Opened 15 years ago
Closed 15 years ago
#435 closed defect (fixed)
Remove score_evenness() functions in matchspy.h
Reported by: | Richard Boulton | Owned by: | Richard Boulton |
---|---|---|---|
Priority: | normal | Milestone: | 1.1.4 |
Component: | Library API | Version: | SVN trunk |
Severity: | normal | Keywords: | |
Cc: | Blocked By: | ||
Blocking: | Operating System: | All |
Description
The score_evenness() functions in matchspy.h were intended to be a way to determine how good a particular set of facet results is. However, after experience using the facet support in real-life situations for a couple of years, I'm very dubious that they ever give useful results: each time I've tested them out, the choice of facets produced by score_evenness() has been pretty poor: a manual fixed choice of order of facets has always been an improvement.
It's possible that the calculation from score_evenness() might be useful in some situations I've not come across, but it's not hard to implement an equivalent outside Xapian.
I therefore propose to move the score_evenness() functions out of Xapian trunk. I'd leave them on the MatchSpy branch for now, but if we've not found a real use for them by the time we've worked out a clean API for the MultiValueCountMatchSpy, I think we can drop them.
Change History (2)
comment:1 by , 15 years ago
Component: | Other → Library API |
---|---|
Type: | task → defect |
comment:2 by , 15 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
Removed from trunk in r13883. Leaving it in limbo on the matchspy branch for now: now's the time for anyone who misses it to ask for its resurrection, but marking this ticket as closed since I'll be stunned if anyone does.
Seems reasonable. Don't forget to remove it from categorisation.rst.