Closed GoogleCodeExporter closed 9 years ago
Sorry for this, I hit enter and there it was...
Feature Request: Search for Favorite Points:
How about filtering Caches with min (/max) number of Fav Points?
Original comment by Christia...@gmail.com
on 23 Mar 2012 at 8:33
You read the recent blog post on Latitude? ;)
The idea isn't new, I've discussed it with Thomas several months ago.
Actually, fav count parsing has been introduced in svn r977 on Sep 8, 2011 :)
There's a serious problem: GeoToad is running out of letters.
"-f" currently is occupied by "fun factor" limits, and I have no idea how many
people are still using it - it made sense when it was invented, but hard to
maintain (in particular, in non-English speaking regions). It's something that
has been in GeoToad for so long that I won't throw it out very quickly.
(What about "-g" for "good"? A lower limit should be enough, right?)
And there's another issue which I consider serious:
Absolute fav counts don't tell you much about a cache - it's rather the ratio
of points to (paying) visitors that would classify a cache as "must do".
Otherwise, new but good caches would be under-represented.
But there's another flaw closely related:
Very old caches would have seen a lot of visitors but only a few would give fav
points retrospectively, for several years - so even the "divide fav count by
visitors" approach would be wrong.
Also, PMO caches would have a higher ratio because non-payers can't give points
- at the same "quality level".
To summarize, fav counts, absolute or relative, are prone to distortion from
several sources. I'd like to leave this issue open for discussion - adding just
another filter isn't too hard once we have fixed the details.
Original comment by Steve8x8
on 25 Mar 2012 at 2:58
Shouldn't this be an "enhancement", not a "defect"?
Original comment by Steve8x8
on 25 Mar 2012 at 7:12
I've got a first attempt at "fav factors" _output_ ready. (No filtering yet,
stay tuned.)
Currently, only the html and text output templates are listing the "fun
factor", so I will add the "fav factor" to them. For GPX output, I will (some
day) work on issue 174 and include that information into a "fake" log entry
(like GSAK does).
The approach at the moment is to ignore the cache's pre-fav (i.e. Feb 1, 2011)
history completely (which is unfair to older caches which have seen the
majority of visitors before 2011 and probably will never catch up with newer
ones).
The formula I've chosen is
5 + ln (favpoints / totalfounds)
of course properly clipped to the 0.0 .. 5.0 range.
This will map the "super duper gets points from everyone" cache to 5.0, and the
average "gets its 1 of 10 share" cache to about 2.5, so is doesn't look
completely bad.
GC19NBJ, as an example (Bonaire is my favorite testing country, with only 14
caches), with 12 fav points out of 79 "found it" (almost one third of them by
basic members) ends up with a fav factor of 3.1 - the fun factor is 1.75 and
perhaps wrong because of different log languages.
I'll upload to my branch when I've done some more testing.
Original comment by Steve8x8
on 28 Mar 2012 at 9:03
Working on it in branch
Original comment by Steve8x8
on 2 May 2012 at 1:23
I've been using this patch for a while now, and found it to work, somehow.
It's far from perfect, mostly because of the late invention of the feature, so
old caches don't get proper points.
During a cache's lifetime, the lion's share of the "found it" entries happens
shortly (first year) after publication.
On the other hand, the "fun factor" calculations are language dependent and
need some training. Anyone come up with training files in Dutch, or Norwegian?
:) I found that fun factors may be terribly misleading...
The only thing we know for sure: a cache which has a fav factor of 0.0 after
one year is... perhaps not a big surprise. One would miss a lot of fun if using
it as a selection criterion.
Should this feature be merged into trunk? What do you think? (And should it
replace the fun factor selection in the text user interface?)
Original comment by Steve8x8
on 4 Sep 2012 at 11:50
With the exception of integrating it into the TUI (replacing "fun factor"),
this feature has been in "stable" starting with 3.16.4 three months ago.
Both "fun" and "fav" factors have their weak points, but I'm very hesitant to
throw the "fun" code out. I'd love some feedback: Please send me an e-mail
- whether you use "fun" or/and "fav" filtering
- whether you still prefer "fun" over "fav", and possibly wrote your own filter
rules
- even if you don't care at all (for example because English isn't the language
of logs in your area)
Original comment by Steve8x8
on 18 Apr 2013 at 1:36
Thanks to your copious feedback, I have become aware that there's public demand
for having fav factors in the TUI as well. (The command line options will still
be -g and -G.)
If you depend on fun factors (because you've invested your last two
years'vacations in tuning the rules), you may still have them via the -f/-F
option.
This has been in the branch for a long time, and will be rolled out with the
upcoming stable release from trunk (in a couple of weeks, I guess).
Original comment by Steve8x8
on 3 Jun 2013 at 12:27
This has been rolled out in both devel and stable 3.16.8 now. Closing.
Original comment by Steve8x8
on 27 Jun 2013 at 9:22
Original issue reported on code.google.com by
Christia...@gmail.com
on 23 Mar 2012 at 8:31