so much work that there's nothing really to be gained by actually filing it except having it clutter up the backlogs. I did want to mention that here, though. :-)
@EBernhardson: I find it somewhat troubling that the test server and production server behave differently in this substantial way. Should I file a separate phab task for that?
Unfortunatly the difference is that we don't have the production mysql databases loaded into labs. Getting a production mysql database load into the labs servers is difficult, I'm not sure what we can do about it really. Can ponder on it but my gut reaction is that this is a necessary evil.
I find it troubling too, but alas, I'm used to it. That we have no production-like test infrastructure is large, systemic problem that we have that is very, very resource-intensive to solve.